SGI's vs PC's?
Frederick Haab
This debate is nearly as meaningless as the PCs vs. Macs debates!
Lets look at a real world example, shall we?
Turner Production/Effects:
2 Onyx RE2 systems
2 Onyx IR systems
3 Indigo2 systems (2 high impact)
2 Indy R4k's (as far as I know)
1 Indy R5k (Indy studio)
The Onyx's do things that *NO* PC is doing right now...I don't
think we have any debate over that. They're running Virtual Set
software, real time character animation of complicated scenes
(believe me, no Alphas/PC's are doing what we're doing with the
Onyx's). For modelling and animation the animators use the
Impacts but then render on Onyx's using Alias and SoftImage.
The Indy's are there for system administration and development
and to pick up little peices when the other machines are busy.
The R5k Indy even has Alias animator (and I defy any Mac to run
Photoshop as fast as this machine does). The two RE's also
run flame and the non-impact Indigo2 runs flint.
Do we have PC's? Some of the people here have PC's for the
business end of the Production department, but none are running
any graphics software, AFAIK.
Do we have Mac's? Most of the Macs in our editing suites are
used for email, sometimes we use them for tranfering files.
Very often clients give us images/textures/whatever on Mac
disks. They're there if we need them, but it's very rare.
Most images done internally are done on Paint Boxes.
I'm *not* saying PC/Alpha/Macs aren't good, but there is *no*
debate when comparing high end to high end. There is little
debate comparing something like an R10k Maximum Impact with
even the best PC ever made...the artists just don't get realtime
feedback/screen updates as fast on any other non-SGI. This
is what SGI's do. If you are raytracing or running a radiosity
solution then any machine with fast processing power is going
to do a good job, but *nobody* beats SGI's for realtime
applications and feedback. YES, PC's are getting faster, and
the feedback may be good enough that it's not worth getting
an SGI for a lot of people, and people without the funds just
have to live with it, but for the real complicated stuff it's
pointless when the animator tries to rotate an object in a
highly complex scene and have to wait forever for the screen
to update.
So what about low end? We have low end SGI's mostly to make
things compatible across the board. I develop on one of the
Indy's...I develop applications that run on the Onyxs, I can't
do that with a PC. I also do some modelling and need to do
R&D with our real time character animation software which will
ONLY run on SGI's - yes it sucks on the Indy, but I develop
pieces on the Indy and we put it all together on the Onyx.
This is a typical real world example of who is using/needs
SGI's. For the high end stuff nobody beats them, but for
non-realtime applications you can probably save tons of money
and use P/A/M (Pentium/Alpha/Mac) based applications.
This is just such a stupid argument, the point I'm making is
that there is a need for low end SGI's as much as there is
a need for high end SGI's, but if you don't NEED an SGI
then a P/A/M render farm is a less expensive and worthy
alternative...it depends who you are and how much money you
have.
Also PLEASE lets not get into a debate about realtime graphics...
there certainly exist P/A/M systems doing realtime graphics,
but not like a high end SGI system can do...not even close.
--
-=-=- Frederick Haab -=- Software Developer -=- Turner Production -=-=-
-=-=- fred...@turner.com -=- 404.885.0317 -=- fax: 404.885.0757 -=-=-
Jean-Paul Maurel
Frederick Haab <fred...@turner.com> wrote:
>This is a typical real world example of who is using/needs
>SGI's. For the high end stuff nobody beats them, but for
>non-realtime applications you can probably save tons of money
>and use P/A/M (Pentium/Alpha/Mac) based applications.
>This is just such a stupid argument, the point I'm making is
>that there is a need for low end SGI's as much as there is
>a need for high end SGI's, but if you don't NEED an SGI
>then a P/A/M render farm is a less expensive and worthy
>alternative...it depends who you are and how much money you
>have.
>Also PLEASE lets not get into a debate about realtime graphics...
>there certainly exist P/A/M systems doing realtime graphics,
>but not like a high end SGI system can do...not even close.
>-=-=- Frederick Haab -=- Software Developer -=- Turner Production -=-=-
>-=-=- fred...@turner.com -=- 404.885.0317 -=- fax: 404.885.0757 -=-=-
No doubt SGI has ruled the high-end graphics up to now.
With the raw speed of DEC's Alpha chips, faster and much cheaper, I
think there are great opportunities for software developers to develop
softwares offerring features and functionalities that so far exist on
high-end systems.
R.S. Little
Gee thats funny most dec's alpha are running DEC"S osf/1 unix
Reto Koradi
Alan Boucek wrote:
> Perhaps you haven't priced the fast Alpha systems, since from what
> I've seen they're neither significantly faster (application
> performance, not clock speed is what you need to look at) nor
> significantly cheaper than SGI machines.
I've never seen an Alpha workstation, but I'm using servers with
Alpha processors, and saying that they don't deliver application
performance is just plain wrong. These beasts are amazingly fast.
Maybe SGI is getting into similar regions with the R10000 now,
but before that they didn't even get close to an Alpha. Of course
there are other factors, but for raw floating point performance
nothing beats the Alpha. Even the cheaper Crays don't have a chance.
--
Reto Koradi (k...@mol.biol.ethz.ch, http://www.mol.biol.ethz.ch/~kor)
George White 6-8509
Several points that I have not seen in this thread:
1. processor speed is often a minor factor in graphics performance.
Memory bandwidth is important in many operations. Some benchmarks
indicate higher memory bandwidth for PCI-bus systems than for low-end
SGI systems.
2. UNIX is very well instrumented, which can be a big help when
trying to understand a processing bottleneck. Under UNIX, several
users can share a system without fighting over the configuration.
Thus a group of Mac or PC users can benefit by pooling some resources
to get a high-end SGI system that can be shared, with priority being
given to the bigger jobs.
3. a Mac, PC, or SGI intended to run Photoshop is going to need a
good monitor, a big chunk of RAM, a big chunk of disk, and a big chunk
of setup and maintenance time. In my experience, the time element is
often the biggest cost, and is often underestimated. The difference in
hardware cost for comparable SGI/PC/Mac systems is minor.
--
George White <aa...@chebucto.ns.ca> <gwh...@bionet.bio.dfo.ca>
Bedford Inst. of Oceanography, Nova Scotia, Canada
Ben Cannon
gwh...@trevnx.bio.dfo.ca (George White 6-8509) wrote:
>Several points that I have not seen in this thread:
>
>1. processor speed is often a minor factor in graphics performance.
>Memory bandwidth is important in many operations.
Absolutely!
>Some benchmarks
>indicate higher memory bandwidth for PCI-bus systems than for low-end
>SGI systems.
??? I guess if you call an Indigo R3000 a low end SGI system, yes.
The Indy has significantly higher memory bandwidth that any PC, though some DEC alphas
using a 128-bit bus may, in theroy move data faster... however, PCI is 32-bits wide.
(sometimes 64)
>3. a Mac, PC, or SGI intended to run Photoshop is going to need a
>good monitor, a big chunk of RAM, a big chunk of disk, and a big chunk
>of setup and maintenance time. In my experience, the time element is
>often the biggest cost, and is often underestimated. The difference in
>hardware cost for comparable SGI/PC/Mac systems is minor.
Good point, although the huge cost bloating SGI has done is hard for even some
buisnesses t swallow. Try the $2,000/year fee if you want a legal copy of a working OS!
Oh, and you're NOT supposed to buy a SGI form SGI!! They deliberatly inflate their prices
to protect the retail VARs!!! buy form a VAR, even SGI wants you to!
The only reason SGI Direct exists is for large orders, you get a discount of about 30%
ordering 10 machines or more, which is about what you'd get buying a single machine form a
VAR. Reason is: VARs can't handle an order of 30 Indigo^2s that need delivery tommorow :)
--
Ben Cannon. art...@a.crl.com
"If you think it more, you will know it." -Joly
________________________________________________________
I have moved! To http://168.75.121.10/ben.htm
Not much, however the old adress no-longer works.
Kevin Palmer
-- There is an excellent article in the new SIGGRAPH issue of Computer
Graphics World on the difference between True Workstations and the
so-called NT Workstations. It is in the August issue on page 21.
____________________________________________________________
Kevin Palmer Applications Engineer/Intern at Alias|Wavefront
Home Indy owner using Alias PowerAnimator/StudioPaint
Wavefront Composer/3Design and ElectroGIG GIG3DGO
____________ http://web-hou.iapc.net/~kpalmer ____________
Jim Mc Morrow
SGI vs NT PC
I look at it this way:
A ford and a Ferrari will both drive you to work. But only one is a
Ferrari and if it's a company car,all the better.
There may be an element of elitism about SGI's.
But secretly most of us want Ferraris to drive, we just don't want (or
can't afford) to pay for them.
Take note SGI, Ferrari is a small company compared to Ford....I wonder
why?
Jim
Erik Johnson
Jim Mc Morrow wrote:
> SGI vs NT PC
> I look at it this way:
> A ford and a Ferrari will both drive you to work. But only one is a
> Ferrari and if it's a company car,all the better.
> There may be an element of elitism about SGI's.
> But secretly most of us want Ferraris to drive, we just don't want (or
> can't afford) to pay for them.
> Take note SGI, Ferrari is a small company compared to Ford....I wonder
> why?
So take a ford stock car racer and pit it against a ferrari on the track.
the ferrari may be elegant and beautiful, but it is overpriced and doesn't
offer the performance one might expect from that price. Not to mention
the absurd cost of parts and maintenance. Your analogy is a good one.
And sure, I'd enjoy a spin in a ferrari but I'm not sure I'd want to own
one (though ford wouldn't be my first choice either).
-Erik Johnson
er...@phidias.colorado.edu
http://phidias.colorado.edu/vgallery.html
John Zulauf
>
> Jim Mc Morrow wrote:
>
> > SGI vs NT PC
> > I look at it this way:
> > A ford and a Ferrari will both drive you to work. But only one is a
> > Ferrari and if it's a company car,all the better.
Erik Johnson wrote:
> So take a ford stock car racer and pit it against a ferrari on
> the track. the ferrari may be elegant and beautiful, but it is
> overpriced and doesn't offer the performance one might expect
> from that price. Not to mention the absurd cost of parts and
> maintenance.
Hmmm, do you mean take a showroom stock Ford on the track against a
Ferrari, or take a Ford NASCAR "stock" car on the track against the
Ferrari? If it's the former be prepared to had over your pink slip, if
it's the latter, be prepared to hand over your wallet. Remember, stock
cars ain't -- and a Winston Cup car makes a Ferrari's price look like
chump change, as for maintenance, your typical Ferrari doesn't need an
engine rebuild every 500 miles (or less).
Well maybe that IS a good analogy...
--
John Zulauf jzu...@aw.sgi.com Alias|Wavefront, Santa Barbara
his opinion and NOT necessarily his employer's
Question Anarchy.
Allen Akin
In article <321B97...@batelco.com.bh>,
Jim Mc Morrow <mo...@batelco.com.bh> wrote:
| I look at it this way:
| A ford and a Ferrari will both drive you to work. But only one is a
| Ferrari and if it's a company car,all the better.
It's a thought-provoking analogy. Recently I read that Saab,
Mercedes, Infiniti, Volvo, Lexus, and BMW each have less than 1% of
the world car market.* Even though these ``vendors'' have dramatically
smaller market shares than GM, Honda, etc., they manage to maintain
viable businesses and provide well-respected products. Perhaps the
computer industry will evolve into a similar situation.
Allen
* This statistic is attributed to the October 1995 issue of Automotive
News.
Ben Cannon
Erik Johnson <er...@phidias.colorado.edu> wrote:
>So take a ford stock car racer and pit it against a ferrari on the track.
>the ferrari may be elegant and beautiful, but it is overpriced and doesn't
>offer the performance one might expect from that price. Not to mention
>the absurd cost of parts and maintenance. Your analogy is a good one.
Well, a Ford Stock car runs about a Half a Million dollars (if I wanted to buy
one right now) for a new one.
Let's not pit apples against oranges and compare the $500K Ferrari F50.
Equivlent prices, yes?
Although I haven't done this myself, the straight line performance of the
Stock car may be slightly faster than the Ferrari (I don't keep up with stock
cars that much, the Ferrari can go 203mph..)
However, throw a few turns into the mix, and the Ferrari will eay your presious
Ford for lunch by a huge margian!
And I can drive to the Quickie Mart to get grocories in my Ferrari :)
Startling how close an alologie this is, the PC(ford) may render(top speed) a bit
faster than the SGI(Ferrari) but realtime graphics(handeling) on the SGI with real
applications (real race trakcs, which have to have 4 turns minimum) is MUCH better!
--
Ben Cannon. art...@a.crl.com
"Hsitory is an Illusion cuased by the passage of Time"
-Douglas Adams.
"Time is an Illusion caused by the passage of History"
-Douglas Adams.
________________________________________________________
http://www.10mb.com/sgi/BEN.HTML The fractional T-3 sure beats
my old 28.8 connection! ( http://168.75.121.10/ben.html )
Allen Akin
In article <4vimfm$s...@nntp.crl.com>, Ben Cannon <art...@a.crl.com> wrote:
|
| Startling how close an alologie this is, the PC(ford) may render(top speed) a bit
| faster than the SGI(Ferrari) but realtime graphics(handeling) on the SGI with real
| applications (real race trakcs, which have to have 4 turns minimum) is MUCH better!
Reminds me of one of John Mashey's favorite benchmarking analogies:
Driver: My car is faster than your plane!
Pilot: Why do you say that?
Driver: I can prove it. Put your plane and my car at
the end of the San Francisco airport runway,
start them at the same time, and my car will
reach the end of the runway before your plane
does.
Pilot: But what about right after that point? My
plane will fly over to Oakland, but your car
will fall into the Bay and sink!
Driver: What does ``fly'' mean?
Note that like PCs and workstations, the car has better
price/performance than the plane. :-)
Allen
Keith J. Cambra
In article <4vihse$h...@fido.asd.sgi.com>, ak...@tuolumne.asd.sgi.com (Allen
Akin) wrote:
>It's a thought-provoking analogy. Recently I read that Saab,
>Mercedes, Infiniti, Volvo, Lexus, and BMW each have less than 1% of
>the world car market.* Even though these ``vendors'' have dramatically
>smaller market shares than GM, Honda, etc., they manage to maintain
>viable businesses and provide well-respected products. Perhaps the
>computer industry will evolve into a similar situation.
>
>Allen
>
>* This statistic is attributed to the October 1995 issue of Automotive News.
yeah, i've read similar info. in fact, Mac Evangelist, Guy Kawasaki, has
made that
point many times to divert attention away from Apples rapidly dropping
marketshare.
there are several differences btwn the automotive industry and computer
industry.
for one, auto are much more "durable," that is last longer w/o going
obsolete, than
computers. autos are a higher ticket item and profits are larger, not by
% but in $ amount.
the auto market is alot larger so it can sustain more players. also, the
auto industry doesn't
have one dominate player like Microsoft & Intel who can impose their
proprietary standards
on the entire industry. eventhough, GM is huge, as far as i can tell, it
doesn't impose
or control any important standard that all other companies must follow and
pay GM
royalties. we have seen some consolidation in the auto industry in the
last few years.
as well as a few low-end manufacturers folding - Daihatsu & Yugo. ok, not
great examples.
also, i think that there's alot more personal preference happening in auto
purchases
than in computer hardware. personal preference comes more into play on
the software side
since you really "customize" your computer system via software.
also notice the manufactures that were mentioned are all considered
high-end autos.
(Saab, Mercedes, Infiniti, Volvo, Lexus, and BMW) they can survive on
lower marketshare
because they're margins are fat. although, last year Mercedes lost $1
billion. Lexus is really doing
a number on them.
also, when Guy mentioned his autos many were brands (or models) and not
entire companies.
maybe each brand/model of a company has a small marketshare but when you
add them up
it's much larger. also, just because they're in business, doesn't
necessarily mean that
they're profitable.
in most markets only the top 2 brands make money. generally speaking you
need a 15-20% marketshare to sustain profitability over the long run. that
would be 5-6 companies if all had similar marketshare. if any one company
has a 70% or larger marketshare they have a monopoly. if the two or three
top company's combine marketshare is 70% or greater they have an
oligopoly.
btw - that info. is from a famous Boston Consulting Group study.
if you look at various markets, you will notice that many times the race
boils down to just two companies. especially in mature markets. just
think - Coke vs. Pepsi, Nike vs. Reebok,
Apple vs. IBM, Hertz vs. Avis, etc.
as for the computer hw industry, it's maturing and consolidating.
anyway, that's what i think.
Hugh LaMaster
Interesting discussion. From my perspective, one of the key features
is price/performance. SGI, having staked out realtime & 3D graphics
as its turf, needs to stay on or above the price/performance curve
that alternatives, such as PCs, can offer.
At the low-end, that means adopting PC technology, like the PCI bus
and peripherals, and getting the entry-level price, including software,
down.
At higher performance levels, that means running key applications
at least 10X as fast, if the box costs 10X as much. It may surprise
some people, but, until the latest (last few weeks) generation
of PCI bus boards came out, I think SGI was often in that position.
Now, the ball is in SGI's court regarding whether it can respond,
cut costs on the low end and raise performance on the high end
to compete with the new generation of Pentium Pro systems with
3D graphics cards.
[IMHO, etc. All disclaimers apply.]
--
Hugh LaMaster, M/S 233-18, Email: Please send ASCII documents to:
NASA Ames Research Center Internet: lama...@ames.arc.nasa.gov
Moffett Field, CA 94035-1000 Or: lama...@george.arc.nasa.gov
Phone: 415/604-1056 Disclaimer: Unofficial, personal *opinion*.
Ben Cannon
lama...@viking.arc.nasa.gov (Hugh LaMaster) wrote:
>Interesting discussion. From my perspective, one of the key features
>is price/performance. SGI, having staked out realtime & 3D graphics
>as its turf, needs to stay on or above the price/performance curve
>that alternatives, such as PCs, can offer.
>
>At the low-end, that means adopting PC technology, like the PCI bus
>and peripherals, and getting the entry-level price, including software,
>down.
I agree, to a point. It would be worthless for SGI to adopt the PCI bus, there would be
almost no cost savings, you would need to pay a huge amount of money for software drivers for
(as an example) a 3Com Network card.
If SGI stays with IRIX, that is.
Now, if they move to NT, then we see a PCI based NT machine with a price premium to SGI
Versus an identical machine I can build with spare parts (MIPS processor, motherboard, etc)
at home for much less.
Why chop the Indy's bus in half? it's already twice as fast PCI! I say we give the Indy 4
processors and the HIO bus form the Onyx! Of course, we can;t just slap the same chipset
in, it will need to be implimented in either a one or two chip set to be affordable and fast.
>At higher performance levels, that means running key applications
>at least 10X as fast, if the box costs 10X as much. It may surprise
>some people, but, until the latest (last few weeks) generation
>of PCI bus boards came out, I think SGI was often in that position.
>Now, the ball is in SGI's court regarding whether it can respond,
>cut costs on the low end and raise performance on the high end
>to compete with the new generation of Pentium Pro systems with
>3D graphics cards.
I want to see single-board Maximum IMPACT level graphics on 4-way SMP Indys
Priced around $15,000. I want to see Infinite reality level performance for $35,000 from
the Indigo^2 running a 400mhz R10k, or three.
And I want to see the Onyx Use Gallium Arsinide ICs and possibly fiber optics to leap beyond
the current and future state of the art.
Belive me, SGI is not sitting on their hands with the affront form PCs, they just didn't
expect so many claims that slow boards were accually "faster" than their IMPACTs!
Trouble is, these outrageous claims make perfect sense to many people :)
Cheers!
Sam Richards
Ben Cannon wrote:
> I agree, to a point. It would be worthless for SGI to adopt the PCI bus, there would be
> almost no cost savings, you would need to pay a huge amount of money for software drivers for
> (as an example) a 3Com Network card.
I dont have a choice
for 100-BaseT cards for Indy or Challange. Only with the Indigo-2
machines with its EISA card slots
can I buy a (3rd party) card. You have to admit that the current SGI
range of machines is really
kinda rediculous when it comes to Buses. The Indys, Indigo-2s and
Challange machines all have different
busses, nothing is compatable between the different ranges, its almost
IBM like.
>Why chop the Indy's bus in half? it's already twice as fast PCI!
The strange thing is that my Indy ATM card out performs the VME ATM card
of the Challange-L!
So you do have a point. However I have know idea what the performance of
the PCI bus ATM card is.
But if they are going to carry on having machines with two different Bus
types, at least make the
standard one PCI, EISA is getting rather old, not to mention VME!
Sam.
Ben Cannon
Sam Richards <s...@blueskyprod.com> wrote:
>Ben Cannon wrote:
>> I agree, to a point. It would be worthless for SGI to adopt the PCI bus, there would be
>> almost no cost savings, you would need to pay a huge amount of money for software drivers for
>> (as an example) a 3Com Network card.
>But at least I might have a chance to buy a 3com network card. Right now
>I dont have a choice
>for 100-BaseT cards for Indy or Challange.
True, though SGI is reportedly about to release a 100baseTX card (I know)
There's always FDDI for the Indy and Challenge.
I would be amazed of SGI's next systems didn't come with 100-baseTX on the motherboard.
And look for Fibre Chanel or maybe 32-bit VeryWide SCSI (with lower clock frequencies than
ultra to keep cabel legents reasonable)
>>Why chop the Indy's bus in half? it's already twice as fast PCI!
>The strange thing is that my Indy ATM card out performs the VME ATM card
>of the Challange-L!
>So you do have a point. However I have know idea what the performance of
>the PCI bus ATM card is.
Well, the Indy's GIO bus transfers data at aobut 256mbytes/sec
Wheras PCI stops at 128mbytes/sec.
Also, on PCI systems, the RAM is on the PCI bus (save for a few Dec Alphas..)
Wheras on the Indy it gets it's own 400mbyte/sec bus!
(note that my predictions are just speculation on my part :)
Kent Sandvik
In article <4vkppb$p...@onramp.arc.nasa.gov>,
lama...@viking.arc.nasa.gov (Hugh LaMaster) wrote:
>At higher performance levels, that means running key applications
>at least 10X as fast, if the box costs 10X as much. It may surprise
>some people, but, until the latest (last few weeks) generation
>of PCI bus boards came out, I think SGI was often in that position.
>Now, the ball is in SGI's court regarding whether it can respond,
>cut costs on the low end and raise performance on the high end
>to compete with the new generation of Pentium Pro systems with
>3D graphics cards.
Something that is important to understand is that modelling the real world
is a tough task, and there's plenty of space to improve the high end of
graphics systems. And that's where the bleeding edge customers are as well.
It's just fine that the 3D graphics industry is growing with the help of
cheap PC solutions, but Hollywood and other companies dealing with projects
that are barely possible need the help of a company that is also interested
in solving the hard problems, not providing cheap 'me too' systems that
handles cases that won't help such companies.
In addition, by solving these hard problems, this company generates
technology that could then be later deployed as cheap entry level systems.
This way the low end market will also gain from such R&D.
Oh, that was my take on promoting the company I work for :-).
Cheers, Kent
Kent Sandvik, Silicon Graphics, Inc. Member of Technical Staff
Web Applied Engineering (SGI internal http://pyros/webappengr/)
Email: san...@sgi.com Phone: +1 415 933-6417
Web: http://reality.sgi.com/sandvik/ (SGI internal http://kent.engr/)
Dave Frederick
Alex P. Madarasz, Jr. wrote:
>
> Erik Johnson wrote:
> > Jim Mc Morrow wrote:
> > > I look at it this way:
> > > A ford and a Ferrari will both drive you to work. But only one is a
> > > Ferrari and if it's a company car,all the better.
> > > But secretly most of us want Ferraris to drive, we just don't want (or
> > > can't afford) to pay for them.
> > > Take note SGI, Ferrari is a small company compared to Ford....I wonder
> > > why?
> >
> > one (though ford wouldn't be my first choice either).
>
>
> Has Ferrari gone belly up just because Chevy makes a Corvette, or Ford
> makes a [?], or Dodge makes a Viper, or ... [what the hell] Yugo makes
> a POS?
>
> Folks who say SGI will roll over and die because "PCs" finally have
> reasonably fast 3-D graphics are just plain missing the point - SGI
> makes
> products which provide the performance at the price points their
> customers want, and their business model allows them to keep doing this
> while growing their business and making a tidy profit too. SGI sells to
> one or two [snicker] other markets besides the "I want lots of static
> rendering power for $9.95" one, and there's _no_ evidence that these
> markets are drying up at the rate suggested in these threads, nor is
> there _any_ evidence that SGI has forgotten how to innovate to keep well
> ahead of "PC" technology.
>
> -----------------------------------------------------------------------
> Alex P. Madarasz, Jr. - Hughes Training, Inc. - al...@eagle.bgm.link.com
But where are the third-party applications for the Ferrari?
The problem lies in the fact that computer systems are purchased on the
basis of how well they solve problems. Those problems are solved by
application software. That giant sucking sound you hear is coming from
Redmond WA, and it's the sound of hundreds of software companies being
drawn into developing applications for NT (non-gui) and WIN32 (gui)
operating systems because of the size of the market. These companies
have a finite amount of development $ to spend, therefore we are seeing
less money spent on SGI versions of the apps. Some companies that used
to develop only for SGI are now working on WIN/NT versions. My guess is
that emerging companies will bypass SGI (unless their application is
specifically aimed at our hardware's capabilities) and go directly into
the light of NT.
--
David C. Frederick User Environment Product Manager
DMS Software Marketing Silicon Graphics Inc.
Phone: (415) 933-7571 Pager E-mail: dc...@pager.sgi.com
Pager: (800) 949-9708 dcf@.sgi.com
Richard Ottolini
In article <32241F...@eagle.bgm.link.com>,
Alex P. Madarasz, Jr. <al...@eagle.bgm.link.com> wrote:
>there _any_ evidence that SGI has forgotten how to innovate to keep well
>ahead of "PC" technology.
It is economics of scale. The Wintel market manufactures and ships
tens of millions units a year. SGI-MIPS is second in market share
having done a respectable job of diversifying, but an order of magnitude
behind Wintel. The other RISC manufacturers come close to neither.
Who is going to have the largest R&D budget then based on units shipped?
I totally agree that non-Wintel hardware and software is more interesting,
but at some point price is a factor.
Noam Ben-Ami
Actually, SGI >is< being forced to respond to the PC market. They
are lowering their prices on their machines, they came out with
the (pathetic) Indy's to try and compete at a lower level, they
bought Alias in order to keep it away from NT...they are coming
out with a set of completely new machines...
Lets face it, PCs are fast closing the gap between themselves
and SGIs, and a year from now, that gap will be even smaller.
--
Noam Ben-Ami Rare Medium, Inc. Multimedia...Well Done
"To Rare Medium's relief, I speak only for myself. Were I speaking for Rare
Medium my opinions would be better informed and much more entertaining."
Alex P. Madarasz, Jr.
Ben Cannon wrote:
> >At the low-end, that means adopting PC technology, like the PCI bus
> >and peripherals, and getting the entry-level price, including software,
> >down.
>
> almost no cost savings, you would need to pay a huge amount of money for software drivers for
> (as an example) a 3Com Network card.
Hmmmm, silly to adopt the PCI bus, huh?
Gee, I guess we'll have to wait and see!
;-)
> Why chop the Indy's bus in half? it's already twice as fast PCI! I say we give the Indy 4
> processors and the HIO bus form the Onyx! Of course, we can;t just slap the same chipset
> in, it will need to be implimented in either a one or two chip set to be affordable and fast.
Well, a 64-bit PCI bus does exists (folks like DEC and Intergraph even
know something about it!), and the potential bandwidth is even in the
same
ballpark as the Indy / Indigo2 GIO buses.
Doesn't mean that PCI is the appropriate bus for SGI graphics tho.
> >Now, the ball is in SGI's court regarding whether it can respond,
> >cut costs on the low end and raise performance on the high end
> >to compete with the new generation of Pentium Pro systems with
> >3D graphics cards.
It's just plain amazing to me how many people in these threads think
that
the PC 3-D market is "pushing" SGI, and that SGI is somehow "forced" to
respond to the "PC" market in order to maintain their business.
Pretty laughable, actually.
jbta...@kuhub.cc.ukans.edu
Alex P. Madarasz, Jr. wrote:
>
> > Alex P. Madarasz, Jr. <al...@eagle.bgm.link.com> wrote:
> > >there _any_ evidence that SGI has forgotten how to innovate to keep well
> > >ahead of "PC" technology.
> >
> > It is economics of scale. The Wintel market manufactures and ships
> > tens of millions units a year.
>
> any way as a cooperative entity that can truly "share" R&D costs across
> the market more efficiently than one innovative company can.
>
> I haven't seen any evidence that the "PC 3-D market" exists in any such
> shape or form - witness the spate of "consumer 3-D accelerators" which
> ship with games customized to their own chipsets / APIs (and the
> constant
> bickering over the "best 3-D API").
>
> The fact is that the "PC market" is a loose conglomeration of
> mass-market
> manufacturers and resellers operating on low profit margins and high
> volume, where cheap is usually "best". I've seen no evidence that such
> a business model is any more cooperative and innovative than, say, the
> automobile parts market.
>
> -----------------------------------------------------------------------
> Alex P. Madarasz, Jr. - Hughes Training, Inc. - al...@eagle.bgm.link.com
As far as graphics APIs go, Open GL seems to be the choice for everyone
in the mid-range except for Max with Heidi. Some of the leading pc
graphics companies are working on additions to the Open GL spec,
including Intergraph, and I think 3D labs. Supporting two major APIs
should not be that big a problem for the wintel platform, their used to
writing Drivers. Also, there maybe a dirth of 3D boards out there, but
the only high end 3D accelerator chips for the PC that I have seen all
come from either 3d Labs (Glint and Delta combos) or Intergraph. By far
most of the high end boards use the Glint chips. BTW, Intergraph is
even starting to offer Glint based solutions on their low end models.
In reality, there is far less diversity in the NT 3D niche then the PC
world is used to.
Mark Taylor
CAD Systems Supervisor
Payless Shoe Source
Alex P. Madarasz, Jr.
Noam Ben-Ami wrote:
> Actually, SGI >is< being forced to respond to the PC market. They
> are lowering their prices on their machines, they came out with
> the (pathetic) Indy's to try and compete at a lower level, they
> bought Alias in order to keep it away from NT...they are coming
> out with a set of completely new machines...
Indy shipped before the current PC 3-D mania started.
And who says SGI is responding to pressure from the PC market - is
that your impression, your guess, your hope, or a quote from an
SGI executive and/or based on a published SGI stategic plan?
Are you sure SGI wasn't / isn't primarily competing with workstations
from HP, Sun et al rather than the PC market?
Are you sure that the applications now being promoted for the PC
platform represent such a significant majority of SGI's revenue stream
that they _have_ to be worried?
> Lets face it, PCs are fast closing the gap between themselves
> and SGIs, and a year from now, that gap will be even smaller.
Nah.
Alex P. Madarasz, Jr.
there _any_ evidence that SGI has forgotten how to innovate to keep well
ahead of "PC" technology.
-----------------------------------------------------------------------
Noam Ben-Ami
"Alex P. Madarasz, Jr." <al...@eagle.bgm.link.com> writes:
>Noam Ben-Ami wrote:
>> Actually, SGI >is< being forced to respond to the PC market. They
>> are lowering their prices on their machines, they came out with
>> the (pathetic) Indy's to try and compete at a lower level, they
>> bought Alias in order to keep it away from NT...they are coming
>> out with a set of completely new machines...
>Indy shipped before the current PC 3-D mania started.
Exactly. And it doesn't have any built in 3D graphics capabilities.
The original Indy was a crappy little machine. The new ones are
usable, but not great. (I coded lots of OpenGL stuff for them and
it was pretty painful.)
>And who says SGI is responding to pressure from the PC market - is
>that your impression, your guess, your hope, or a quote from an
>SGI executive and/or based on a published SGI stategic plan?
Its common sense.
>Are you sure SGI wasn't / isn't primarily competing with workstations
>from HP, Sun et al rather than the PC market?
Of course I am. Show me places that have a mix of SGIs and HPs and Suns,
and I'll show you an educational institution. Show me a place with a
mix of SGIs, PCs, and Macs, and I'll show you a design/creative company.
>Are you sure that the applications now being promoted for the PC
>platform represent such a significant majority of SGI's revenue stream
>that they _have_ to be worried?
Yes. Lets face it, After Effects may not be taking Flame seats, but it
does let people say "hey, I don't HAVE to have Flame to do this." In
a couple of years it will start taking Flame seats. However, for now,
no, PCs do not yet represent a majority.
>> Lets face it, PCs are fast closing the gap between themselves
>> and SGIs, and a year from now, that gap will be even smaller.
>Nah.
If you really think that, you're way out of touch.
Then again, the new SGIs come out in October, and if SGI doesn't
pull a Commodore, they might keep the gap from narrowing a lot.
The gap HAS been narrowing, of course, and anyone who denies that
is not up to date on the way of things. My Pentium Pro here feels
as fast and responsive as my SGI2Extreme, though not with interactive
graphics, of course. It does, however, render faster.
Robert A Schmitt
nbe...@uiuc.edu (Noam Ben-Ami) writes:
>Actually, SGI >is< being forced to respond to the PC market. They
>are lowering their prices on their machines, they came out with
>the (pathetic) Indy's to try and compete at a lower level, they
>bought Alias in order to keep it away from NT...they are coming
>out with a set of completely new machines...
>Lets face it, PCs are fast closing the gap between themselves
>and SGIs, and a year from now, that gap will be even smaller.
PCs == Crays? Think not dude. Product line turnover proceeds
at its usually hairy pace (still using your '486s to do your
rendering??) All you're saying to me is "I wish SGIs cost what
a PC costs, 'cause then I could buy an SGI!"
The Indy architecture is 4 years old. Show me a 4 yr old PC that
you can pull the CPU out of and stick in the latest model.
I still use my "pathetic" Indy 4 times more than my PPro
GLINTed NT box. The software is better, the interface is better,
the busses are better, etc. etc. etc.
"It's the WHOLE system, stupid!" You can quote my paraphrase.
or
"Oooooh, it's almost a UNIX box!"
--
Bob
------------------------------------------------------------------------
Robert A. Schmitt RESolution Graphics Inc.
Information Visualization Consulting r...@world.std.com
peter gruhn
> Indy shipped before the current PC 3-D mania started.
So, you're saying that SGI isn't visionary?
Alex P. Madarasz, Jr.
Richard Ottolini wrote:
> Alex P. Madarasz, Jr. <al...@eagle.bgm.link.com> wrote:
> >there _any_ evidence that SGI has forgotten how to innovate to keep well
> >ahead of "PC" technology.
>
> tens of millions units a year.
This means nothing unless you can show that the "Wintel market" acts in
any way as a cooperative entity that can truly "share" R&D costs across
the market more efficiently than one innovative company can.
I haven't seen any evidence that the "PC 3-D market" exists in any such
shape or form - witness the spate of "consumer 3-D accelerators" which
ship with games customized to their own chipsets / APIs (and the
constant
bickering over the "best 3-D API").
The fact is that the "PC market" is a loose conglomeration of
mass-market
manufacturers and resellers operating on low profit margins and high
volume, where cheap is usually "best". I've seen no evidence that such
a business model is any more cooperative and innovative than, say, the
automobile parts market.
-----------------------------------------------------------------------
Ben Cannon
nbe...@uiuc.edu (Noam Ben-Ami) wrote:
>
>Actually, SGI >is< being forced to respond to the PC market. They
>are lowering their prices on their machines, they came out with
>the (pathetic) Indy's to try and compete at a lower level, they
>bought Alias in order to keep it away from NT...they are coming
>out with a set of completely new machines...
I'm sure PCs are eating into SGI's market, idoits will buy anything nowadays...
But the Indy is anything but "pathetic"!
More like a mericle that GSI could take a $30,000 workstation (Indigo^2 XZ)
Add ISDN, Video digitizing and a digital video camera then cut the price in half!
--
Ben Cannon. art...@a.crl.com
"History is an Illusion caused by the passage of Time"
peter gruhn
> My Pentium Pro here feels
> as fast and responsive as my SGI2Extreme, though not with interactive
> graphics, of course. It does, however, render faster.
You seem level headed and rational, I'll ask you.
I'm not surprised about the interactive performance, but then, what kind
of acceleration do you have for that?
Interesting about the rendering faster. Can you give me a better idea of
what hardware (I can't keep track of what chips and memories are in
which SGI) and what kind of rendering (what software, levels of
rendering, you know the processor intensive stuff) you are doing?
TNX - gruhn
peter gruhn
> Who is going to have the largest R&D budget then based on units shipped?
Not easy for me to answer. How many companies are in Wintel R&D? How
many of them are cooperating? What kind of profit margins do the Wintels
have vs. SGI? What are the R&D departments targeting? How much of
Wintel's R&D budget is spent on perfecting the UI? SGI's?
etc.
Simon Powers
In article <502abh$8...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
Ben-Ami) wrote:
>"Alex P. Madarasz, Jr." <al...@eagle.bgm.link.com> writes:
>
>>> Lets face it, PCs are fast closing the gap between themselves
>>> and SGIs, and a year from now, that gap will be even smaller.
>
>>Nah.
>
>If you really think that, you're way out of touch.
>
>Then again, the new SGIs come out in October, and if SGI doesn't
>pull a Commodore, they might keep the gap from narrowing a lot.
>
>The gap HAS been narrowing, of course, and anyone who denies that
>is not up to date on the way of things.
Well I'm afraid I'd deny it. 2 years ago my desktop machine was an Indigo2
Extr. No PC at the time (P90 being pretty standard) could get close to it
for real time polygon manipulation.
Now my desktop is an Indigo^2 Max Impact, and surprisingly enough no PC
can get close to it again. We've got Glint based cards in Pentium Pro's
that are beginning to look closer to the SGI's (now old) XZ/Extreme
product line, but they're still not quite there yet.
If PC's hadn't finally started getting hardware 3D acceleration we'd have
been talking about the gap widening, rather than shrinking. As it is, in
my area (real time polygon manipulation) PC's continue to lag very
significantly and don't appear to be getting closer.
> My Pentium Pro here feels
> as fast and responsive as my SGI2Extreme, though not with interactive
> graphics, of course.
> It does, however, render faster.
For some people, interactive graphics are the whole reason for SGI's.
Rendering (as I believe your applying the term) is processor limited, and
I suspect a new R10K would give you nice zippy performance.
>
>--
>
>Noam Ben-Ami Rare Medium, Inc.
Simon Powers
--
G: Never mind all that, take a card. All opinions are my own
D: Card? What do I do with the card? and not BT's.
G: You can keep it I've got 51 left.
Duck Soup
Peter Shenkin
In article <aboucek-2908...@p44.ts1.walrus.com>,
Alan Boucek <abo...@walrus.com> wrote:
>In article <502abh$8...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
>Ben-Ami) wrote:
> > ...Show me places that have a mix of SGIs and HPs and Suns,
> > and I'll show you an educational institution. Show me a place with a
> > mix of SGIs, PCs, and Macs, and I'll show you a design/creative company.
No way. Computational chemists generally use SGIs and Macs. For
example, our group has about 10 SGIs and 3 Macs. We develop
computational chemistry software. 90% of our users run it on SGI.
When we visit industrial or academic sites where computational
chemistry is done (for example, molecular modeling groups in
drug or chemical companies or in academic departments), we see SGIs
and Macs.
The SGIs are used for scientific software; the Macs are used for
word-processing, spreadsheets, XY plotting, and making line-drawings
of chemical structures (using ChemDraw, from Cambridge Scientific,
which is almost single-handedly responsible for popularizing the Mac
among chemists). Example: we write our manuals on the Mac using
FrameMaker, and when we publish articles we use MS/Word to prepare
the manuscript.
The other 10% of our users run on IBM/AIX. A few folks out there
prefer PCs to Macs, and there is now beginning to be interest in
the computational chemistry community in using PentiumPro machines
for number-crunching.
But at this moment, SGI and Mac happen to be what computational
chemists favor, in academe and the industrial world.
By the way, computational chemistry alone is a larger market for
SGI than film/advertising/design, if I recall correctly.
-P.
--
********************* (Note new snail-mail address.) **********************
* Peter S. Shenkin, Chemistry, Columbia U., 3000 Broadway, Mail Code 3153,*
** NY, NY 10027; she...@columbia.edu; (212)854-5143; FAX: 678-9039 ***
MacroModel WWW page: http://www.cc.columbia.edu/cu/chemistry/mmod/mmod.html
Simon Powers
In article <504co7$k...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
Ben-Ami) wrote:
>Simon....@bt-sys.bt.co.uk (Simon Powers) writes:
>
>>In article <502abh$8...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
>>Ben-Ami) wrote:
>
>>>"Alex P. Madarasz, Jr." <al...@eagle.bgm.link.com> writes:
>>>
>
>>Well I'm afraid I'd deny it. 2 years ago my desktop machine was an Indigo2
>>Extr. No PC at the time (P90 being pretty standard) could get close to it
>>for real time polygon manipulation.
>>Now my desktop is an Indigo^2 Max Impact, and surprisingly enough no PC
>>can get close to it again. We've got Glint based cards in Pentium Pro's
>>that are beginning to look closer to the SGI's (now old) XZ/Extreme
>>product line, but they're still not quite there yet.
>>If PC's hadn't finally started getting hardware 3D acceleration we'd have
>>been talking about the gap widening, rather than shrinking. As it is, in
>>my area (real time polygon manipulation) PC's continue to lag very
>>significantly and don't appear to be getting closer.
>
>
>Hehehe, you're forgetting a crucial fact: the cost of GLint based
>pentium pro is about a tenth of the cost of an Indigo Max Impact...
>well, more like a fifth. And it's CPU is pretty competitive.
What's cost got to do with your argument? As far as I can see you were
stating PC's were closing the gap on SGI's. I state they clearly aren't,
since the SGI standard desktop machine keeps getting faster, and suddenly
your onto who gives per bangs per buck. This always has been the PC, and
always will be the PC, since it has the benefit of mass market (actually
my PSX probably gives more polys per sec per dollar, but thats a different
argument :) ). However, in terms of the gap in real time poly performance,
its still as wide as ever.
As for the CPU, the gap was never massive. In my case, a couple of years
ago it was a R4400 150Mhz against a 90MHz Pentium, now its a 250MHz R4400
against a 200MHz Pentium, and next time it'll be a R10K against a Pentium
Pro. SGI always has an edge, but never a particularly massive one.
>
>Now, how long do you figure till GLint cards that can compete with the
>Impact graphics come out? No too damn long...and even if they cost
>$5k-10k to start, that's still reasonable compared to the cost of the Impact
>graphics upgrade for something like a 2Extreme.
I've been hearing PCs will compete with SGI's for a few years now and it
still hasn't happened. Getting Impact performance levels is a lot more
work than simply bolting GLint chips onto a card. Of course in a few years
PC's may be as fast a max impact, but then I'll have an infinite reality
on my desk :).
p.s. I develop on unaccelerated PC's, accelerated PC's and SGI's, so I'd
hope to be fairly unbiased on this issue.
>Noam Ben-Ami Rare Medium, Inc. Multimedia...Well Done
Richard Ingram
In article <Dwv03...@world.std.com> r...@world.std.com (Robert A Schmitt)
writes:
nbe...@uiuc.edu (Noam Ben-Ami) writes:
>Lets face it, PCs are fast closing the gap between themselves
>and SGIs, and a year from now, that gap will be even smaller.
PCs == Crays?
Who was saying that ? Cray is no longer, all it will become is a rebadged SGI
box, how many years before the CRAY name is removed from SGI and becomes just
a distant memory ?
Now PC technology can be used to make Cray beaters, what about the boxes Intel
are making ? Let me reiterate no one is saying a PC can beat a CRAY, but they
can beat an SGI Indy, and how much does a new processor module cost for an
Indy ? The same as a PentiumPro box with 17" Multisync monitor and 32 MB RAM !!
Richard.
David H Dennis
My SGI Personal Iris 4D/30TG with 32MB RAM is quite a bit slower than
my Pentium 120 PC with 48MB RAM running Windows NT, and my Sun SPARC
5 system. And yet I still prefer using it because I love the SGI
OS and environment.
I could have gotten a much faster Unix box for my $ 2k, but I fell
in love with the SGI operating environment and have never regretted
my decision to get a SGI.
I wonder how many people here who talk about their PCs being faster
would /really/ rather have a PC when push came to shove, even if
it was a bit faster. Development of hardware and sofware are so
fast that there will always be a computer faster than the one you
own. But if it's not a SGI, well, it won't be a SGI.
I think a healthy percentage of the people here would really rather
have a SGI, but just can't afford one just yet. I've seen PCs and
I've seen SGIs, and I think SGIs are well worth it.
If I only had the money for a WebForce Indy ... but I guess I now
know I should wait until October, eh?
D
John Zulauf
Ben Cannon wrote:
>
> I'm sure PCs are eating into SGI's market, idoits will buy anything
> nowadays...
Actually SGI volumes are continuing to grow -- now if your claim is that
PC are eating into the market SHARE of the $20,000 and under
workstation, you might be right -- but the total market is growing
violently.
--
John Zulauf jzu...@aw.sgi.com Alias|Wavefront, Santa Barbara
speaking for himself
Richard Ingram
I'm sure PCs are eating into SGI's market, idoits will buy anything
nowadays...
But the Indy is anything but "pathetic"!
More like a mericle that GSI could take a $30,000 workstation (Indigo^2 XZ)
Add ISDN, Video digitizing and a digital video camera then cut the price
in half!
Hey its just a shame they can not manage to sell a 1Gig SCSI HD for less than
a PentiumPro PC with 32 Megs of RAM and 2 Gig Hard disc !!
Check out UK prices for Indy Hard Discs from SGI 1400UKP plus 17.5% VAT and
thats for 1Gig!!
Better than that why don't they put a 3Gig HD in the base level Indy ?? Of
course that will up the price by at least 3K !
Richard.
Richard Ottolini
For its first decade 1980-1990, SGI was primarily a scientific/engineering
graphics workstation company. It is only when computer animation started
its resurgence @1989 that SGI became a prominent entertainment graphics
workstation. Some of SGI's size offerings were strongly influenced
by its original market.
Noam Ben-Ami
Simon....@bt-sys.bt.co.uk (Simon Powers) writes:
>In article <502abh$8...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
>Ben-Ami) wrote:
>>"Alex P. Madarasz, Jr." <al...@eagle.bgm.link.com> writes:
>>
>Well I'm afraid I'd deny it. 2 years ago my desktop machine was an Indigo2
>Extr. No PC at the time (P90 being pretty standard) could get close to it
>for real time polygon manipulation.
>Now my desktop is an Indigo^2 Max Impact, and surprisingly enough no PC
>can get close to it again. We've got Glint based cards in Pentium Pro's
>that are beginning to look closer to the SGI's (now old) XZ/Extreme
>product line, but they're still not quite there yet.
>If PC's hadn't finally started getting hardware 3D acceleration we'd have
>been talking about the gap widening, rather than shrinking. As it is, in
>my area (real time polygon manipulation) PC's continue to lag very
>significantly and don't appear to be getting closer.
Hehehe, you're forgetting a crucial fact: the cost of GLint based
pentium pro is about a tenth of the cost of an Indigo Max Impact...
well, more like a fifth. And it's CPU is pretty competitive.
Now, how long do you figure till GLint cards that can compete with the
Impact graphics come out? No too damn long...and even if they cost
$5k-10k to start, that's still reasonable compared to the cost of the Impact
graphics upgrade for something like a 2Extreme.
--
Noam Ben-Ami Rare Medium, Inc. Multimedia...Well Done
Bryant Reif
Alan Boucek wrote:
> When you look at comparable complete systems, the pricing difference is
> nowhere near that big. High Impact is around $30-35K. Comparably
> configured Intergraph boxes that might have a chance of competing start at
> about half that. Few people are going to use mass market PPro boxes as
> workstations- many of them max out at 128mb RAM- these days, that's a base
> configuration for the things that most people do with SGIs.
That's yesterday's news. Today you can get a Goliath Quad PPro
motherboard that maxes out at 1GB of onboard RAM. In fact the quad
150MHz version (That's with FOUR 150MHz PPro CPUs!) sells for under
$2600. Zoinks!! How much was that Indigo again?!
> never forget, you get what you pay for.
If you say so. :)
> --
> Alan Boucek abo...@walrus.com New York, New York
--
Bryant Reif
mailto:reif...@pilot.msu.edu
http://www.aiesec.org/~bryant
Steph Greenberg
jbta...@kuhub.cc.ukans.edu wrote:
: writing Drivers. Also, there maybe a dirth of 3D boards out there, but
: the only high end 3D accelerator chips for the PC that I have seen all
: come from either 3d Labs (Glint and Delta combos) or Intergraph. By far
: most of the high end boards use the Glint chips. BTW, Intergraph is
: even starting to offer Glint based solutions on their low end models.
: In reality, there is far less diversity in the NT 3D niche then the PC
: world is used to.
Yes, but there are an interesting variety of things on the lower end. Certainly
not very high performance, but good if you need to do some minor work at home
instead of staying late. To wit:
#9 Imagine 128 (Imagine 2D/3D chip)
Matrox Millenium (their own chip)
3Dfx (Voodoo Grafix chip)
Diamond Edge (NVidia which appears to do direct draw of patches,
but which doesn't appear to support open GL)
NetPower (Glint Permedia/Delta)
There are quite a few chip makers out there doing 3D acceleration of one sort
or another, though most appear to lean on the CPU alot. I'd say that's a fair
amount of diversity actually.
: Mark Taylor
: CAD Systems Supervisor
: Payless Shoe Source
--
-------------------------------------------------------------------------------
Steph Greenberg, 3DCGIMD CGI Character Orthopedic Surgeon,
st...@primenet.com Chiropractor, and Podiatrist.
Copyright 1996. All Rights Reserved. Permission granted for non-commercial
electronic republication only, such as Usenet and Email, and
non-commercial educational purposes such as charge free WWW pages.
Express permission is required for any other use. When in doubt, ask.
-------------------------------------------------------------------------------
Michael Schulman
In fact, about 15 % of our revenues are Entertainment related. The largest
market by revenue is selling to Manufacturing companies ( cad,cam,cae,database
servers, etc. ).
In article <504rr2$l...@sol.ctr.columbia.edu>, she...@still3.chem.columbia.edu (Peter Shenkin) writes:
|>
|> In article <504f0h$8...@news.unocal.com>,
|> According to the SGI people I've been in contact with, entertainment
|> graphics, though it gets lots of publicity, is still a far smaller
|> part of SGI's market than is scientific computation.
|>
|> -P.
|>
|> --
|> ********************* (Note new snail-mail address.) **********************
|> * Peter S. Shenkin, Chemistry, Columbia U., 3000 Broadway, Mail Code 3153,*
|> ** NY, NY 10027; she...@columbia.edu; (212)854-5143; FAX: 678-9039 ***
|> MacroModel WWW page: http://www.cc.columbia.edu/cu/chemistry/mmod/mmod.html
--
Michael Schulman
Manager, Technical Marketing
Engineering Applications
Silicon Graphics Computer Systems
2011 N. Shoreline Blvd.
Mountain View, CA 94039
415-390-3308
Peter Shenkin
jos...@cimatron.co.il
Keith J. Cambra wrote:
>
> In article <4vihse$h...@fido.asd.sgi.com>, ak...@tuolumne.asd.sgi.com (Allen
> Akin) wrote:
>
> >It's a thought-provoking analogy. Recently I read that Saab,
> >Mercedes, Infiniti, Volvo, Lexus, and BMW each have less than 1% of
> >the world car market.* Even though these ``vendors'' have dramatically
> >smaller market shares than GM, Honda, etc., they manage to maintain
> >viable businesses and provide well-respected products. Perhaps the
> >computer industry will evolve into a similar situation.
> >
> >Allen
> >
> >* This statistic is attributed to the October 1995 issue of Automotive News.
>
> yeah, i've read similar info. in fact, Mac Evangelist, Guy Kawasaki, has
> made that
> point many times to divert attention away from Apples rapidly dropping
> marketshare.
>
> there are several differences btwn the automotive industry and computer
> industry.
Very good analysis snipped: see original.
One point NOT mentioned is that after purchase, cars only need gas and service, which
are both available as commodities from several sources.
Computers need software and peripherals after the purchase. Software is not a
standardized commodity, nor are peripherals. If Apple's market share drops, software
development for Macs will cease to be profitable. Ditto for boards, drives, etc. When
users see that they pay a premium for Mac software - and development doesn't keep pace
with other platforms - they LEAVE. People don't switch cars because of the spark plugs
or gas that is available for the car. They do in computers.
I see the Unix stations on the desks of the techies where I work - then go back to my
Pentium PC in the marketing group. I am, by training, an engineer. I understand the
technical advantages. But the OS battle is being decided by marketing moves. Microsoft
deep-sixed OS/2 not on technical merits, but by squeezing Windows onto as many computers
as they could, and providing relatively cheap development tools. At one point, IBM was
charging US$2,000 for the OS/2 developer's kit while Microsoft was giving the Windows
kit away for free.
It's a pity none of the keepers of various Unix flavors has been able to get it together
to do the same. And in the same light, what I've read here about the SGI product line -
non-standard buses, etc. - gives me great pause. I think the era of buying a
purpose-built computer (sort of how many graphic artists view their Mac) is past. Large
numbers of users are more sophisticated. They want to be able to expand and build on
their investment. SGI doesn't need to compete with NTs to realize that the market for
expensive, single-task systems is smaller than the market for the same power in a more
open arrangement.
My guess is that many people will endure the less elegant, less powerful solutions of
the PC world because they can leverage their investment with constantly improving
software and periphals. Both initial purchase and subsequent add-ons are at lower price
points than UNIX, Mac, and other options. There is no reason why UNIX could not do the
same thing at a higher tier of performance - perhaps at a slightly higher price than PCs
despite economies of scale, but more commodity-like than currently.
Richard Ottolini
In article <504rr2$l...@sol.ctr.columbia.edu>,
Peter Shenkin <she...@still3.chem.columbia.edu> wrote:
>In article <504f0h$8...@news.unocal.com>,
>Richard Ottolini <stg...@sugarland.unocal.COM> wrote:
>>For its first decade 1980-1990, SGI was primarily a scientific/engineering
>>graphics workstation company. It is only when computer animation started
>>its resurgence @1989 that SGI became a prominent entertainment graphics
>>workstation. Some of SGI's size offerings were strongly influenced
>>by its original market.
>
>According to the SGI people I've been in contact with, entertainment
>graphics, though it gets lots of publicity, is still a far smaller
>part of SGI's market than is scientific computation.
Trick answer: Does that include the video game market of which SGI/MIPS
is ramping up in?
Frederick Haab
You know, I'd love to see this debate end and so don't like
to butt in, but then you see something you just have to argue
about....
Alan Boucek wrote:
>
> In article <502abh$8...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
> Ben-Ami) wrote:
>
>> "Alex P. Madarasz, Jr." <al...@eagle.bgm.link.com> writes:
>>
>>>Noam Ben-Ami wrote:
>>>> Actually, SGI >is< being forced to respond to the PC market. They
>>>> are lowering their prices on their machines, they came out with
>>>> the (pathetic) Indy's to try and compete at a lower level, they
>>>> bought Alias in order to keep it away from NT...they are coming
>>>> out with a set of completely new machines...
>>
>>>Indy shipped before the current PC 3-D mania started.
>>
>> Exactly. And it doesn't have any built in 3D graphics capabilities.
>> The original Indy was a crappy little machine. The new ones are
>> usable, but not great. (I coded lots of OpenGL stuff for them and
>> it was pretty painful.)
>
> and they're nearing the end of their life cycle. Interestingly, 4-5
> year old purple Indigos with 4400 processor upgrades are still *very*
> useful machines. Indys will remain useful for quite a while. How
> useful are 3-5 y/o PCs?
I have to argue with this point...Indy's are great machines that
many people are finding new uses for everyday. The R5000 came
out earlier this year and is specifically being used in Indys,
the R10000 in Indigo^2's. Doesn't sound like a machine nearing
the end of it's life cycle just because Indigos didn't last
as long as they should have.
And I really need to argue about the Indy's usefulness. It has
*never* been a crappy little machine. Certainly when one is
sitting next to an Indigo^2 Extreme or better, but set it next
to a PC and it hardly seems crappy. And it also seems to me that
Indys have had XZ graphics available for *years*, if not from
the very beginning. Right now all but the very high end accelerators
for PC's don't even touch XZ graphics.
And one of the major points of having an Indy is for development,
and specifically for low end applications that you don't need
a $100,000 machine for (I've never seen even a PowerMac run
PhotoShop as quickly as my R5000 Indy). Development of applications,
for example, can easily be done on an Indy and then run beautifully
on an Onyx. Try to program for a low end PC graphics card and then
get it to run on a high end card, and take advantage of the better
features...there's just no comparison.
> The Indy will be replaced. The machine that replaces it will be
> impressive. It will still cost more than a PC.
And Onyx's will be replaced, and Indigo^2's will be replaced,
and PC's will be replaced as well...pretty soon it'll be the
Pentium Power Pro, and the Quadruple Pentium "Can't touch this
with a stick" Superduper Power Pro. So I'm not sure I understand
this point.
>>> And who says SGI is responding to pressure from the PC market - is
>>> that your impression, your guess, your hope, or a quote from an
>>> SGI executive and/or based on a published SGI stategic plan?
>>
>> Its common sense.
>
> Is it? More likely, the PC world sees the sexiness of some high
[...info about only a small part of SGI's being in entertainment...]
I'm having a difficult time determining whose side you are on!
My impression is that SGI is targeting certain people, and some of
these people might overlap with people targeted by PC companies
in some low end areas. There *is* competition with high end Alphas
and Pentium Pros running apps like SoftImage. They *are* responding,
look at the first few pages of the latest 3D Design magazine if you
don't think so. But I still don't see what point this proves
except that really high end PC's are capable of doing some things
SGI's were doing years ago.
[...lots of stuff about what computers are used for deleted...]
>>>> Lets face it, PCs are fast closing the gap between themselves
>>>> and SGIs, and a year from now, that gap will be even smaller.
>>
>>>Nah.
>>
>> If you really think that, you're way out of touch.
[...]
>> The gap HAS been narrowing, of course, and anyone who denies that
>> is not up to date on the way of things. My Pentium Pro here feels
>> as fast and responsive as my SGI2Extreme, though not with interactive
>> graphics, of course. It does, however, render faster.
>
> nah, the gap will probably grow a bit. You're comparing a contemporary
> Pentium Pro machine with an SGI that's 3-4 years old.
Again, lets compare apples and apples...in the high end the gap
will continue to grow, it hasn't been narrowing. And as Alan
wrote, the gap may narrow when you compare 3 to 4 year old low
end SGI technology and new PC technology, but when you compare even
the latest low end technology from SGI (if you can even call Impact
"low end") to the latest high end technology on PC's then I think
it'll be a pretty steady, if not widening, gap.
I use a PC at home, I'd *love* to write apps on it, it's not that
I don't *want* a lot of competition to bring prices down, and get
some nice 3D graphics on PC's. But I've always hated developing on
PC's, esepecially graphics. A few years ago the issue was whose video
card are you going to support, after several years and standards
emerging and layers to handle the hardware (something SGI has always
done), you can finally sit down and write a program that uses high
resolution graphics and will run on just about any PC. Now introduce
a set of 3D graphics accelerators...here comes those feelings again.
OK, so I hear the argument that these accelerators will be OpenGL
compatible, but at what level? Do the people porting OpenGL
have to be hardware aware of every card? Do I need company
X's OpenGL port to run on their card, and do I need to recompile
with another library to run on another card and take advantage of
that hardware? Or do we get stuck in a quagmire of O.S. level
interpretation where each company will have to supply drivers for
their hardware, like Windows does with standard graphics cards?
And then there's the issue of what O.S. to even run on!
Maybe that's why I'm at work writing this on an Indy. We have
some Macs and PC's around for email and word processing and stuff,
some of the graphics guys do some painting on the Macs. That's
about it.
--
-=-=- Frederick Haab -=- Software Developer -=- Turner Production -=-=-
-=-=- ha...@efx7.turner.com -=- 404.885.0317 -=- fax: 404.885.0757 -=-=-
Robert A Schmitt
r...@morticia.inmos.co.uk (Richard Ingram) writes:
>In article <Dwv03...@world.std.com> r...@world.std.com (Robert A Schmitt)
>writes:
> nbe...@uiuc.edu (Noam Ben-Ami) writes:
> >Lets face it, PCs are fast closing the gap between themselves
> >and SGIs, and a year from now, that gap will be even smaller.
> PCs == Crays?
>Who was saying that ? Cray is no longer, all it will become is a rebadged SGI
Why would SGI buy Cray?
1. Market share / customer base
2. FOR THE TECHNOLOGY
3. Dominance in the high-end supercomputing market
This generation of SGI will not see any boosts from Cray inspired
architectures, but add-ons and follow-ons will. And again the SGI
line will be head and shoulders above the PC market (as always).
Noam stated that PCs will soon equal SGIs. I personally doubt it
(unless SGI goes full force into consumer electronics and starts
fighting Sony, Samsung, et al. by building intelligent entertainment
devices that are Webbed and gets out of the workstation market,
but I digress).
>Now PC technology can be used to make Cray beaters, what about the boxes Intel
Then they are no longer PCs. I can say the same thing about workstation
technology.
>are making ? Let me reiterate no one is saying a PC can beat a CRAY, but they
>can beat an SGI Indy, and how much does a new processor module cost for an
Yesterday's technology verses today's. Can't wait to see SGIs today's
technology in Oct. And my Indy still offers a better system's level
'feel' than NT 3.51 (can't go to NT 4.0 'cause the performance sucks
compared to NT 3.51).
>Indy ? The same as a PentiumPro box with 17" Multisync monitor and 32
MB RAM !!
Join the SGI Developers Group (max $395 per year) and get 30%-50% off
hardware purchases. And a development environment (ProDev Workshop,
C, C++, Fortran compilers, RapidApp, Developer's Toolbox, DevSupport, ...)
I don't see Compaq/Gateway/etc. giving away dev tools (but hey, I get
Office9* outta the deal :). At those prices, PCs do not compare to
SGIs in any way, shape, or form.
Try haggling. SGI sales people will try to get you as much
discount as possible (they really do). Try that with a PC mail-order
shop.
Ben Cannon
r...@morticia.inmos.co.uk (Richard Ingram) wrote:
>In article <502iba$s...@nntp.crl.com> Ben Cannon <art...@a.crl.com> writes:
>
> I'm sure PCs are eating into SGI's market, idoits will buy anything
> nowadays...
>
> But the Indy is anything but "pathetic"!
>
> More like a mericle that GSI could take a $30,000 workstation (Indigo^2 XZ)
> Add ISDN, Video digitizing and a digital video camera then cut the price
> in half!
>
>Hey its just a shame they can not manage to sell a 1Gig SCSI HD for less than
>a PentiumPro PC with 32 Megs of RAM and 2 Gig Hard disc !!
I never siad they did everyhting right -:)
SGI needs to do some serious price revisions, or I'm gonna keep buying RAM and disks
elseware :)
Hey, at least the SIGs can take PC simms!! Glad I don't have a SUN right now.
Jeff S. Ford
{snip}
> most of the high end boards use the Glint chips. BTW, Intergraph is
> even starting to offer Glint based solutions on their low end models.
> In reality, there is far less diversity in the NT 3D niche then the PC
> world is used to.
Intergraph is *NOT* selling Glint based solutions on our low end
models. We *are* selling a scaled back version of our graphics
accelerators called Intense 3D. This board is being sold to OEMs
(Gateway and others) and as an add-in board. It is much faster and has
more features than the 500T Glint boards at a comparable price. There
may be some Intergraph resellers out there bundling Glint in our
systems, but I don't know why they would.
Ben Cannon
nbe...@uiuc.edu (Noam Ben-Ami) wrote:
>Hehehe, you're forgetting a crucial fact: the cost of GLint based
>pentium pro is about a tenth of the cost of an Indigo Max Impact...
>well, more like a fifth. And it's CPU is pretty competitive.
>Now, how long do you figure till GLint cards that can compete with the
>Impact graphics come out?
Probally twice as long as it takes SGI to put Infinite Reality on the desktop
for $25,000.
:)
Ben Cannon
Accually, I think he was saying that PC makers weren't the visonary ones :)
SGI started the 3-D craze....
David Andriessen
In article <4vimfm$s...@nntp.crl.com>, Ben Cannon <art...@a.crl.com> wrote:
> Well, a Ford Stock car runs about a Half a Million dollars (if I wanted
to buy
> one right now) for a new one.
>
> Let's not pit apples against oranges and compare the $500K Ferrari F50.
> Equivlent prices, yes?
>
> Although I haven't done this myself, the straight line performance of the
> Stock car may be slightly faster than the Ferrari (I don't keep up with stock
> cars that much, the Ferrari can go 203mph..)
>
> However, throw a few turns into the mix, and the Ferrari will eay your
presious
> Ford for lunch by a huge margian!
>
> And I can drive to the Quickie Mart to get grocories in my Ferrari :)
>
> --
> Ben Cannon. art...@a.crl.com
>
Lets not forget the babes we can pickup along the way in a Ferrari:)
--david
Alex P. Madarasz, Jr.
Ben Cannon wrote:
> peter gruhn <gr...@apricot.com> wrote:
> >> Indy shipped before the current PC 3-D mania started.
> >
> >So, you're saying that SGI isn't visionary?
>
> Accually, I think he was saying that PC makers weren't the visonary ones :)
> SGI started the 3-D craze....
Actually, the way this sub-thread started is:
Noam Ben-Ami wrote:
> Actually, SGI >is< being forced to respond to the PC market. They
> are lowering their prices on their machines, they came out with
> the (pathetic) Indy's to try and compete at a lower level, they
> bought Alias in order to keep it away from NT...they are coming
> out with a set of completely new machines...
which says that SGI developed the Indy to respond to the advances in the
PC market.
The point I was making is that SGI shipped the Indy (in August 93)
before, for example, 3Dlabs was even formed (April 1994) - i.e.,
stating that the development of Indy (and SGI's _years_-old policy of
regularly lowering the price of older technology as they roll out new)
is in response to the PC market is little more than a trip through your
own _virtual_ reality.
People keep trying to convince themselves that SGI _must_ operate
according to their personal perception of the way the PC marketplace
operates, completely apart from any evidence that SGI's business (and
technology) model has anything to do with the "PC market".
-----------------------------------------------------------------------
Alex P. Madarasz, Jr. - Hughes Training, Inc. - al...@eagle.bgm.link.com
Alex P. Madarasz, Jr.
Noam Ben-Ami wrote:
> >Are you sure that the applications now being promoted for the PC
> >platform represent such a significant majority of SGI's revenue stream
> >that they _have_ to be worried?
>
> Yes. Lets face it, After Effects may not be taking Flame seats, but it
> does let people say "hey, I don't HAVE to have Flame to do this." In
> a couple of years it will start taking Flame seats. However, for now,
> no, PCs do not yet represent a majority.
Whelp, I won't argue with "common sense" - take a look at:
http://www.sgi.com/Headlines/investor/marketsegments.new.html
and note that SGI says that "Entertainment" represents only 20% of their
market. Even if the entire entertainment world lost complete faith in
SGIs and rushed out to their local "PC" store to buy "PCs", SGI would
still have 80% of their markets available to continue their revenue
stream.
Then, browse around the rest of their financial data and pull out some
numbers that back up the assertions that SGI is wilting under the
pressure from the "PC market".
Ray Chen
In article <322533...@cimatron.co.il>, <jos...@cimatron.co.il> wrote:
>Keith J. Cambra wrote:
>It's a pity none of the keepers of various Unix flavors has been able
>to get it together to do the same. And in the same light, what I've
>read here about the SGI product line - non-standard buses, etc. -
>gives me great pause. I think the era of buying a purpose-built
>computer (sort of how many graphic artists view their Mac) is past.
>Large numbers of users are more sophisticated. They want to be able to
>expand and build on their investment. SGI doesn't need to compete with
>NTs to realize that the market for expensive, single-task systems is
>smaller than the market for the same power in a more open
>arrangement. My guess is that many people will endure the less
>elegant, less powerful solutions of the PC world because they can
>leverage their investment with constantly improving software and
>periphals. Both initial purchase and subsequent add-ons are at lower
>price points than UNIX, Mac, and other options. There is no reason why
>UNIX could not do the same thing at a higher tier of performance -
>perhaps at a slightly higher price than PCs despite economies of
>scale, but more commodity-like than currently.
Some people just don't get it. The reason why SGI can deliver
more powerful machines is precisely because the hardware is more
proprietary. For example, we couldn't do the graphics we do now
(or for that matter, the stuff we did 3 years ago) if we were
limited to using today's PCI-32 bus. Not enough bandwidth.
The graphics would chew up so much there wouldn't be enough
left to do other useful work.
You can't have it both ways. Either you run on commodity hardware
and are limited to the performance of that hardware -- and everyone
else can deliver roughly the same performance -- or you can try and
blow the doors off what's out there using proprietary hardware.
If you look at the PC graphics board companies, they're doing the
same thing SGI is. They use custom (proprietary) chips in an
effort to stay ahead of the competition (S3, ATI Mach X, etc.).
But they can run under a software standard (Windows) that masks
the hardware differences.
We do our own buses in addition to custom graphics hardware, and our
software standards are OpenGL and Unix (IRIX is X/Open-branded,
POSIX compliant, etc.).
Ray Chen
r...@sgi.com
Ray Chen
In article <3228F4...@eagle.bgm.link.com>,
Alex P. Madarasz, Jr. <al...@eagle.bgm.link.com> wrote:
>Actually, the way this sub-thread started is:
>
>Noam Ben-Ami wrote:
>> Actually, SGI >is< being forced to respond to the PC market. They
>> are lowering their prices on their machines, they came out with
>> the (pathetic) Indy's to try and compete at a lower level, they
>> bought Alias in order to keep it away from NT...they are coming
>> out with a set of completely new machines...
>
>which says that SGI developed the Indy to respond to the advances in the
>PC market.
>
>The point I was making is that SGI shipped the Indy (in August 93)
>before, for example, 3Dlabs was even formed (April 1994) - i.e.,
>stating that the development of Indy (and SGI's _years_-old policy of
>regularly lowering the price of older technology as they roll out new)
>is in response to the PC market is little more than a trip through your
>own _virtual_ reality.
>
>People keep trying to convince themselves that SGI _must_ operate
>according to their personal perception of the way the PC marketplace
>operates, completely apart from any evidence that SGI's business (and
>technology) model has anything to do with the "PC market".
Correct. SGI didn't bring out the Indy to compete against PCs.
SGI brought out the Indy in an effort to expand its traditional
strengths -- graphics intensive desktops -- down the food chain.
Mostly, that's technical computing and content creation.
Indys made it possible to develop SGI-based applications at
1/3 the price of the Indigo/Indigo2 line. And adding more
IRIX users made it easier to persuade 3rd-party application
companies to port their software to SGI hardware. That's
one reason why SGI tries so hard to maintain *binary*
compatibility from the desktop to the supercomputer.
Ray Chen
r...@sgi.com
jos...@cimatron.co.il
Ray Chen wrote:
>
> In article <322533...@cimatron.co.il>, <jos...@cimatron.co.il> wrote:
> >Keith J. Cambra wrote:
> >It's a pity none of the keepers of various Unix flavors has been able
> >to get it together to do the same. And in the same light, what I've
> >read here about the SGI product line - non-standard buses, etc. -
> >gives me great pause. I think the era of buying a purpose-built
> >computer (sort of how many graphic artists view their Mac) is past.
> >Large numbers of users are more sophisticated. They want to be able to
> >expand and build on their investment. SGI doesn't need to compete with
> >NTs to realize that the market for expensive, single-task systems is
> >smaller than the market for the same power in a more open
> >arrangement. My guess is that many people will endure the less
> >elegant, less powerful solutions of the PC world because they can
> >leverage their investment with constantly improving software and
> >periphals. Both initial purchase and subsequent add-ons are at lower
> >price points than UNIX, Mac, and other options. There is no reason why
> >UNIX could not do the same thing at a higher tier of performance -
> >perhaps at a slightly higher price than PCs despite economies of
> >scale, but more commodity-like than currently.
>
> Some people just don't get it.
I "got" it - and agree with your statements (which follow). But
this thread is about comparing PCs and SGI. And market approach
is one valid point of comparison.
The reason why SGI can deliver
> more powerful machines is precisely because the hardware is more
> proprietary. For example, we couldn't do the graphics we do now
> (or for that matter, the stuff we did 3 years ago) if we were
> limited to using today's PCI-32 bus. Not enough bandwidth.
> The graphics would chew up so much there wouldn't be enough
> left to do other useful work.
>
> You can't have it both ways. Either you run on commodity hardware
> and are limited to the performance of that hardware -- and everyone
> else can deliver roughly the same performance -- or you can try and
> blow the doors off what's out there using proprietary hardware.
And wind up releasing your proprietary advantage as a public standard
when it's already been outpaced (like Apple). Or get enough of a buzz
going so that sufficient people buy into your innovation to make it a de
facto standard - while subsequently cooking up the next round of
innovation, which becomes a must-have standard because people have
bought into it....like Microsoft/Intel. These are just two of the many
market strategies possible in today's computer world.
You're expounding a niche strategy. I agree with you that SGI is, and
will remain, an upper echelon product line. But there are other
approaches. And I pointed out that, for example, UNIX never coalesced
into a standard with the market force of Windows - due solely to
marketing decisions. We can only guess what profits have been forfeited
by the developers and managers of UNIX.
>
> If you look at the PC graphics board companies, they're doing the
> same thing SGI is. They use custom (proprietary) chips in an
> effort to stay ahead of the competition (S3, ATI Mach X, etc.).
> But they can run under a software standard (Windows) that masks
> the hardware differences.
>
> We do our own buses in addition to custom graphics hardware, and our
> software standards are OpenGL and Unix (IRIX is X/Open-branded,
> POSIX compliant, etc.).
Which is why many of the SGI supporters who have posted here say
something like, "I just LOVE using my SGI machine for chemical
modeling/animation/other specific application, and when I need to write
a technical article/run a simple paint program/prepare a budget/run
games for my kids I do THAT on my trusty (name of second machine)."
Again, this may be a conscious decision on the part of SGI, as valid as
any other market positioning. But it is perfectly a propos this thread
to point out that this means SGI machines, for all their power, will not
draw much 3rd party development beyond its niche markets, and SGI users
must (and do) look elsewhere for the broader usefulness that comes from
such development.
For example, I know many people who are running technical and
non-technical businesses and consultancies. Those with software
backgrounds drool at the thought of having a UNIX station running
their SOHO operations. But it won't happen - because all the peripherals
and collateral software aren't there. You can't run a business with a
CPU or an OS. The same holds for SGI.
Again, this may be their strategy. In the larger context, it means that
the technology innovators aren't leading the broader computing
community: the big decisions are being left to snazzy marketers with
patchy solutions while innovators like Apple, and now SGI, stay in niche
markets. With all the annual drops in computing costs, isn't it a pity
the best technology never becomes mainstream?
Joshua
david
I got one question, and one statement
Question: how much does Photoshop and other none 3d programs cost
for SGI?
Statement: it seems to me that the people with all the higher then
hell SGI equipment are fooling themselves into thinking that SGI is
the best. All I know is a friend just got a new Indy all decked out
side by side my cheap home built intel box ruined his day, he's
sending the shit back, to slow for the money. done deal.
Jean-Eric Hénault
> Something that is important to understand is that modelling the real world
> is a tough task, and there's plenty of space to improve the high end of
> graphics systems. And that's where the bleeding edge customers are as well.
> It's just fine that the 3D graphics industry is growing with the help of
> cheap PC solutions, but Hollywood and other companies dealing with projects
> that are barely possible need the help of a company that is also interested
> in solving the hard problems, not providing cheap 'me too' systems that
> handles cases that won't help such companies.
So, in that case, what does SGI think of the new Digital PowerStorm
graphics architecture for the Dec Alpha. Apparently, this new
architecture can outperform a High-Impact machine at a much lower price.
This is probably just the beginning, but companies like Digital also
have lots of ressources, and I wouldn't call these products "cheap"
hardware solutions. Also, smaller companies like Deskstation bring
original new solutions to new problems. Like the new Multiflex
architecture, with up to 4-500 Mhz Dec alpha processors on a single
motherboard. Talk about performance... It is obvious to everyone that
SGI is suffering from an identity crisis. Its existence itself is
threatened. The new ads "We'd hate to bust your bubble..." is the first
evidence of this.
To survive, I don't expect SGI to sink as low as to offer $1,000 clones
to the general public, but it will have to widen its technology
spectrum. Offering Windows NT or Dec Alpha processors are perhaps
solutions. Lowering prices may be an other one. SGI have had the
exclusivity of this high-end market for a long time now, but these days
are gone. Time to adapt and make changes.
Sincerelly,
Jean-Eric Hénault
Ben Cannon
r...@tilt.engr.sgi.com (Ray Chen) wrote:
>In article <322533...@cimatron.co.il>, <jos...@cimatron.co.il> wrote:
>Some people just don't get it. The reason why SGI can deliver
>more powerful machines is precisely because the hardware is more
>proprietary. For example, we couldn't do the graphics we do now
>(or for that matter, the stuff we did 3 years ago) if we were
>limited to using today's PCI-32 bus. Not enough bandwidth.
>The graphics would chew up so much there wouldn't be enough
>left to do other useful work.
And Indy cars (the kind with piston engines, not the SGI version :)
burn Methanol, not Gasolene.
I haven't tried it yet, but on any sort of road-like surface....... you get the idea ;)
--
Ben Cannon. art...@a.crl.com
"History is an Illusion caused by the passage of Time"
Ben Cannon
r...@tilt.engr.sgi.com (Ray Chen) wrote:
>Indys made it possible to develop SGI-based applications at
>1/3 the price of the Indigo/Indigo2 line.
Not only that, but you didn;t hack apart the Indigo^2 either!
80% of the performance at 40% of the price of an Extreme....
You guys rule!
>And adding more
>IRIX users made it easier to persuade 3rd-party application
>companies to port their software to SGI hardware. That's
>one reason why SGI tries so hard to maintain *binary*
>compatibility from the desktop to the supercomputer.
Keep it up, and I promise I'll put an Onyx in my garrage someday :)
George White
Alex P. Madarasz, Jr. (al...@eagle.bgm.link.com) wrote:
: Folks who say SGI will roll over and die because "PCs" finally have
: reasonably fast 3-D graphics are just plain missing the point - SGI
: makes
: products which provide the performance at the price points their
: customers want, and their business model allows them to keep doing this
: while growing their business and making a tidy profit too. SGI sells to
: one or two [snicker] other markets besides the "I want lots of static
: rendering power for $9.95" one, and there's _no_ evidence that these
: markets are drying up at the rate suggested in these threads, nor is
: there _any_ evidence that SGI has forgotten how to innovate to keep well
: ahead of "PC" technology.
It ain't necessarily so. A large share of SGI's market now has options
that were not available before. The total market is increasing, but
it isn't clear that SGI's share will continue to grow as other vendors
cut into the low end. At one time, vendors saw the high-end hardware
as a sort of advertisement for the low-end of their range, which was
where they made money. I'm sure SGI benefits from this effect, which
means they don't really need to recover the full cost of the advanced
hardware. This could change if the low-end market becomes more
competitive.
--
George White <aa...@chebucto.ns.ca> <gwh...@bionet.bio.dfo.ca>
Andy Rawling
Richard Ingram wrote:
>
> Check out UK prices for Indy Hard Discs from SGI 1400UKP plus 17.5% VAT and
> thats for 1Gig!!
>
Hmmm. Now, I've never gone to SGI for a hardware quote but I'll give you
the benefit of the doubt & assume that's what you're talking about.
For any other UK dealer you're out by a factor of 3 to 4.
Last quote I had was around 1200 for 4GB.
You really should shop around.
> Better than that why don't they put a 3Gig HD in the base level Indy ?? Of
> course that will up the price by at least 3K !
>
> Richard.
Robert A Schmitt
> It's a pity none of the keepers of various Unix flavors has been able to get it together
> to do the same. And in the same light, what I've read here about the SGI product line -
> non-standard buses, etc. - gives me great pause. I think the era of buying a
> purpose-built computer (sort of how many graphic artists view their Mac) is past. Large
Totally wrong. Specialization of computers will always exist just as
specialization of any product exists. People buy automobiles based on
their needs: some buy sedans, some buy minivans, some buy sportscars.
They are all of the same class, automobile, but they are specialized
towards their most important function. This holds for any item I can
think of purchasing.
Would you NOT buy a car because it's powertrain was more powerful
than other cars of a similar type? Your company bought UNIX workstations
for your engineers while buying PCs for marketing. The right tools
for the job needed to be accomplished. I can stuff 7 people into
a sedan, but a minivan would do the job better at a higher price
point. But if I'm constantly stuffing 7 people into a vehicle it
makes sense to go with a minivan.
> numbers of users are more sophisticated. They want to be able to expand and build on
> their investment. SGI doesn't need to compete with NTs to realize that the market for
First and foremost, companies want to execute on their business plans.
If they fail that, forget about expansion.
> expensive, single-task systems is smaller than the market for the same power in a more
> open arrangement.
An SGI is not a single-task system. I've used them in the financial
services sector and the CAD sector, but to get the same results:
high-speed interactive 3D graphics which is the specialization of
an SGI computer.
Your definition of 'open' is strange to me. Software and hardware
architectures controlled by 2 companies is anything but open. If my
needs differ from the business models used by those 2 companies, I'm
SOL in your scheme of computing.
The markets maybe smaller for an SGI, but those markets have shown
that they will pay the higher premium. Also, PCs are not of the same
'power' that an SGI is.
> My guess is that many people will endure the less elegant, less powerful solutions of
> the PC world because they can leverage their investment with constantly improving
When IBM ruled the computing world and did not solve the needs of
different market segments, people did not endure. They innovated and
created wholly new solutions to solve their business problems. Why
would people give up that most successful strategy?
> software and periphals. Both initial purchase and subsequent add-ons are at lower price
> points than UNIX, Mac, and other options. There is no reason why UNIX could not do the
> same thing at a higher tier of performance - perhaps at a slightly higher price than PCs
> despite economies of scale, but more commodity-like than currently.
It's already done. SCSI devices, PostScript printers, Ethernet cards
all work on most UNIX platforms. It's the business model of UNIX
companies that is different from PC companies. Lower volume/higher
margins vs higher volume/lower margins.
The big question is: "Can this type of computer perform the job
I need to accomplish better and faster than other types of computers?"
jos...@cimatron.co.il
Robert A Schmitt wrote:
(heavily snipped)
> > I think the era of buying a
> > purpose-built computer (sort of how many graphic artists view their Mac) is past.
>
> Totally wrong. Specialization of computers will always exist just as
> specialization of any product exists. People buy automobiles based on
> their needs: some buy sedans, some buy minivans, some buy sportscars.
> They are all of the same class, automobile, but they are specialized
> towards their most important function. This holds for any item I can
> think of purchasing.
>
> Would you NOT buy a car because it's powertrain was more powerful
> than other cars of a similar type?
I would NOT buy a car that didn't use the same gas sold everywhere, and
that only had limited access to service and spare parts. Would you?
THAT is the appropriate automotive analogy to proprietary technology in
computing. Software (the magic gas which makes my computer a station
wagon when I need it, then a sportscar when I need THAT) is not readily
available. Neither are peripherals ( = spare parts).
This the Apple/Amiga syndrome: a marketing decision not to open the
standard chokes market share. To this day, Apple peripherals and software
have a price premium. Despite inventing the PC, they are niche players.
They have lost the battle for the home computer and business market -
although they possessed technology, mass production capability, and
look-and-feel that should have made them a shoo-in. Why? Because people,
given a choice, don't want to be beholden to privately held technology.
> The big question is: "Can this type of computer perform the job
> I need to accomplish better and faster than other types of computers?"
I agree that NT machines do NOT match the power of SGI machines. But
computing power steadily increases: more and more tasks fall within the
commodity computer's capabilities, fewer and fewer require specialized
product. Not coincidentally, software becomes more and more important to
most users as hardware power becomes a given. Ever heard the phrase
"killer app"?
> Bob
Joshua
Ian CR Mapleson
In article <322933...@odyssee.net> Jean-Eric Hénault <vide...@odyssee.net> writes:
>To survive, I don't expect SGI to sink as low as to offer $1,000 clones
>to the general public, but it will have to widen its technology
>spectrum. Offering Windows NT or Dec Alpha processors are perhaps
>solutions. Lowering prices may be an other one. SGI have had the
>exclusivity of this high-end market for a long time now, but these days
>are gone. Time to adapt and make changes.
There is another possibility that you haven't thought of, but from the comments
above I didn't think you would.
As Fistandantilus once said, I'll leave that to future history...
Ian.
The Doom Help Service (DHS): http://doomgate.gamers.org/dhs/
SGI Network Admin, University of Central Lancashire, Preston, England, PR1 2HE.
Doom: mapl...@gamers.org | Tel: (+44) 01772 893297, Fax: (+44) 01772 892913
Misc: mapl...@cee.hw.ac.uk | Office machine: SGI Indy R4400-200Mhz, 1MB cache
64MB RAM, 2GB SCSI disk, 24bit 17" 1280x1024, CDROM+DAT, IndyCam Digital Camera.
Richard Ottolini
>In article <322933...@odyssee.net> Jean-Eric Hénault <vide...@odyssee.net> writes:
>>To survive, I don't expect SGI to sink as low as to offer $1,000 clones
>>to the general public, but it will have to widen its technology
>>spectrum. Offering Windows NT or Dec Alpha processors are perhaps
>>solutions. Lowering prices may be an other one. SGI have had the
>>exclusivity of this high-end market for a long time now, but these days
>>are gone. Time to adapt and make changes.
Sorry to break the news, but they already did, and for around $250.
It is one of the 3D video game players (Nintendo64?) coming to market soon.
SGI was hawking them at SIGGRAPH.
This product may sell more computers on a per-unit basis than the rest
of the product lines of the company, but I don't know what kind of
profitibility it will have.
Noam Ben-Ami
stg...@sugarland.unocal.COM (Richard Ottolini) writes:
>Sorry to break the news, but they already did, and for around $250.
>It is one of the 3D video game players (Nintendo64?) coming to market soon.
>SGI was hawking them at SIGGRAPH.
>This product may sell more computers on a per-unit basis than the rest
>of the product lines of the company, but I don't know what kind of
>profitibility it will have.
None, in and of itself. It is a common saying in the games hardware industry,
that you ship a $100 bill with every machine sold. The market is that
competitive.
--
www.raremedium.com
Noam Ben-Ami Rare Medium, Inc. Multimedia...Well Done
"To Rare Medium's relief, I speak only for myself. Were I speaking for Rare
Medium my opinions would be better informed and much more entertaining."
Ben Cannon
Richard Ottolini wrote:
>
> Sorry to break the news, but they already did, and for around $250.
> It is one of the 3D video game players (Nintendo64?) coming to market soon.
> SGI was hawking them at SIGGRAPH.
> This product may sell more computers on a per-unit basis than the rest
> of the product lines of the company, but I don't know what kind of
> profitibility it will have.
Nintendo sells those machines at a substantial LOSS to make the whole
thing look "cheap" to 12-year olds. They then charge $100 per game to
accually make money.
The U64 chipset is neat, but nothing really special in the grand scheme
of things. Sure, it's totaly amazing for $250 but a 5+ year old Indigo
(back when CGA was the standard on PCs) will have better 3D performance.
Best Regards!
Ben.
Ben Cannon
jos...@cimatron.co.il wrote:
> I would NOT buy a car that didn't use the same gas sold everywhere, and
> that only had limited access to service and spare parts. Would you?
Yourt mileage may vary, I have a car of my own (it's small, almost "go
kart" like
but it'll do 120+mph :) that runs on Ethonal (and NOT gasolene!) that I
welded together
myself (read, I'm IT for spare parts)
Your mileage may vary :)
> THAT is the appropriate automotive analogy to proprietary technology in
> computing. Software (the magic gas which makes my computer a station
> wagon when I need it, then a sportscar when I need THAT) is not readily
> available. Neither are peripherals ( = spare parts).
What peripherals? SGIs use PC hardware (disk drives, CD-roms, and RAM)
Which is more than I can say for SUN. Just because there aren't ads for
it
in the back of "PC Magasine" dosen't mean it dosen't exist :)
> This the Apple/Amiga syndrome: a marketing decision not to open the
> standard chokes market share.
Well, that clears things up!
SGI, by it's very nature, can't be a mass-market company. To do so
would mean
offering second rate hardware at cheap prices. SGI makes the best of
the best, and their prices are pretty reasonable compared with the likes
of SUN, HP and IBM.... And Intergraph....
Best Regards
Ben.
Bryant Reif
Ben Cannon wrote:
> SGI, by it's very nature, can't be a mass-market company. To do so
> would mean
> offering second rate hardware at cheap prices. SGI makes the best of
> the best, and their prices are pretty reasonable compared with the likes
> of SUN, HP and IBM.... And Intergraph....
I heard the EXACT same arguement back in 1992... except it was coming
from a Macintosh user. The NT train came right at Apple in the
following years and before they even knew what happened, they were
steamrolled. Now it's heading for SGI and it's funny to begin to see
history repeating itself. :)
> Best Regards
> Ben.
--
Bryant Reif
mailto:reif...@pilot.msu.edu
http://www.aiesec.org/~bryant
Jim Mc Morrow
key.You can only play Mario on Nintendo.The reason people bought
Nintendo machines was because of the SOFTWARE available for it.If Ultra
64 has a game so good you just can't live without it, then it will sell
millions of machines and Nintendo will dominate the known Universe
again.
The question with Silicon Graphics is....Is Alias good enough that you
just have to have it if you are doing High end animation?If the answer
is yes then they will keep on selling machines.SGI have to keep spending
money to maintain Alias as (I think anyway), the best 3D animation
software around,even if you have to pay through the nose for it.
Regards Jim
Noam Ben-Ami
Jim Mc Morrow <mo...@batelco.com.bh> writes:
>In a way the Nintendo market reflects the SGI market.Exclusivity is the
>key.You can only play Mario on Nintendo.The reason people bought
>Nintendo machines was because of the SOFTWARE available for it.If Ultra
>64 has a game so good you just can't live without it, then it will sell
>millions of machines and Nintendo will dominate the known Universe
>again.
This is partly true. Marketing has a lot to do with it as well...you're
right though, Mario drives nintendo sales, and Nintendo knows this. In
fact, this fact is central to their developer strategy for the Ultra
64. No more half-assed games houses. You have to be VERY wealthy to
become a Nintendo Ultra 64 developer.
>The question with Silicon Graphics is....Is Alias good enough that you
>just have to have it if you are doing High end animation?If the answer
>is yes then they will keep on selling machines.SGI have to keep spending
>money to maintain Alias as (I think anyway), the best 3D animation
>software around,even if you have to pay through the nose for it.
This is absolutely wrong. First, as has been quite rightly pointed out,
SGI does not live off the commercial animation market. Its simply the
reason people think SGIs are so sexy...other than the great box colors
(though I still think the Indy's little plastic logos are terribly cheesy,
and I still have a yellow Extreme button lying around. :). )
Furthermore, an SGI running LW/Softimage/Vertigo/Prisms, at the absolute
high end, is superior to pretty much anything else you can buy, period.
Yes, they're expensive, yes, at the low end they are no longer terribly
competitive, but if you have upwards of $30,000 to burn on a machine and
want to do heavy animation, 2D or 3D, SGI is the clear choice.
Lastly, I sort of resent the "paying through the nose" comment. Do you
have any clue as to how complex a package like Alias is? Any idea of
the incredible, exorbitant costs of developing such a package for as
small and as demanding a market as alias targets? (I'm talking about
Alias Studio and Alias PowerAnimator, specifically) Alias is very
reasonably priced these days, for what it is.
Just my neurotic devil's advocate 2 cents.
Soren
In article <322FC3...@pilot.msu.edu>, Bryant Reif
<reif...@pilot.msu.edu> wrote:
> Ben Cannon wrote:
>
> > SGI, by it's very nature, can't be a mass-market company. To do so
> > would mean
> > offering second rate hardware at cheap prices. SGI makes the best of
> > the best, and their prices are pretty reasonable compared with the likes
> > of SUN, HP and IBM.... And Intergraph....
>
> I heard the EXACT same arguement back in 1992... except it was coming
> from a Macintosh user. The NT train came right at Apple in the
> following years and before they even knew what happened, they were
> steamrolled. Now it's heading for SGI and it's funny to begin to see
> history repeating itself. :)
As an apple user, my opinion is of course biased, but I'll state it
anyway: For those who don't yet know it, The Mac is still THE graphics
design system, like it or not, it has the most high-end hardware and
software specific to the fields of 2-d, 3-d, and video. If you what a
system that will run windows NT may I suggest buying a PPCP early next
year, you know, the system design by Apple and associates that will run
both the MacOS and WinNT, OS/2, plus a couple other things like Solaris,
etc. then you truely would have a system that would allow you total
compatiability accross all markets. If a piece of software is not
available for your OS of choice, it most certainly should be available for
at least one of the available OS's. (If it isn't, buying a SUN, SGI, or
IBM wouldn't put you in any better position, in fact, the hardware would
cost more and in fact probably be less powerful.)
--
Internet: tmd...@ups.edu
Bryant Reif
Soren wrote:
> As an apple user, my opinion is of course biased, but I'll state it
Oh no... I didn't expect to start another OS war. ;)
Well, I'm a former Apple user. I used them for 9 years before switching
to the PC for it's price/performance and especially software base. My
friends were shocked, but now most of them use PCs too. :)
> anyway: For those who don't yet know it, The Mac is still THE graphics
> design system, like it or not, it has the most high-end hardware and
> software specific to the fields of 2-d, 3-d, and video. If you what a
Uh... I don't think so. 3D on the Mac is really pathetic. Except for
EIAS and Form^Z there is nothing (At least until LW is ported). I was
forced to use Strata Studio Pro and Infini-D on the Mac for years and I
will never forgive them for it. :p
> system that will run windows NT may I suggest buying a PPCP early next
> year, you know, the system design by Apple and associates that will run
> both the MacOS and WinNT, OS/2, plus a couple other things like Solaris,
> etc. then you truely would have a system that would allow you total
> compatiability accross all markets. If a piece of software is not
> available for your OS of choice, it most certainly should be available for
> at least one of the available OS's.
You forgot one important point. Who is going to re-compile all the NT
apps to run on the PPC? Or are you talking about emulation? If the
latter is the case the performance hit would make most NT apps
impractical running on a PPC.
> (If it isn't, buying a SUN, SGI, or IBM wouldn't put you in any better
> position, in fact, the hardware would cost more and in fact probably be less
> powerful.)
That's debatable. But what isn't is that the PC platform has the best
price/performance ratio of any platform. Period. This point isn't
debatable.
P.S. Do not mistake this for Mac bashing please. Anyone who knows me
knows I have a soft spot for Apple (I used them for 9 years! You just
can't toss that away like an old shoe ;)) But I'm also realistic. The
Apple has a lock on DTP, graphic design, and multimedia development, but
I believe this is primarily due to inertia. It's not because the Mac is
inherantly superior in these fields it's just that industries don't
switch platforms overnight. :)
> --
> Internet: tmd...@ups.edu
Noam Ben-Ami
tmd...@ups.edu (Soren) writes:
>As an apple user, my opinion is of course biased, but I'll state it
>anyway: For those who don't yet know it, The Mac is still THE graphics
For ONE reason: inertia. Even the die hard mac fanatics here are conceding
that when we upgrade machines next, it'll be to NT.
>design system, like it or not, it has the most high-end hardware and
>software specific to the fields of 2-d, 3-d, and video. If you what a
Bullshit. 2-d? All adobe products are either available for NT or will
be, shortly. 3-d software? The Mac has electric image and Strata and
LW. PC's have LW, Max, Softimage, and many others. The mac isn't even
close to being able to compete in tha arena. Video? Natch.
>system that will run windows NT may I suggest buying a PPCP early next
>year, you know, the system design by Apple and associates that will run
>both the MacOS and WinNT, OS/2, plus a couple other things like Solaris,
>etc. then you truely would have a system that would allow you total
>compatiability accross all markets. If a piece of software is not
Except that it'll probably cost too much and WinNT would need to have
software ported to it to run on the machine...sorry, I'll stick to
my Pentium Pro.
>available for your OS of choice, it most certainly should be available for
>at least one of the available OS's. (If it isn't, buying a SUN, SGI, or
>IBM wouldn't put you in any better position, in fact, the hardware would
>cost more and in fact probably be less powerful.)
Thats highly arguable.
Now, personally, I think the PowerPC chip is significantly better than
the Intel chips, but NT/PPro is a tried and true system with tremendous
industry support. Furthermore, we're talking about the MacOS and Mac
systems today, not next year...and not running yet another flavor of
NT.
Apple never had a clue as to how to write a great operating system. Its
too bad because on the surface, the Mac GUI is pretty nice.
Noam Ben-Ami
WOW! Hey Bryant, our two posts were COMPLETELY identical! What a cool coincidence.
Kind of figures, since our top ten list of musical bands show us both to be
brilliant people. :)
Paul Christenson [N3EOP]
jos...@cimatron.co.il wrote:
> > Would you NOT buy a car because it's powertrain was more powerful
> > than other cars of a similar type?
>
> I would NOT buy a car that didn't use the same gas sold everywhere, and
> that only had limited access to service and spare parts. Would you?
>
> THAT is the appropriate automotive analogy to proprietary technology in
> computing. Software (the magic gas which makes my computer a station
> wagon when I need it, then a sportscar when I need THAT) is not readily
> available. Neither are peripherals ( = spare parts).
Sorry, not a valid comparison. Gasoline is consumable; you have to
continue to buy it to use your car. (Most) software isn't like that;
you buy it once, and you can keep using it forever. (I know that some
software licensing requires periodic payments.) The gasoline metaphor
compares to electricity. If the SGI required 208V 3 phase, then the
comparison would be valid, but it doesn't.. it uses the same 110V that
the PC uses.
The class of vehicle IS a valid comparison. Take PC/pickup truck vs.
SGI/sports car. If the speed is important, the choice is obvious, as it
is if you need to pull a camper or haul kitchen appliances.
> To this day, Apple peripherals and software have a price premium.
> Despite inventing the PC, they are niche players.
Better check your history again. Apple wasn't even close to inventing
the "home computer". The PC, or "Personal Computer", was solely an
IBM invention (which is a trade mark, BTW), and came YEARS after the
Apple II. The Apple II was among the first computers to come
preassembled. Heathkit's H8 system predated the Apple II by a couple
of years. (I don't recall the first computer created for the home/
hobbyist market; I want to say the Altair, but I may be wrong.)
> I agree that NT machines do NOT match the power of SGI machines. But
> computing power steadily increases: more and more tasks fall within the
> commodity computer's capabilities, fewer and fewer require specialized
> product. Not coincidentally, software becomes more and more important to
> most users as hardware power becomes a given.
It's true that the low end is getting more powerful, and tasks that were
once relegated to the high end are becoming possible. However, keep in
mind that the high end is moving just as fast (if not more quickly);
things that were impossible several years ago are now commonplace. In
fact, it's the development of the high end that pushes the "old" high
end stuff into the low end.
Take the much-touted movie industry. A few years ago, "Jurassic Park"
set all kinds of standards for reality-based computer animation.
Already, that level is commonplace.. and "state of the art" is several
times as advanced as JP. Remember the hoopla over "Star Wars"? It
looks downright primitive now.. and was state of the art in 1977.
Back in school, we had a lab with Indys, and a lab of Pentum-133 NT
machines. Both labs had much of the same software. We were doing some
processing as a part of our coursework that was simply not feasable
on the NT machines, due to the runtimes involved were measured in hours
on the NT machines. (And just try to get the ME students to use the
Pentiums for CAD work.)
--
+---------------------------------------------------+
| Technical Support Engineer, Cyclades Corporation |
| 800/88-CYCLADES (882-9252) or (510)770-9727, x258 |
| Maker of High Performance Multiport Serial Cards |
+---------------------------------------------------+
Ben Cannon
Soren wrote:
> As an apple user, my opinion is of course biased, but I'll state it
> anyway: For those who don't yet know it, The Mac is still THE graphics
> design system, like it or not, it has the most high-end hardware and
> software specific to the fields of 2-d, 3-d, and video.
Um... Photoshop is faster on my Indy than most any mac (the new quad
processor clones may edge it out, I haven't hcecked)
It's silly to say Macs have even decent 3-d, and if you take a look
at a package called "FLAME" for the SGI...... Macs may have the best
video available at CompUSA, but that dosen't count.
> If you what a
> system that will run windows NT may I suggest buying a PPCP early next
> year, you know, the system design by Apple and associates that will run
> both the MacOS and WinNT, OS/2, plus a couple other things like Solaris,
> etc.
If I were to go for an NT box it would be a DEC. They've got the
processing power down. Kind of.
If I wanted to shoot myself in the foot, I'd ust a .357 Magnum.
Not sure which one I'd rather do :)
> If a piece of software is not
> available for your OS of choice, it most certainly should be available for
> at least one of the available OS's.
Accually, SGI's are usually bought to run software available only on the
SIG platform :)
>(If it isn't, buying a SUN, SGI, or IBM wouldn't put you in any better position,
>in fact, the hardware would cost more and in fact probably be less powerful.)
Er... I get the cost more part, but be less powerful?
Wow! I just simply can't think of an adequate responce here...
Best Regards!
Ben.
Ben Cannon
Noam Ben-Ami wrote:
> This is absolutely wrong. First, as has been quite rightly pointed out,
> SGI does not live off the commercial animation market. Its simply the
> reason people think SGIs are so sexy...other than the great box colors
> (though I still think the Indy's little plastic logos are terribly cheesy,
> and I still have a yellow Extreme button lying around. :). )
SGI does have the best box design, and the monitors are neat, too :)
> Furthermore, an SGI running LW/Softimage/Vertigo/Prisms, at the absolute
> high end, is superior to pretty much anything else you can buy, period.
> Yes, they're expensive, yes, at the low end they are no longer terribly
> competitive, but if you have upwards of $30,000 to burn on a machine and
> want to do heavy animation, 2D or 3D, SGI is the clear choice.
> Lastly, I sort of resent the "paying through the nose" comment. Do you
> have any clue as to how complex a package like Alias is? Any idea of
> the incredible, exorbitant costs of developing such a package for as
> small and as demanding a market as alias targets? (I'm talking about
> Alias Studio and Alias PowerAnimator, specifically) Alias is very
> reasonably priced these days, for what it is.
Couldn't have said it better myself.
I think I'll drop out of this thread on a high note, they seem to few
and far
between during platform wars :)
Best!
Ben.
Joe Monahan
In article <50qco9$k...@vixen.cso.uiuc.edu>, nbe...@uiuc.edu (Noam
Ben-Ami) wrote:
> tmd...@ups.edu (Soren) writes:
> >As an apple user, my opinion is of course biased, but I'll state it
> >anyway: For those who don't yet know it, The Mac is still THE graphics
> that when we upgrade machines next, it'll be to NT.
Nonsense - I had a student call the top 25 graphic design/multimedia
studioa in the Phoenix metro area and all but 2 used Macs exclusively.
Only one was contemplating switching to PC's and most stated that the new
Mac/PC computers offered by Mac would meet their needs. Its only
non-graphics folk who dismiss the Mac, cause they haven't a clue.
> Bullshit. 2-d? All adobe products are either available for NT or will
> be, shortly. 3-d software? The Mac has electric image and Strata and
> LW. PC's have LW, Max, Softimage, and many others. The mac isn't even
> close to being able to compete in tha arena. Video? Natch.
You obviously have never done any multimedia production. You can talk all
you want about what apps "will" be available for the PC, but as one who
has spent three years doing multimedia, let me state clearly - don't get
lost in the silly assertions of computer programmers - the Mac is still
the best way to go. Its ease of use and built in capabilities with video,
etc make it an easy choice for all except the most nurdy.
> Except that it'll probably cost too much and WinNT would need to have
> software ported to it to run on the machine...sorry, I'll stick to
> my Pentium Pro.
And just keep buying those upgrade cards, and SCSI cards, and sound boards and
scanner cards and printer updates and calling tech support to see how the
hell its all supposed to fit together. Its fine if your a programmer type,
but if you want to produce - get a Mac.
> Apple never had a clue as to how to write a great operating system. Its
> too bad because on the surface, the Mac GUI is pretty nice.
Which would make one wonder why Microsoft has spent a decade trying to
immatate Macs. Ten years ago I could allocate memory to specific
applications, view a monitor whith thousands of colors, force quit an
application instead of rebooting, plug and play any number of peripherals,
listen to a music CD or work with digitized video. Microsoft is finally
catching up...to a decade old Mac.
JMonahan
Reto Koradi
Joe Monahan wrote:
> Which would make one wonder why Microsoft has spent a decade trying to
> immatate Macs. Ten years ago I could allocate memory to specific
Believe it or not, there are operating systems where every application
just allocates as much memory as it needs, without the user having
to do anything about it. And they have been existing for 25 years.
Reto (who loves OS wars and has been hating Macs for 10 years...)
--
Reto Koradi (k...@mol.biol.ethz.ch, http://www.mol.biol.ethz.ch/~kor)
Matt Jurcich
Bryant Reif wrote:
>
> Ben Cannon wrote:
>
> > SGI, by it's very nature, can't be a mass-market company. To do so
> > would mean
> > offering second rate hardware at cheap prices. SGI makes the best of
> > the best, and their prices are pretty reasonable compared with the likes
> > of SUN, HP and IBM.... And Intergraph....
>
> I heard the EXACT same arguement back in 1992... except it was coming
> from a Macintosh user. The NT train came right at Apple in the
> following years and before they even knew what happened, they were
> steamrolled. Now it's heading for SGI and it's funny to begin to see
> history repeating itself. :)
So you're saying you want SGI to become a mass-market company? I agree,
it would be cool to see an Indy on every desk, but they would have to do
quite a bit of redesigning and restructuring of their company to facilitate
it--not to mention (if you're going to do it right) release hardware info
to 3rd party vendors and get MS to port Office to Irix and all that
junk. They seem to be doing well catering to certain markets, and as
far as pricing and availability goes, I think it's not too bad for what
you're getting.
--
// matt jurcich // transwarp // mailto:inv...@goldengate.net //
\\ Silicon Graphics \\ Compaq \\ Novell \\ Netscape \\ NEC \\
// http://www.goldengate.net/~invisix //
Redistribution of this message via the Microsoft Network is prohibited
Matt Jurcich
Had to jump in here.
> > Except that it'll probably cost too much and WinNT would need to have
> > software ported to it to run on the machine...sorry, I'll stick to
> > my Pentium Pro.
>
> And just keep buying those upgrade cards, and SCSI cards, and sound boards and
> scanner cards and printer updates and calling tech support to see how the
> hell its all supposed to fit together. Its fine if your a programmer type,
> but if you want to produce - get a Mac.
Or SGI.
> > Apple never had a clue as to how to write a great operating system. Its
> > too bad because on the surface, the Mac GUI is pretty nice.
>
> Which would make one wonder why Microsoft has spent a decade trying to
> immatate Macs. Ten years ago I could allocate memory to specific
> applications, view a monitor whith thousands of colors, force quit an
> application instead of rebooting, plug and play any number of peripherals,
> listen to a music CD or work with digitized video. Microsoft is finally
> catching up...to a decade old Mac.
You imply Microsoft knows what they're doing. Enough said.
Marc Wilhite
> For ONE reason: inertia. Even the die hard mac fanatics here are conceding
> that when we upgrade machines next, it'll be to NT.
>
I'd consider myself a mac fanatic and when I upgrade I certainly will not
waste my time with NT, but will be upgrading to SGI.
Regards,
Marc
Noam Ben-Ami
peter gruhn
> > I heard the EXACT same arguement back in 1992... except it was coming
> > from a Macintosh user.
>
No, he's saying he thinks that SGI offers nothing more than attractive
advertising.
peter gruhn
> > >As an apple user, my opinion is of course biased, but I'll state it
> > >anyway: For those who don't yet know it, The Mac is still THE graphics
> > For ONE reason: inertia. Even the die hard mac fanatics here are conceding
> > that when we upgrade machines next, it'll be to NT.
> Nonsense - I had a student call the top 25 graphic design/multimedia
When you say "Nonsense", do you mean:
a) That an Apple user's opinion could not possibly be biased
b) That he'll state his opinion anyway
c) That the Mac is the THE graphics [...]
d) That the dis hard Mac fanatics there are conceding [etc.]
> studioa in the Phoenix metro area and all but 2 used Macs exclusively.
Or was it maybe:
e) An example of nonsense regarding the Phoenix metro area follows.
Well, which one? Come on, out with it. The suspense is driving me to
distraction.
> you want about what apps "will" be available for the PC, but as one who
> has spent three years doing multimedia, let me state clearly - don't get
Sure, I'll let you state clearly. Especially if you state it more
clearly than you stated that the person to whom you were replying has
spent three years doing multimedia.
> etc make it an easy choice for all except the most nurdy.
The most nerdy are buying BeBoxes.
> And just keep buying those upgrade cards, and SCSI cards, and sound boards and
> scanner cards and printer updates and calling tech support to see how the
Are you saying that if I buy a Mac I'll be stuck with what I buy today
and will not be able to buy cool new stuff when it comes out?
> Its fine if your a programmer type,
Your last girlfriend was a programmer, right?
[Lots of stupid stuff about what somebody's brother told him Windows was
like and stuff that Apple told him the Mac OS was like deleted because
we've all read it a thousand times before and in a day or two we'll all
get the chance to read the lies that MS and somebody else's brother told
somebody even more else about the Mac OS and Windows.]
Look, if you are going to try to open the lame old war, use real ammo,
don't shoot blanks that sound like whoopee cusions. Or better yet, try
to find the truth so we can build a better tomorrow rather than fight
about why yesterday sucked.
Steph Greenberg
Joe Monahan <jmon...@netzone.com> wrote:
: Which would make one wonder why Microsoft has spent a decade trying to
: immatate Macs. Ten years ago I could allocate memory to specific
: applications, view a monitor whith thousands of colors, force quit an
: application instead of rebooting, plug and play any number of peripherals,
: listen to a music CD or work with digitized video. Microsoft is finally
: catching up...to a decade old Mac.
Not if you were using a Mac. Macs didn't have color until around 1988. Or sound.
Now if you are talking about using an Amiga, which had a multitasking OS,
graphic *and* command line interface, thousands of colors, sound and
NTSC and Pal syncable video, well they *did* have that 10 years ago. Oh yeah,
they also had plug and play.
IMHO, Macs and PCs have spent the last 5 years or so imitating Amigas, and only
in the last couple have they, through increased power in the hardware,
succeeded to any extent. The Amiga failed because Commodore was an incredibly
stupid company who wouldn't know a clue if it fell on them.
Come to think of it, Ataris and even the lowly Apple IIgs had color before the
Mac. And when the Mac did first get color, it was the same 256 colors that
the PC of the same era had. And it had the same 8 bit sound as the Amiga
and Atari. The Amiga could display 4096 colors at the time, though with flaws.
In any case, only the original Mac was innovative, its successors came in
about a year behind current tech until about 1991. And the Mac OS itself
was copied off of X-windows developed at Xerox Park a decade earlier. As
Apple's suit against Microsoft revealed.
Nonetheless, I'd rather use an NT machine than a Mac. Also, I don't know if your
student bothered to ask which model of Macs designers were using in the Phoenix
area, but I'd bet the majority are still using Quadras.
Microsoft is busily trying to immitate SGIs now. And if you've seen
digital editing systems like Razor Pro using a modest PVR card on NT, or any
of the Matrox digital studio cards, you wouldn't be bragging about the Macs
digital video. Only the rather expensive Avid and Imix boxes are really pro
in their video output.
If you are into 3D, an area Macs have always been weak at, PC/NT machines
are a much better bet with a wider variety of good software than Macs. Let's
see, Softimage, Lightwave, 3DS Max, Truespace (Animation Master/MH3D is available
on both PCs and Macs). Tons of 3D hardware cards from around $299 up.
I think graphic designers using Macs are a bit like those die hards who were
still touting Amigas a year ago. For 2D work, Macs are still fine machines, but
nearly every piece of software available for them is also available at least on
PCs and probably NT machines as well. And NT machines running the fastest
DEC Alpha chips kick the crap out of anything below SGI for sure, and even
some of the SGI models.
And at least 5 devoted Mac fans that I know switched to NT this year to use
Softimage on it.
Micro$oft may not be innovative, and they may even be evil incarnate, but
machines running NT can definately get some serious work done at an affordable
price.
: JMonahan
--
-------------------------------------------------------------------------------
Steph Greenberg, 3DCGIMD CGI Character Orthopedic Surgeon,
st...@primenet.com Chiropractor, and Podiatrist.
Copyright 1996. All Rights Reserved. Permission granted for non-commercial
electronic republication only, such as Usenet and Email, and
non-commercial educational purposes such as charge free WWW pages.
Express permission is required for any other use. When in doubt, ask.
-------------------------------------------------------------------------------
Arijan Siska
Jim Mc Morrow (mo...@batelco.com.bh) wrote:
: The question with Silicon Graphics is....Is Alias good enough that you
: just have to have it if you are doing High end animation?If the answer
We are not talking about an absolute "performance" of the system. We are talking
about performance/$ and with the certan $ limit in mind right?
Sure Alias is great, better than 3dsMAX, but when I went to the local
dealer, and when he droped a package with a $20k price tag I immediately
cooled off.
There are very few buisnesses that will be profitable enough to be able
to buy Alias. And really big animation studios (read the articles in CGW) use
50% (on average) of their own software, making the high end 3d market even
smaller.
SGI will have to produce a pretty spectacular app. for it's machines to make
them appeal to wider public.
--
Arijan
E-mail: ari...@kette.fe.uni-lj.si
WWW: http://luz.fe.uni-lj.si/users/arijan-eng.html
Reto Koradi
Steph Greenberg wrote:
> And the Mac OS itself was copied off of X-windows developed at Xerox
> Park a decade earlier.
X windows was developed at MIT. The first X11 release was in
September '87, long after the first Mac. It's true that the first
window system was developed at Xerox, but this didn't have a
direct relation with X windows.
Richard Ottolini
In article <323552...@spectrospin.ch>,
Reto Koradi <k...@spectrospin.ch> wrote:
>Steph Greenberg wrote:
>> And the Mac OS itself was copied off of X-windows developed at Xerox
>> Park a decade earlier.
>
>X windows was developed at MIT. The first X11 release was in
>September '87, long after the first Mac. It's true that the first
>window system was developed at Xerox, but this didn't have a
>direct relation with X windows.
Sorry, X was invented at Stanford by Paul Asente.
His thesis advisor wanted to build a distributed graphics
system for the UNIX ethernet that was just starting at
that time (early 80s). Paul called his language "W" for
windowing system and the fashion of giving single letter
names to UNIX software packages.
The first window systems were developed at SRI and Evans
and Sutherland.
You may be thinking about raster graphics windows systems.
Twenty years from now these will all be dusty footnotes
to MicroSoft Windows :-( :-( :-(
Joshua Boyd
Arijan Siska wrote:
IMHO, PC's have taken the low end market from SGI (ie Indy and Indigo2
computers). At the high end, for shear rendering power, I would rather
use DEC alphas than Onyxs(Challenger is still offers more power than an
Alpha because you can use so many more processors.).
The only market left IMO for sgi is in realtime VR driven by the onyx.
For everything else I think that PC's have stolen the SGI market by
having equal power and a lower price. And yes I have really used the
Onyx and othe SGI workstations.
Ben Cannon
Noam Ben-Ami wrote:
> Well, nobody's perfect.
You should know :)
Noam, have you ever really used a SGI? I don't mean spendig ten minuites
compiling your OGL code on a PI either..
(I'm talking Indy r4400-200 or better)
I ask this simply because you don't think they're God's gift to 3D :)
Everyone I have ever met who has played around with at leawst Lightwave
on
the SGI agrees that they're about the best thing since swiss cheese!!!
Man, if you don't absolutely love IRIX after using it for a day,
something is
seriously wrong wiht you and you are wierd and strange :)
Just my $0.02.... no need to pay me back, consider it a gift :)
Best!
Ben.
Ben Cannon
Hey, some people think mice are the supreme beings on the face of this
Earth.....
If the shoe fits..... I can think of at least ten people I'd like to
smack with it!
Ben.
PS, I really think we should end this thread, it's become contageous, I
think...
Noam Ben-Ami
Joshua Boyd <bo...@zansiii.millersv.edu> writes:
Your lack of knowledge about SGI's platforms, market, and even PCs
is astounding.
Over and out.
Martin Preston
Arijan said:
> We are not talking about an absolute "performance" of the system. We are
> talking about performance/$ and with the certan $ limit in mind right?
I think thats probably the main reason for much of the disagreement in this
thread. If you need high end graphics then the actual cost of the machinery
isn't all that important (salaries are expensive, hardware is cheap). If you
need a particular thing doing in N weeks, then if only one machine can do
that..
thats what you buy!
So you could probably split the animation market into two groups:
Low to mid-end : Price/Performance counts, for a variety of reasons.
Mid to high-end: Performance counts above all else. (Price, as well as quality
of service, after sales help, on-site 12 hour maintenance etc. only come into
it when more than one manufacturer can meet your requirements).
A depressing number of posters to this thread believe that price/performance
is all, and can't understand why anyone would pay lots of money for a machine
that had higher performance ("But my amiga gives more bang for the buck").
(I'm not the first person to say this, and I won't be the last, but...)
Can we kill this thread please? Nothing of any insight has been posted for
weeks now, and we're just going over old ground.
Martin
------------------------------------------------------------------------------
|Martin Preston, Computer Graphics Unit, MC, | T: (+44) 161 275 6045 |
|University of Manchester, Manchester, M13 9PL, U.K. | F: (+44) 161 275 6040 |
|W: http://info.mcc.ac.uk/CGU/staff/preston/ | E: pre...@mcc.ac.uk |
------------------------------------------------------------------------------
50th Anniversary of stored program computing: http://www.cs.man.ac.uk/mark1/