-=* Industry Insighter Reports! *=-
Jose Elias
[ Author was Jose Elias ]
[ Posted on 5 Sep 1994 20:02:38 GMT ]
As a computer field insider, I'll talk about:
1) the 100-500MHZ chip rumors
2) the future of the new machines
(Saturn, U64, SuperFX, PSX, Jaguar, 3Do, etc...)
3) A warning about the Video Games industry
Let's begin:
Guys, I'm a computer engineer with some connections in the
computer field (specially chip manufacturers), and i'm going
to tell you a few things to settle this 100-500MHZ thing:
Today's most powerfull desktop chips cost well over $250
in mass quantities, not only that, but only now are 90MHZ
versions of these chips available. If you think that you
will get a 500MHZ chip in the U64 I might as well take
out my old Atari 2600 and say it's better than any arcade
machine out today. As a matter of fact, there isn't _ONE_
chip on the market that even comes _close_ to 500MHZ...
Second, stop the RISC/CISC battle. RISC is mostly hype,
but that doesn't mean I think CISC is better. RISC
(Reduced Instruction Set Computing) and CISC (Complex
Instruction Set Computing) are just two ways of arriving
at the same thing, and as a matter of fact there isns't
even ONE chip on the market that really fits the definition
of either one, they just try to get as close to it as possible,
but both CISC and RISC design borrow ideas from each other
in actual implementations.
Third, I do not doubt that Nintendo/SGI could comke up with
something around $300, and it could be VERY powerfull, but
it's not going to be the KILLER 1-million Gigaherts 64-trilion
colors Nintendo is making everyone beleive, as a matter of fact,
if you had the chance of playing the prototype U64 and Sega's
Saturn you'll see how Nintendo is just hyping up, in order
to make up for the time they're spending while SEGA and SONY
come out with products. It's all marketing.
Fourth, PLEASE do NOT consider me a SEGA fan who hates Nintendo,
I actually love them both, I'm just saying this because people
who're not in the industry should know what REALLY is behind
doors.
OK, now the next part is my personal opinion on the future of
the new consoles:
o Atari Jaguar, it's undoubtly the most powerfull thing out today,
but _as_it_is_ it won't make it, due mainly to shortage of games,
and no new games. Let's face it, Tempest 2000 is awesome, but
people want MKII and SFII, not just a fifth remake of pacman.
Also Atari MANY times in the past has been known for letting
users in the cold. They abandoned the 2500, 2700, and Atari Lynx.
If they want to succeed, they have to convince the mayor players
into making current arcade games for the jaguar, and LET people
know about the jaguar.
o Sega 32X, if it arrives before the end of this year as promised
it will be a big success, but ONLY if SEGA somehow comes up with
a promotion actually showing people what its powerfull hardware
can do. Why? because people were VERY dissapointed with the SEGA-CD,
sure it sold a lot of units, but the games don't really show a leap
forward in technology as the 32X really can.
o 3DO, is just "hanging" there with a lot of hype, it was suffering a
slow and painfull death until very recently when a new wave of software
has been annouced as almost completed, also Super Street Fighter II Turbo
will place it on the map as a "real" video games machine. HOWEVER, if 3DO
doesn't lower the price SIGNIFICANTLY _AND_ releases all the killer software
BEFORE all the big players come out with their new units, then you can kiss
3DO good bye for ever (remember, SEGA and Nintendo _CAN_ afford to have
moderetelly-priced units out because they're already stablished in the
gaming community. Would you buy a $400 3DO or a $400 SEGA Saturn???)
o SEGA Saturn, will be a killer, due mainly to all the hype being created
around it (remember the movie Batman? we were injected advertisement
for it a YEAR before it came out, when it finally came out it sucked,
but still did amazing at the box office, HOWEVER, this time the Saturn
WON'T be a dissapointment, at least prototype games of Daytona have the
promise of being BETTER than the actual arcade). HOWEVER, an abusive
prising from SEGA's part would be dangerous, since at the time it comes
out here all the big players will also be asking for high prices in a
saturated market, and "daddy" doesn't feel like banging out $450 bucks
for a new games machine for christmas...
o Nintendo U64, the same I said for the Saturn above applies to the U64,
HOWEVER, Nintendo will be in more danger than SEGA since by the time it
comes out (it seems) all the big players will be already out with games
to spare. But as we all know big-N has been known for coming out with
great games at the right time.
o Neo-Geo, this one's dead, from the beginning they should have released the
home version with a CD instead of tape. Who's going to be paying between
$150 and $200 for each new game??? (a CD holds 5-10 times as much as a
Neo-Geo cart, yet it costs 1 dollar (yes, one dollar) to manufacture.
o NEC SuperFX, accoring to NEC, what they have will blow away everything out
of the face of the earth, problem is, software. in japan they could do good,
but in the US it would be tougher, they'd have to come up with a TOTALLY
new name for it, a new design, make it look like something REALLY new,
or people won't buy it. And don't forget the most important thing: games.
o SONY PSX, Looks hot, and could be a hit in consumer stores. Dealers trust
SONY, and this is a BIG plus, dealers will push the machine, and SONY will
advertise like crazy. This could become the third big kid of the block
together with SEGA and Nintendo. Also their machine has impressive specs,
all they need is the games now to invade the US.
One last remark, the video games market is in for a BIG surprise. I can
forsee something "ugly" happening within 1 to 3 years. We won't see a
crash, but many companies will be hurt. People will just NOT like the
idea of paying $350 for a games machine every 6 months, and MOST important,
games are getting WAY to expensive. $80 for a fighting game is OUTRAGEOUS,
I only hope that we see games in the $25-35 range again (this is more than
just possible on CD distribution) or the video games market will die in
its own greediness...
Jose Elias
HaCkEr
hac...@acs.bu.edu
p.s.: Permision is granted to cross-post to other newsgroups,
but other use of my opinions expressed here has to be consulted
with me first.
.---'\/\/\/`--- - - - - - - - - - - ---'\/\/\/`---.
| hac...@acs.bu.edu |
`----------[ Imagination is more important than knowledge... ]----------'
--
.---'\/\/\/`--- - - - - - - - - - - ---'\/\/\/`---.
| hac...@acs.bu.edu |
`----------[ Imagination is more important than knowledge... ]----------'
Scott Campbell
own that atari needs games, and nintendo won't have a 500 mhz chip in the
ultra 64.
Ralph Barbagallo
2500 and 2700. I've never seen a SINGLE game for it! How does Atari
expect to succeed with that kind of support? Geesh. And talk about
bad distribution...did you EVER see one in the stores? It's all over.
--
Ralph A. Barbagallo III --- rbar...@cs.uml.edu --- Only AMIGA makes it
Possible...
R.I.P. Jay Miner: June 20th 1994 Father of the 2600, Atari 8bit, Amiga
-=-VAPS Member=-=-Classic Game Collector=-=-Broke Freelance Journalist=-
Glenn M. Saunders
: o Atari Jaguar, it's undoubtly the most powerfull thing out today,
: but _as_it_is_ it won't make it, due mainly to shortage of games,
: and no new games. Let's face it, Tempest 2000 is awesome, but
: people want MKII and SFII, not just a fifth remake of pacman.
: Also Atari MANY times in the past has been known for letting
: users in the cold. They abandoned the 2500, 2700, and Atari Lynx.
^^^^^^^^^^^
Ignorance alert!
: o Neo-Geo, this one's dead, from the beginning they should have released the
: home version with a CD instead of tape. Who's going to be paying between
^^^^
: $150 and $200 for each new game??? (a CD holds 5-10 times as much as a
: Neo-Geo cart, yet it costs 1 dollar (yes, one dollar) to manufacture.
Ignorance alert!! I hope you don't call carts tapes and think they are
synonymous. I stopped thinking carts were tapes when I was 11.
Jochen Scharrlach
> o Atari Jaguar, it's undoubtly the most powerfull thing out today,
> but _as_it_is_ it won't make it, due mainly to shortage of games,
> and no new games. Let's face it, Tempest 2000 is awesome, but
[...]
> o 3DO, is just "hanging" there with a lot of hype, it was suffering a
> slow and painfull death until very recently when a new wave of software
> has been annouced as almost completed, also Super Street Fighter II Turbo
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Have you ever heard of some new Jag games which will come in the near
future? Ok, you could say, it doesn't matter, because they aren't here
*now*, but please use the same argument for the 3D0!
Bye,
Jochen
--
------------------------------------
EMail: jsch...@ba-stuttgart.de
or: aco...@ftp.uni-stuttgart.de
------------------------------------
Erik Fortune
Except, of course, that the Ultra 64 *will* have memory clocked at 500Mhz.
-- Erik
Wesley Potter
Erik Fortune <er...@westworld.esd.sgi.com> wrote:
>
>Except, of course, that the Ultra 64 *will* have memory clocked at 500Mhz.
>
>-- Erik
Would you mind telling us morons out in the real world why you would
choose to waste the consumer's money on 500 MHz ram? That would have to
be handpicked to say the least. Besides, unless you have a 500MHz
processor (which according to our computer engineer friend doesn't exist)
or you have 5 100MHz processors reading or writing to the same RAM in
the same clockcycle (which would be inconvenient at best), it is worthless.
Erik Fortune
> Would you mind telling us morons out in the real world why you would
> choose to waste the consumer's money on 500 MHz ram? That would have to
> be handpicked to say the least. Besides, unless you have a 500MHz
> processor (which according to our computer engineer friend doesn't exist)
> or you have 5 100MHz processors reading or writing to the same RAM in
> the same clockcycle (which would be inconvenient at best), it is worthless.
We've been over this about a half-dozen times and people who know a lot more
about Rambus memory than I do have posted fairly detailed descriptions. Is
there an FAQ we can put this into?
Rambus memory takes some normal length of time (56ns?) to deliver a random
byte in memory and then delivers sequential bytes every 2ns. You can
(probably) think of this as 32-bits every 8 nanoseconds (i.e. comparable
to SRAM), if that makes you happier. The chief advantage of Rambus memory
is that it delivers high throughput over a narrow bus (9 bits, 30 pins if
I remember correctly) which translates into lower manufacturing costs.
Rambus memory costs 20% more than standard DRAM in quantity (and Nintendo
will assuredly be buying memory in quantity).
It seems to me that video games will tend to access memory sequentially (for
images, textures, etc) more often than arbitrary applications, making Rambus
memory an even more attractive choice.
If you really care and you have the technical background to understand it,
why don't you contact Rambus directly and see if they have some information
they can send you. They're in Mountain View (CA) -- I'm sure you can
get their number from information at 1-415-555-1212.
-- Erik
Public Image
That's okay.
My Atari 2600 joystick was clocked at twice that :)
Memory cannot be clocked in MHz. it doesn't make any sense.
Don't flame me, I read the pointless debating in alt.atari-jag
-Scott
: -- Erik
--
----------------------------------------------------------------------
|Scott Call aka Zapo Zapper | Public Image Limited BBS 510-831-3634 |
|VBBS Support/anime stuff | Owner: Atari Jag and Gravis Ultrasound |
----------------------------------------------------------------------
Jose Elias
[ Author was Jose Elias ]
[ Posted on 7 Sep 1994 06:21:01 GMT ]
Jose Elias (hac...@bu.edu) wrote:
: [ Article crossposted from rec.games.video.sega ]
: Let's begin:
: o Atari Jaguar, it's undoubtly the most powerfull thing out today,
: but _as_it_is_ it won't make it, due mainly to shortage of games,
: and no new games. Let's face it, Tempest 2000 is awesome, but
: people want MKII and SFII, not just a fifth remake of pacman.
: Also Atari MANY times in the past has been known for letting
: users in the cold. They abandoned the 2500, 2700, and Atari Lynx.
: If they want to succeed, they have to convince the mayor players
: into making current arcade games for the jaguar, and LET people
: know about the jaguar.
: o Sega 32X, if it arrives before the end of this year as promised
: it will be a big success, but ONLY if SEGA somehow comes up with
: a promotion actually showing people what its powerfull hardware
: can do. Why? because people were VERY dissapointed with the SEGA-CD,
: sure it sold a lot of units, but the games don't really show a leap
: forward in technology as the 32X really can.
: o 3DO, is just "hanging" there with a lot of hype, it was suffering a
: slow and painfull death until very recently when a new wave of software
: has been annouced as almost completed, also Super Street Fighter II Turbo
: will place it on the map as a "real" video games machine. HOWEVER, if 3DO
: doesn't lower the price SIGNIFICANTLY _AND_ releases all the killer software
: BEFORE all the big players come out with their new units, then you can kiss
: 3DO good bye for ever (remember, SEGA and Nintendo _CAN_ afford to have
: moderetelly-priced units out because they're already stablished in the
: gaming community. Would you buy a $400 3DO or a $400 SEGA Saturn???)
: o SEGA Saturn, will be a killer, due mainly to all the hype being created
: around it (remember the movie Batman? we were injected advertisement
: for it a YEAR before it came out, when it finally came out it sucked,
: but still did amazing at the box office, HOWEVER, this time the Saturn
: WON'T be a dissapointment, at least prototype games of Daytona have the
: promise of being BETTER than the actual arcade). HOWEVER, an abusive
: prising from SEGA's part would be dangerous, since at the time it comes
: out here all the big players will also be asking for high prices in a
: saturated market, and "daddy" doesn't feel like banging out $450 bucks
: for a new games machine for christmas...
: o Nintendo U64, the same I said for the Saturn above applies to the U64,
: HOWEVER, Nintendo will be in more danger than SEGA since by the time it
: comes out (it seems) all the big players will be already out with games
: to spare. But as we all know big-N has been known for coming out with
: great games at the right time.
: o Neo-Geo, this one's dead, from the beginning they should have released the
: home version with a CD instead of tape. Who's going to be paying between
: $150 and $200 for each new game??? (a CD holds 5-10 times as much as a
: Neo-Geo cart, yet it costs 1 dollar (yes, one dollar) to manufacture.
: o NEC SuperFX, accoring to NEC, what they have will blow away everything out
jon stern
I read, in the electronics press, that Rambus RAMS (made by Hitachi) cost about
twice normal DRAM costs, not the 20% more you quote.
Who is correct? RAM cost is going to be a very significant factor in the new
consoles, so this is quite important.
--
--------------------------------------------------
Jon Stern
Dept. Electronic & Electrical Enginnering
University of Sheffield
Sheffield S1 3JD
UK
__ __ __ __
/___ heffield /___ ilicon /___ ystems / _ roup
___/ ___/ ___/ \__/
---------------------------------------------------
tCW
>: [ Article crossposted from rec.games.video.sega ]
>: [ Author was Jose Elias ]
>: [ Posted on 5 Sep 1994 20:02:38 GMT ]
>
>: As a computer field insider, I'll talk about:
>: 1) the 100-500MHZ chip rumors
>: 2) the future of the new machines
>: (Saturn, U64, SuperFX, PSX, Jaguar, 3Do, etc...)
>: 3) A warning about the Video Games industry
>
>: computer field (specially chip manufacturers), and i'm going
>: to tell you a few things to settle this 100-500MHZ thing:
If you need to go to this much work to make us think you are credible,
somenthing is wrong. I am also a Computer Engineer. That would have been
enough. Or you could have said you know somenthing about computer
architecture. You did not say that.
Thats not my point. Read on. You don't appear to understand Rambus really
well, so you should not try to say "I am going to settle the issue".
Furthermore, in my opinion, you say several other questionable things that
are misleading and/or incorrect.
>: Today's most powerfull desktop chips cost well over $250
>: in mass quantities, not only that, but only now are 90MHZ
>: versions of these chips available. If you think that you
>: will get a 500MHZ chip in the U64 I might as well take
>: out my old Atari 2600 and say it's better than any arcade
>: machine out today. As a matter of fact, there isn't _ONE_
>: chip on the market that even comes _close_ to 500MHZ...
You have missed the majority of the CPUs currently used. Many (most)
workstations use CPUs faster that 90MHz. You seem to only understand the
Pentium (or maybe PowerPC).
DECs Alpha CPU runs at > 200MHz. HP 9000s run at 100 to 150 MHz
(HP-PA). Maybe you don't consider those "desktop" chips, but the
distinction is blurring. And so called "desktop" chips have nothing to do
with the video game industry, as game machines are not tied down to the
1970s Intel x86 architecture like the home PC. So anyhow, many chips run
faster than 90MHz. Likewise, many cost more than $250 in quantity.
More importantly you have missed the whole point of Rambus. With fast
risc processors, there is a problem of code and data starvation: the memory
is not fast enough to feed the pipelines of the CPU(s). Rambus is one of
several solutions to this problem. It can transfer ***500MB/sec*** over a
8 bit bus. The memory used is from Rambus, and you use a rambus core in your
chip to interface with this memory.
I don't really blame you for the 500MHz line. That had been going around,
and is misleading. 500MB on a 8 bit bus cooresponds to a 500MHz frequency,
if at maximum bandwidth. There is no reason you *NEED* to use the data that
quickly. That is peak! In fact, sustained rates are more along the 250MB/sec.
Now a little math:
90MHz x 4bytes/cycle = 360MB/sec.
That is the maximum rate of a 32 bit CPU at 90MHz with 1 cycle 32 bit memory
access. Now 500MB/sec doesn't sound so far fetched, huh?
I am not going to go into more detail unless requested. Basically the point
of Rambus is to allow the CPU to get data as fast as it needs it. It is sort
or like using only fast static RAM, but cheaper and more feasable. Rambus and
other such solutions offer a clear advantage to DRAM, which at 50ns (very
fast), again on a 32 bit bus offers only about 80MB/sec, which is too slow
for some applications.
Whether or not a game machine needs to move 500MB/sec is up to debate. But
things like Rambus can be used very efficiently in these applications, and
it does not require a 500MHz CPU to use!
>: Second, stop the RISC/CISC battle. RISC is mostly hype,
>: but that doesn't mean I think CISC is better. RISC
>: (Reduced Instruction Set Computing) and CISC (Complex
>: Instruction Set Computing) are just two ways of arriving
>: at the same thing, and as a matter of fact there isns't
>: even ONE chip on the market that really fits the definition
>: of either one, they just try to get as close to it as possible,
>: but both CISC and RISC design borrow ideas from each other
>: in actual implementations.
This is horrible. Did you really say you had a degree? The Pentium is
sort of a blur between RISC and CISC, but it is a fairly ugly architecture,
and is only considered risc in so far as some internal architecture and
microcode architecture.
RISC is not mostly hype. I won't say it is necessarily better (time will
show that it is, but . . .) It is a general architecture to eliminate
problems with CISC machines. In particular, it is more suited to super
scalar processing, parallel processing, etc. If you didn't know, the
Pentium can have opcodes from 1 byte to over 15 bytes. This is a nightmare
when you try to do things like dynamic branch prediction, super-scalar
pipelining, etc. Imagine you need to fill two pipelines, but you have no
idea where one instruction starts and the next ends. You need to
actually decode portions of the op-code to determine the length of it. In
RISC, you have a pre set size, so this issues is non-existant.
You can bet people like Intel are going to find ways to get very good
performance out of their sort of hybrid RISC/CISC architecture.
Anyhow, to say there is no chip that fits the definition of "RISC" is
very very naive. The definition is simple (basically just what the
name implies, as well as a paradigm of how this is beneficial).
Maybe few new chips are pure "CISC", but many are obviously "RISC"
>: o Atari Jaguar, it's undoubtly the most powerfull thing out today,
>: but _as_it_is_ it won't make it, due mainly to shortage of games,
>: and no new games. Let's face it, Tempest 2000 is awesome, but
>: people want MKII and SFII, not just a fifth remake of pacman.
>: Also Atari MANY times in the past has been known for letting
>: users in the cold. They abandoned the 2500, 2700, and Atari Lynx.
>: If they want to succeed, they have to convince the mayor players
>: into making current arcade games for the jaguar, and LET people
>: know about the jaguar.
There has never been a 2500 or a 2700. Is this a joke? But it is true
that software released over the next few months will probably make or break
the Jag, as the competition lurks on the horizon.
>: o Sega 32X, if it arrives before the end of this year as promised
>: it will be a big success, but ONLY if SEGA somehow comes up with
>: a promotion actually showing people what its powerfull hardware
>: can do. Why? because people were VERY dissapointed with the SEGA-CD,
>: sure it sold a lot of units, but the games don't really show a leap
>: forward in technology as the 32X really can.
The 32x is a quick effort to have somenthing out in the 32 bit realm before
the market is devoured. It is not super powerful, but yes, it will sell very
well. There are a lot of them pre-ordered.
>: o SEGA Saturn, will be a killer, due mainly to all the hype being created
>: around it (remember the movie Batman? we were injected advertisement
>: for it a YEAR before it came out, when it finally came out it sucked,
>: but still did amazing at the box office, HOWEVER, this time the Saturn
>: WON'T be a dissapointment, at least prototype games of Daytona have the
>: promise of being BETTER than the actual arcade). HOWEVER, an abusive
>: prising from SEGA's part would be dangerous, since at the time it comes
>: out here all the big players will also be asking for high prices in a
>: saturated market, and "daddy" doesn't feel like banging out $450 bucks
>: for a new games machine for christmas...
>
>: o Nintendo U64, the same I said for the Saturn above applies to the U64,
>: HOWEVER, Nintendo will be in more danger than SEGA since by the time it
>: comes out (it seems) all the big players will be already out with games
>: to spare. But as we all know big-N has been known for coming out with
>: great games at the right time.
Nintendo has more money in the bank than most big companies make (in sales)
in a year. This gives them significant marketplace clout. Also SGIs
involvement will be very benneficial (more so for SGI than Nintendo, but . . )
>: o Neo-Geo, this one's dead, from the beginning they should have released the
>: home version with a CD instead of tape. Who's going to be paying between
>: $150 and $200 for each new game??? (a CD holds 5-10 times as much as a
>: Neo-Geo cart, yet it costs 1 dollar (yes, one dollar) to manufacture.
This has been dead for a long time.
>: o NEC SuperFX, accoring to NEC, what they have will blow away everything out
>: of the face of the earth, problem is, software. in japan they could do good,
>: but in the US it would be tougher, they'd have to come up with a TOTALLY
>: new name for it, a new design, make it look like something REALLY new,
>: or people won't buy it. And don't forget the most important thing: games.
NEC will have a super big problem competing with Nintendo and Sega.
>: o SONY PSX, Looks hot, and could be a hit in consumer stores. Dealers trust
>: SONY, and this is a BIG plus, dealers will push the machine, and SONY will
>: advertise like crazy. This could become the third big kid of the block
>: together with SEGA and Nintendo. Also their machine has impressive specs,
>: all they need is the games now to invade the US.
This is going to be fairly expensive. Possible the same problem as the
3D0.
Erik Fortune
In article <34n2sr$l...@hippo.shef.ac.uk>, j...@shef.ac.uk writes:
> Erik,
>
> I read, in the electronics press, that Rambus RAMS (made by Hitachi) cost
> about twice normal DRAM costs, not the 20% more you quote.
>
> Who is correct? RAM cost is going to be a very significant factor in the new
> consoles, so this is quite important.
We're probably both correct. My understanding is that all other things
being equal, Rambus memory is about 20% more expensive to manufacture
than DRAM. Given that Rambus memory is not being manufactured in
huge quantities at the moment, you pay a premium for it. Once Nintendo (and
others) starting using Rambus memory in quantity the price should drop
to 120% of normal DRAM prices.
Pure speculation, but it fits the claiims made by all parties and certainly
seems reasonable.
-- Erik