RPC II technical
Andreas Dehmel
Right, some more questions on my part:
RO won't support multiprocessing, right? But the hardware should
be capable of it. Wasn't there some talk about Acorn adopting the
threads module or something? Will RO at least provide an interface
in the form of a module so apps written with MP in mind can at
least use multiple processors?
The only real problem on the software side I see with multiprocessing
support (in the simple form of different processes on different
processors) is the old issue of re-entrant modules (giving each
processor its own SVC / IRQ stacks should be handled by the OS quite
well). Couldn't this be helped by allowing only one processor to
access a module (via SWI or CLI)? SWIs should be really simple, just
patching the SWI handler and using a ``chunk in use'' bitmap should
do; commands shouldn't be a problem either. If a processor tries
entering an occupied module it's put to sleep in a waiting queue and
every time a module is exited some code is called that checks the
waiting queue and wakes the next processor. Crudeish but shouldn't
be hard to implement.
That leaves the problem of different environments (like printing to
screen / printer) - that would require the biggest changes to RO, no
doubt, since it involves all the work that would have been required
for PMT as well. But it'd help if you could just thread out something
like TaskWindows for a start, that should definitely be possible without
too many changes to RO. Up to 5 concurrent compile-jobs sound pretty
good to me, at least.
OK, assuming (realistically) that RO will never support MP, what about
the commitment of the RiscBSD team in that respect? Because I expect
to run some UNIX on my next computer anyway it'd certainly help if
I could use multiple processors at least under one OS. Considering
that UNIX already is PMT the changes involved should be trivial compared
to the ones necessary in RO (bear in mind I said ``compared to'').
Galileo: Acorn are certainly working (or planning) on some sort of
RO shell sitting on top of Galileo. Since Galileo was designed
with MP in mind that should give us some form of MP involving RO.
Whether that means one independent RO-session for each processor or
a (hopefully) better integrated solution is mere speculation.
Now let's forget about Galileo with all its bells and whistles for
the time being, when can we expect the bare Galileo essentials to
run RO on top of Galileo and get MP support? Specifying the year would
be OK as an answer, quartals wouldn't hurt either.
The bus: still not sure what width it is, 32 or 64 bits? Since it
supports Pentiums it'd have to be 64 bit, right? Would this be the
reason for the small amount of SRAM on the mainboard, to connect
the SRAM with the bus with 64 bits and SRAM <-> ARM with 32? If
the SRAM's clocked at double the bus speed that'd give us the
same throughput as if the ARM could use the 64bit bus directly.
What's the SRAM clocked at?
VRAM: I suppose that'll be similarily clocked as the main RAM, i.e.
66MHz. VIDC 20 used 64 bits (I don't think they changed that in the
new version), so that would give us roughly 500MB/s for the video
hardware, enough for 1152*896 @24bpp at over 120Hz, right?
FP: OK, Peter Bondar said RiscPCs would never have HFP, but did he
say RiscPC II? Also wouldn't it be rather stupid not to use a chip
like the SA1500/AMP once it's available? If it can't drive the
video system, how about as an additional processor (back to the MP
issue)? One thing's for sure, I won't buy a computer without HFP,
be it as nice as it may in other areas. Also am I the only one who
considers it rather strange that on one hand Acorn/ARM can design
a chip that does an entire FFT in hardware (Piccolo) but can't make
an FPU?
The most interesting aspect of the RiscPC II is MP, as far as I'm
concerned. If it can't do that even via specialised apps I'll think
twice before buying one.
OK, that's it for now. Some answers, please?
Andreas
Tim Tyler
Andreas Dehmel (deh...@Informatik.TU-Muenchen.DE) wrote:
[snip]
: OK, assuming (realistically) that RO will never support MP, what about
: the commitment of the RiscBSD team in that respect?
Certainly Neil Carson, who is involved in the BSD port, was also responsible
for developing some of the software for Simtec's Hydra card...
ATM AFAICS there's not much point in writing MP software for Acorn machines
because they don't tend to have multiple processors. Even a fully populated
Hydra goes at the same speed as a single SA, and both are having bandwidth
problems with the system's bus.
Until Acorn makes a machine with more than one processor this seems unlikely
to change. When they do so, I would expect Casuality to have the angle
covered...
: Galileo: Acorn are certainly working (or planning) on some sort of
: RO shell sitting on top of Galileo. Since Galileo was designed
: with MP in mind that should give us some form of MP involving RO.
/If/ there's some suitable hardware...
: Now let's forget about Galileo with all its bells and whistles for
: the time being, when can we expect the bare Galileo essentials to
: run RO on top of Galileo and get MP support? Specifying the year would
: be OK as an answer, quartals wouldn't hurt either.
My bet goes on the year 200x. I'm not making any bets about the x.
: FP: OK, Peter Bondar said RiscPCs would never have HFP, but did he
: say RiscPC II?
Pretty much, yes, at least until some suitable ARM with it is made.
: Also wouldn't it be rather stupid not to use a chip
: like the SA1500/AMP once it's available?
It was explicitly said that this was not considered suitable for the RPC.
I don't see quite what the problem is, myself...
: If it can't drive the
: video system, how about as an additional processor (back to the MP
: issue)? One thing's for sure, I won't buy a computer without HFP,
: be it as nice as it may in other areas. Also am I the only one who
: considers it rather strange that on one hand Acorn/ARM can design
: a chip that does an entire FFT in hardware (Piccolo) but can't make
: an FPU?
Hardware FFTs are cool and useful. I (for one) have little use for FP.
ARM can (and have) made FPUs. Acorn have put them inside their computers, too.
: The most interesting aspect of the RiscPC II is MP, as far as I'm
: concerned.
It would be so for me, perhaps, if it showed any sign of supporting it...
--
__________
|im |yler The Mandala Centre http://www.mandala.co.uk t...@cryogen.com
Andreas Dehmel
deh...@Informatik.TU-Muenchen.DE (Andreas Dehmel) writes:
Following up my own article... do you get blind from this sort of thing?
>The only real problem on the software side I see with multiprocessing
>support (in the simple form of different processes on different
>processors) is the old issue of re-entrant modules (giving each
>processor its own SVC / IRQ stacks should be handled by the OS quite
>well). Couldn't this be helped by allowing only one processor to
>access a module (via SWI or CLI)? SWIs should be really simple, just
>patching the SWI handler and using a ``chunk in use'' bitmap should
>do; commands shouldn't be a problem either. If a processor tries
>entering an occupied module it's put to sleep in a waiting queue and
>every time a module is exited some code is called that checks the
>waiting queue and wakes the next processor. Crudeish but shouldn't
>be hard to implement.
... and I remembered yesterday that a simple system like this can
give you deadlocks which are absolutely unacceptable for running
processes. That leaves two options:
1) Instantiate each module for each process that accesses it, thus
essentially creating a process-specific environment. This has
severe disadvantages, like you need some mechanism to sort of
inherit the setup from a default module, and of course you'll
need busloads of memory (just imagine SparkFS used by 10 processes.
Brrrr...). Not a good idea, IMHO.
2) Make each module declare all its ``resources'' (modules it
accesses) and it may only be entered if all of these are free.
Considering all modules use OS_... calls a special treatment
for those would be called for or we'd be back to ``only one
processor in the system'' which is the crudest implementation
of them all. OTOH considering OS_... calls shouldn't call any
SWIs but OS_... calls themselves I suppose they can't cause
deadlocks so may be excluded from the list (i.e. processor A
can enter a module even if another processor B is just in OS_...
code, although A will have to wait for B to leave the OS
before it's allowed to enter the OS (or any other ``occupied''
module) itself).
...which requires a new module interface, but I suppose we all
guessed that and besides it's no big change.
>That leaves the problem of different environments (like printing to
>screen / printer) - that would require the biggest changes to RO, no
>doubt, since it involves all the work that would have been required
>for PMT as well. But it'd help if you could just thread out something
>like TaskWindows for a start, that should definitely be possible without
>too many changes to RO. Up to 5 concurrent compile-jobs sound pretty
>good to me, at least.
A little note on that too: TaskWindows are currently not running any
tasks that change the environment in a way that would affect other
tasks (like you can't print from a TaskWindow). So stuff that's running
from one even now should also run in true parallel mode without too
much hassle.
Andreas
Jon Harrop
In article <63ugnb$7...@sunsystem5.informatik.tu-muenchen.de>, Andreas Dehmel
<URL:mailto:deh...@Informatik.TU-Muenchen.DE> wrote:
> deh...@Informatik.TU-Muenchen.DE (Andreas Dehmel) writes:
>
> Following up my own article... do you get blind from this sort of thing?
What about Java running in parallel?
Cheers,
Jon.
Eddie Edwards
In article <63s101$h...@sunsystem5.informatik.tu-muenchen.de>, Andreas Dehmel
<URL:mailto:deh...@Informatik.TU-Muenchen.DE> wrote:
> Also am I the only one who
> considers it rather strange that on one hand Acorn/ARM can design
> a chip that does an entire FFT in hardware (Piccolo) but can't make
> an FPU?
That'd be a fixed-point FFT, I should think ... ARM does have a nice
integer multiplier, but FP is (I guess) a different kettle of fish.
--
Eddie Edwards (ed...@powerslv.demon.co.uk)
Games & Entertainment software for RISC OS/ARM/N64
Daniel J Blueman
Andreas Dehmel wrote:
> The bus: still not sure what width it is, 32 or 64 bits? Since it
> supports Pentiums it'd have to be 64 bit, right? Would this be the
> reason for the small amount of SRAM on the mainboard, to connect
> the SRAM with the bus with 64 bits and SRAM <-> ARM with 32? If
> the SRAM's clocked at double the bus speed that'd give us the
> same throughput as if the ARM could use the 64bit bus directly.
> What's the SRAM clocked at?
> VRAM: I suppose that'll be similarily clocked as the main RAM, i.e.
> 66MHz. VIDC 20 used 64 bits (I don't think they changed that in the
> new version), so that would give us roughly 500MB/s for the video
> hardware, enough for 1152*896 @24bpp at over 120Hz, right?
OK, the data and address busses on Pheobe will both be 32bits wide.
The PCI bus on Pheobe and all PCs, PowerMacs etc. are all 32bits wide.
The CPU bus speed will be 66Mhz, and the PCI bus speed will be 33Mhz.
The SRAM used is a pseudo-cache, and it will run at syncronous to the CPU
bus speed.
RISC OS will use it to store frequently accessed code runs, data stacks,
etc. (this is employed since SDRAM is excellent at long bursts, but not as
good for random access).
Hope that helps,
Daniel J Blueman
Ernst Dinkla
I posted a message some days ago with the subject title
'comparable specifications ?'
it contained the specifications for the 3 new Apple models,
I removed any Apple indication to prevent a brands discussion.
It also shows Apple is no longer the expensive alternative for
Wintel machines.
To prove that the specification of the RPC2 isn't up to the
1998 standard part of the specification is quoted here.
You will also see that on Powermacs the information you give
is not correct.
Each Power Macintosh G3 system features a 233 or 266 MHz
PowerPC G3 processor with a 512K level 2 memory cache
running at 117 or 133 MHz; a 66 MHz system and main
memory bus; and, 32MB of RAM (expandable up to 192 or
384MB). Built into every configuration are ATI 3D RAGE ll+
64-bit graphics and multimedia accelerator chip with 2MB of
graphics memory (expandable to 6MB), three standard 12-inch
PCI slots, SCSI, IDE, Ethernet, stereo audio, Apple Desktop
Bus and serial interfaces. Storage capabilities include an
internal 4GB or 6GB IDE hard drive; Internal high capacity
floppy disk drive; an internal 24x-speed CD-ROM drive
(maximum); an internal 100MB Zip drive in some
configurations and expansion bays for additional storage
devices. Optional items include video capture and a 56K
modem.
* Power Macintosh G3 (desktop model): U.S. $2,000
233MHz/32MB RAM
(supports up to 192MB)/4GB Hard Drive/24x-speed
CD-ROM drive (maximum)/2MB Video
SGRAM/512K level 2 backside cache/66MHz system
bus speed/117MHz backside bus speed/two 3.5-in SCSI
bays for additional storage devices
Ernst
--
Even a blind pig finds an acorn sometimes
Ernst Dinkla Serigrafie,Zeefdruk edi...@inter.nl.net
Jon Harrop
In article <ant17120...@edinkla.inter.nl.net>, Ernst Dinkla
<URL:mailto:edi...@inter.nl.net> wrote:
>
> * Power Macintosh G3 (desktop model): U.S. $2,000
> 233MHz/32MB RAM
> (supports up to 192MB)/4GB Hard Drive/24x-speed
> CD-ROM drive (maximum)/2MB Video
> SGRAM/512K level 2 backside cache/66MHz system
> bus speed/117MHz backside bus speed/two 3.5-in SCSI
> bays for additional storage devices
Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
accelerator running Linux for 1350 quid.
The Risc PC II is too little, too late. The specs were
bad before, now they have become even worse and the
release date has been put back.
I suspect that there is now little or no money to be
made in RISC OS computers and that Acorn will give up
on them over the next year. Just look at the software
market...
Cheers,
Jon.
Barry Wickett
In article <ant19153...@isiselec.demon.co.uk>, Jon Harrop
<URL:mailto:j...@isiselec.demon.co.uk> wrote:
> In article <ant17120...@edinkla.inter.nl.net>, Ernst Dinkla
> <URL:mailto:edi...@inter.nl.net> wrote:
> >
> > * Power Macintosh G3 (desktop model): U.S. $2,000
> > 233MHz/32MB RAM
> > (supports up to 192MB)/4GB Hard Drive/24x-speed
> > CD-ROM drive (maximum)/2MB Video
> > SGRAM/512K level 2 backside cache/66MHz system
> > bus speed/117MHz backside bus speed/two 3.5-in SCSI
> > bays for additional storage devices
>
> Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
> bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
> accelerator running Linux for 1350 quid.
>
> The Risc PC II is too little, too late. The specs were
> bad before, now they have become even worse and the
> release date has been put back.
There is more to a computer than merely the specs. I don't care how
fast a PC can be, if it has got W95 then forget it. I'll go for a Mac
or Acorn.
>
> I suspect that there is now little or no money to be
> made in RISC OS computers and that Acorn will give up
> on them over the next year. Just look at the software
> market...
>
The NCs market will be worth quite a lot... I don't think they will
give up on those over the next year. Those little beauties use RISCOS
even if it is not so obvious!
Let's not simply talk the market down, and down!
On our business network we have PCs, Macs AND Acorns! I'm lucky
enough to be able to work from a SA RPC and will be getting the RPC
II on my desk for sure! You wouldn't believe some of the ways we use
Acorns here.
Anyway, I'm getting off topic a little here!
Regards,
--
Barry Wickett
Systems Co-ordinator
TECHNIQUEST
Stuart Street
Cardiff CF1 6BW
T 01222 475475
F 01222 482517
Jon Harrop
In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
<URL:mailto:Ba...@tquest.org.uk> wrote:
> In article <ant19153...@isiselec.demon.co.uk>, Jon Harrop
> <URL:mailto:j...@isiselec.demon.co.uk> wrote:
> > Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
> > bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
> > accelerator running Linux for 1350 quid.
> >
> > The Risc PC II is too little, too late. The specs were
> > bad before, now they have become even worse and the
> > release date has been put back.
>
> There is more to a computer than merely the specs. I don't care how
> fast a PC can be, if it has got W95 then forget it. I'll go for a Mac
> or Acorn.
Like I said, Linux...
> > I suspect that there is now little or no money to be
> > made in RISC OS computers and that Acorn will give up
> > on them over the next year. Just look at the software
> > market...
>
> The NCs market will be worth quite a lot... I don't think they will
> give up on those over the next year. Those little beauties use RISCOS
> even if it is not so obvious!
I don't believe they will continue to use RISC OS as it is
not pre-emptive which makes writing efficient web software
a pig.
> On our business network we have PCs, Macs AND Acorns! I'm lucky
> enough to be able to work from a SA RPC and will be getting the RPC
> II on my desk for sure! You wouldn't believe some of the ways we use
> Acorns here.
Acorns just don't hack it anymore for me.
Cheers,
Jon.
Greg Hennessy
On Thu, 20 Nov 1997 18:15:04 +0000 (GMT), Ernst Dinkla
<edi...@inter.nl.net> wrote:
>more and more acceptable. The second choice seems a good idea
>for networking and in Holland slightly older models DEC Alpha's
>are cheap. Could add some NC's to it too :-)
>
The following was posted into the vendor conference on CIX a few weeks
back.
Alpha PC164SX motherboard with 21164PC 533MHz cpu
1mb level 2 cache
64mb SDRAM
3.5" FD
3.2GB EIDE hard drive
24 speed EIDE CD Rom drive
Matrox Millenium II 4mb PCI video card
Full tower ATX case
Cherry keyboard
Microsoft mouse
NT4 Workstation
CTX 15" SVGA colour monitor
#1,975 + vat (#2,320.63 inc. vat)
The above will run Linux just dandy as well.
Given that the price being discussed for the RPC II is around the same
figure, there is no competition here. I would estimate that Acorn
will sell maybe 1000-1500 machines at most to the dwindling base of
enthusiasts.
greg
>
>
>
Barry Wickett
In article <ant20195...@isiselec.demon.co.uk>, Jon Harrop
I believe they will continue to use RISC OS for quite a while. The
amount of software around is not huge :-(. The people who create the
great software out there will not rewrite it because of a new OS!
>
> > On our business network we have PCs, Macs AND Acorns! I'm lucky
> > enough to be able to work from a SA RPC and will be getting the RPC
> > II on my desk for sure! You wouldn't believe some of the ways we use
> > Acorns here.
>
> Acorns just don't hack it anymore for me.
Well I manage to work every day VERY nicely on mine (much to the envy
of most other people in the office). I don't see any reason why this
shouldn't continue. We intend to introduce NCs in the office in the
near future anyway.
NCWrite using Word format hacks it for me!
Barry Wickett
In article <ant20180...@edinkla.inter.nl.net>, Ernst Dinkla
<URL:mailto:edi...@inter.nl.net> wrote:
> Barry Wickett wrote:
> > On our business network we have PCs, Macs AND Acorns! I'm lucky
> > enough to be able to work from a SA RPC and will be getting the RPC
> > II on my desk for sure! You wouldn't believe some of the ways we use
> > Acorns here.
>
> So it is possible to have Macs and Acorns on a network without
> problems?
Err, not quite. They sit on the same network, they just don't talk to
each other very well.
We're waiting for the AppleTalk part of Omniclient 2 to be released.
Meanwhile I use the TCP/IP protocol to ftp files around when I need
to. I use POP3 to fetch my mail from our mail server.
You can connect Macs and Acorns to a suitable fileserver (NTserver
does it!). Just don't try it using Novell.
Daniel J Blueman
Jon Harrop wrote:
> In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
> <URL:mailto:Ba...@tquest.org.uk> wrote:
> > In article <ant19153...@isiselec.demon.co.uk>, Jon Harrop
> > <URL:mailto:j...@isiselec.demon.co.uk> wrote:
[Snip, snip]
> Like I said, Linux...
Linux has it's problems ;-)
> > > I suspect that there is now little or no money to be
> > > made in RISC OS computers and that Acorn will give up
> > > on them over the next year. Just look at the software
> > > market...
> >
> > The NCs market will be worth quite a lot... I don't think they will
> > give up on those over the next year. Those little beauties use RISCOS
> > even if it is not so obvious!
>
> I don't believe they will continue to use RISC OS as it is
> not pre-emptive which makes writing efficient web software
> a pig.
That's where you are wrong. If you are a good programmer, and use aync.
events you can have very efficient web software. Eg. I am developing a
webserver for RISC OS which can blast data out with little CPU overhead,
with very low latency.
Clearly, multitasking and context switching needn't even be involved much
for a dedicated server, if you write it optimally.
> > On our business network we have PCs, Macs AND Acorns! I'm lucky
> > enough to be able to work from a SA RPC and will be getting the RPC
> > II on my desk for sure! You wouldn't believe some of the ways we use
> > Acorns here.
>
> Acorns just don't hack it anymore for me.
>
> Cheers,
Yeah, cheers
> Jon.
Daniel J Blueman
Richard Walker
In message <ant20195...@isiselec.demon.co.uk>
Jon Harrop <j...@isiselec.demon.co.uk> wrote:
> In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
> <URL:mailto:Ba...@tquest.org.uk> wrote:
> >
> > The NCs market will be worth quite a lot... I don't think they will
> > give up on those over the next year. Those little beauties use RISCOS
> > even if it is not so obvious!
>
> I don't believe they will continue to use RISC OS as it is
> not pre-emptive which makes writing efficient web software
> a pig.
Oh, I dunno... why do you think Acorn are developing RISC OS 4?
They've realised that Galileo isn't going to be ready for desktop/NC
use until after the year 2000 (slightly exaggerated!), and /power/
apps like web browsing and Java interpreting are rather let down by
RISC OS 3... the easiest solution is to do some low-level admin on
good ol' RISC OS! :-)
Note that the above paragraph is my own opinion/theory!
--
Richard.
_____________________________________________________________________________
my software: http://www.manorsch.demon.co.uk/rw (LanManFS front-end)
*updated!*
Sveinung W. Tengelsen
In article <ant21122...@gromit.tquest.org.uk>, Barry Wickett
<URL:mailto:Ba...@tquest.org.uk> wrote:
> In article <ant20195...@isiselec.demon.co.uk>, Jon Harrop
> <URL:mailto:j...@isiselec.demon.co.uk> wrote:
> > In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
> > <URL:mailto:Ba...@tquest.org.uk> wrote:
> > > <URL:mailto:j...@isiselec.demon.co.uk> wrote:
> > > > Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
> > > > bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
> > > > accelerator running Linux for 1350 quid.
Nice server/programming-workstation. "Casual user" machine? Hah.
> > > > The Risc PC II is too little, too late. The specs were
> > > > bad before, now they have become even worse and the
> > > > release date has been put back.
What spex? The only I'm a mite disappointed over is the (alleged)
33MHz PCI-bus (byebye online video-editing) plus that the system-
bus at 66MHz *may* be a bit slow for slapping on four next-generation
StrongARM's, totalling a theoretical processing-power (SA core + AMP)
in the order of ~6000 bloody MIPS. Am I the only person getting goose-
bumps all over from *thinking* about this kind of ultra-performance? 8)
Though, if we get a reasoneably flexible case, one may be allowed to
upgrade the motherboard sometime late 1998. I hope.
> > > There is more to a computer than merely the specs. I don't care how
> > > fast a PC can be, if it has got W95 then forget it. I'll go for a Mac
> > > or Acorn.
RISC OS is still nicer to use than MacOS (8). It has more facilities,
sure, but Macs are still sold with these awful one-button mice, which
IMO is pure madness.
> > Like I said, Linux...
> >
> > > > I suspect that there is now little or no money to be
> > > > made in RISC OS computers and that Acorn will give up
> > > > on them over the next year. Just look at the software
> > > > market...
> > >
> > > The NCs market will be worth quite a lot... I don't think they will
> > > give up on those over the next year. Those little beauties use RISCOS
> > > even if it is not so obvious!
Depends how much NCOS-a-la-RISC OS- specific software gets written -
"multiplatform" it isn't, and certainly limits the possible number
of sales. But once written for RISC/NC OS, it's quite natural to
also do a RISC OS version. Isn't there an english saying which goes
along the lines "when it rains on the boss, it drips on the employee"
or something to that effect? 8) If/when the NC-concept takes off,
this will be the premise since the NC/RISC OS-markets will be
in an umpteen-to-one relationship.
> > I don't believe they will continue to use RISC OS as it is
> > not pre-emptive which makes writing efficient web software
> > a pig.
We need pre-emption anyway.
> I believe they will continue to use RISC OS for quite a while. The
> amount of software around is not huge :-(. The people who create the
> great software out there will not rewrite it because of a new OS!
I guess a RISC OS/old-architecture emulator is possible, with the sort
of processor-ooomph we may expect around "goodbye RISC OS, hello
Galileo"-time. :)
> > > On our business network we have PCs, Macs AND Acorns! I'm lucky
> > > enough to be able to work from a SA RPC and will be getting the RPC
> > > II on my desk for sure! You wouldn't believe some of the ways we use
> > > Acorns here.
> >
> > Acorns just don't hack it anymore for me.
Too bad for you. 8)
> Well I manage to work every day VERY nicely on mine (much to the envy
> of most other people in the office). I don't see any reason why this
> shouldn't continue. We intend to introduce NCs in the office in the
> near future anyway.
Likewise. .:)
> NCWrite using Word format hacks it for me!
S'okay - let's see the next version without Oracle's feature-mangling. 8)
--
Regards,
Sveinung W. Tengelsen
-----------------------------------------------------------------------
mailto:pixe...@sn.no | I have one illusion;
http://www.sn.no/~pixeleye/Index.htm | I have no illusions.
Sveinung W. Tengelsen
<snip>
> So it is possible to have Macs and Acorns on a network without
> problems?
If you put a Peecee between'em, no problem.
PS - OmniClient2 has a greyed-out "Appletalk" option. Have no idea
when they plan to put something "under" it. Someone@acorn care to
hint at an approximate releasedate? :)
Alex Hayward
Daniel J Blueman (daniel.j...@stud.umist.ac.uk) wrote:
: That's where you are wrong. If you are a good programmer, and use aync.
: events you can have very efficient web software. Eg. I am developing a
: webserver for RISC OS which can blast data out with little CPU overhead,
: with very low latency.
OK, so how do you do async disc IO under RISC OS? Write your own driver??
--
al...@hayward.u-net.com; xe...@ferret.lmh.ox.ac.uk
Sveinung W. Tengelsen
In article <ant22152...@isiselec.demon.co.uk>, Jon Harrop
<URL:mailto:j...@isiselec.demon.co.uk> wrote:
> In article <ant21212...@pixeleye.online.no>, Sveinung W. Tengelsen
> <URL:mailto:pixeleye@remove_this.sn.no> wrote:
> > > > > > ...
> > > > > > Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
> > > > > > bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
> > > > > > accelerator running Linux for 1350 quid.
> >
> > Nice server/programming-workstation. "Casual user" machine? Hah.
>
99% of all users (I.e. those who don't love to tinker with
Linux etc.).
> > > > > > The Risc PC II is too little, too late. The specs were
> > > > > > bad before, now they have become even worse and the
> > > > > > release date has been put back.
> >
> > What spex?
>
> The ones on the Acorn Cybervillage. The most optimistic
> place in the world...
Oh that, yes. 8)
> > The only I'm a mite disappointed over is the (alleged)
> > 33MHz PCI-bus (byebye online video-editing) plus that the system-
> > bus at 66MHz *may* be a bit slow for slapping on four next-generation
> > StrongARM's,
>
> Not the speed of the CPUs and the lack of multiple
> processors, graphics acceleration, floating point,
> etc?
Let's just say that certain developments *could* have been brought
to the market a bit faster, and then I don't mean the RPC2. I'm
lowly becoming a fan of the idea of postponing the RPC2-launch,
in order to wait out the availability of certain nice components.
> Sveinung, you disappoint me! ;-)
Spaced if you do, spaced if you don't... 8)
> > totalling a theoretical processing-power (SA core + AMP)
>
> And what makes you think that Acorn will be using
> a processor with an AMP core in the Risc PC II?
Logic. It will not be the SA1500 (for reasons explained in this group).
Hint: do you *really* believe Acorn would work with Dec on the "new
super-chip" (as PB has mentioned umpteen times lately, latest at AW '97)
without also doing a "desktop" version? Huh. 8)
> > in the order of ~6000 bloody MIPS.
>
> 'Bloody MIPs' I would agree with, but not the rest... ;-)
Explicitly?
> > Am I the only person getting goose-
> > bumps all over from *thinking* about this kind of ultra-performance? 8)
>
> Yes. ;-)
Okay, so you're a cool one - good 4 U. 8)
> > Though, if we get a reasoneably flexible case, one may be allowed to
> > upgrade the motherboard sometime late 1998. I hope.
>
> I don't care about upgrading the motherboard, I just
> want a decent spec in the first place.
What one wants and what may be attainable aren't always synonymous.
If you want the RPC2 *fast* (which I certainly do) this will (probably)
be a necessary sacrifice. Not Acorn's fault.
> > > > > The NCs market will be worth quite a lot... I don't think they will
> > > > > give up on those over the next year. Those little beauties use RISCOS
> > > > > even if it is not so obvious!
> >
> > Depends how much NCOS-a-la-RISC OS- specific software gets written -
> > "multiplatform" it isn't, and certainly limits the possible number
> > of sales. But once written for RISC/NC OS, it's quite natural to
> > also do a RISC OS version. Isn't there an english saying which goes
> > along the lines "when it rains on the boss, it drips on the employee"
> > or something to that effect? 8) If/when the NC-concept takes off,
> > this will be the premise since the NC/RISC OS-markets will be
> > in an umpteen-to-one relationship.
>
> Sure, but it is the 'If/when' that matters.
From my point of observation, it looks like a 80/20 chance in Acorn's
favour. Can't expound the reasons for this belief, since I'm not
allowed to.
> > > > I don't believe they will continue to use RISC OS as it is
> > > > not pre-emptive which makes writing efficient web software
> > > > a pig.
> >
> > We need pre-emption anyway.
>
> Agreed.
But of course. :)
Ian Lynch
ghen...@lhr-sys.dhl.com (Greg Hennessy) wrote:
> I would estimate that Acorn will sell maybe 1000-1500 machines at most
> to the dwindling base of enthusiasts.
And perhaps that is their plan. So what?
--
Ian
Greg Hennessy
On Sat, 22 Nov 1997 10:11:58 GMT, i...@ianlynch.demon.co.uk (Ian Lynch)
wrote:
>And perhaps that is their plan. So what?
It speaks volumes about the products life expectancy as a desktop
platform .
greg
Richard Walker
In message <ant22143...@isiselec.demon.co.uk>
Jon Harrop <j...@isiselec.demon.co.uk> wrote:
> In article <fce5acec47%walk...@manorsch.demon.co.uk>, Richard Walker
> <URL:mailto:runny...@mindless.com> wrote:
> > In message <ant20195...@isiselec.demon.co.uk>
> > Jon Harrop <j...@isiselec.demon.co.uk> wrote:
> >
> > > In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
> > > <URL:mailto:Ba...@tquest.org.uk> wrote:
> > > >
> > > > The NCs market will be worth quite a lot... I don't think they will
> > > > give up on those over the next year. Those little beauties use RISCOS
> > > > even if it is not so obvious!
> > >
> > > I don't believe they will continue to use RISC OS as it is
> > > not pre-emptive which makes writing efficient web software
> > > a pig.
> >
> > Oh, I dunno... why do you think Acorn are developing RISC OS 4?
>
> I don't think that!
Why not?
> > They've realised that Galileo isn't going to be ready for desktop/NC
> > use until after the year 2000 (slightly exaggerated!), and /power/
> > apps like web browsing and Java interpreting are rather let down by
> > RISC OS 3... the easiest solution is to do some low-level admin on
> > good ol' RISC OS! :-)
>
> No, power apps are let down by the hardware.
OK. Floating point. There, said it - no new version of RISC OS is
going to fix that, but for what Acorn are trying to get RISC OS to
do at the moment (in desktops, and other things) version 4 is just
what the doctor ordered.
> Acorn are trying to keep a few gullible, err sorry,
> loyal followers with them to the end...
So what? At the end of the day, I prefer to use RISC OS to Windows.
I can get my work done /so/ much faster, and remain calmer at the end
of it. I don't consider it a miracle when I get a single A4 page
printed OK, or my 'sound card works with game X'...
Mechanoid
In article <3476380b....@news.ncc.dhl.com>, ghen...@lhr-sys.dhl.com
(Greg Hennessy) wrote:
[specs snipped]
> The above will run Linux just dandy as well.
It doesn't run RiscOS software though does it?
> Given that the price being discussed for the RPC II is around the same
> figure, there is no competition here.
Compared to the current RiscPC, the RiscPC 2 would seem to compete
strongly. Sales of the current RiscPC have not declined.
> I would estimate that Acorn will sell maybe 1000-1500 machines at
> most to the dwindling base of enthusiasts.
Multiply that figure by 50 and I might agree.
Cheers,
--
Dan "Mech" Maloney.
__ _______ ______ __
/ |/ / __/ ___/ /_/ / # Mechanoid the Haemorroidal Android
/ /|_/ / _// /__/ __ / # Disclaimer: I may have been wrong deliberately
/_/ /_/___/\___/_/ /_/ # mailto:mech...@usa.net
Jon Harrop
In article <ant21212...@pixeleye.online.no>, Sveinung W. Tengelsen
<URL:mailto:pixeleye@remove_this.sn.no> wrote:
> > > > > ...
> > > > > Try dual 180MHz Pentium Pro, 512k L2 cache, 60MHz data
> > > > > bus, 64Mb SDRAM, 8x CD-ROM, 8Mb 128-bit 3D graphics
> > > > > accelerator running Linux for 1350 quid.
>
> Nice server/programming-workstation. "Casual user" machine? Hah.
Who you callin' casual user? ;-)
> > > > > The Risc PC II is too little, too late. The specs were
> > > > > bad before, now they have become even worse and the
> > > > > release date has been put back.
>
> What spex?
The ones on the Acorn Cybervillage. The most optimistic
place in the world...
> The only I'm a mite disappointed over is the (alleged)
> 33MHz PCI-bus (byebye online video-editing) plus that the system-
> bus at 66MHz *may* be a bit slow for slapping on four next-generation
> StrongARM's,
Not the speed of the CPUs and the lack of multiple
processors, graphics acceleration, floating point,
etc?
Sveinung, you disappoint me! ;-)
> totalling a theoretical processing-power (SA core + AMP)
And what makes you think that Acorn will be using
a processor with an AMP core in the Risc PC II?
> in the order of ~6000 bloody MIPS.
'Bloody MIPs' I would agree with, but not the rest... ;-)
> Am I the only person getting goose-
> bumps all over from *thinking* about this kind of ultra-performance? 8)
Yes. ;-)
> Though, if we get a reasoneably flexible case, one may be allowed to
> upgrade the motherboard sometime late 1998. I hope.
I don't care about upgrading the motherboard, I just
want a decent spec in the first place.
> > > > The NCs market will be worth quite a lot... I don't think they will
> > > > give up on those over the next year. Those little beauties use RISCOS
> > > > even if it is not so obvious!
>
> Depends how much NCOS-a-la-RISC OS- specific software gets written -
> "multiplatform" it isn't, and certainly limits the possible number
> of sales. But once written for RISC/NC OS, it's quite natural to
> also do a RISC OS version. Isn't there an english saying which goes
> along the lines "when it rains on the boss, it drips on the employee"
> or something to that effect? 8) If/when the NC-concept takes off,
> this will be the premise since the NC/RISC OS-markets will be
> in an umpteen-to-one relationship.
Sure, but it is the 'If/when' that matters.
> > > I don't believe they will continue to use RISC OS as it is
> > > not pre-emptive which makes writing efficient web software
> > > a pig.
>
> We need pre-emption anyway.
Agreed.
Cheers,
Jon.
Jon Harrop
In article <fce5acec47%walk...@manorsch.demon.co.uk>, Richard Walker
<URL:mailto:runny...@mindless.com> wrote:
> In message <ant20195...@isiselec.demon.co.uk>
> Jon Harrop <j...@isiselec.demon.co.uk> wrote:
>
> > In article <ant19170...@gromit.tquest.org.uk>, Barry Wickett
> > <URL:mailto:Ba...@tquest.org.uk> wrote:
> > >
> > > give up on those over the next year. Those little beauties use RISCOS
> > > even if it is not so obvious!
> >
> > not pre-emptive which makes writing efficient web software
> > a pig.
>
I don't think that!
> They've realised that Galileo isn't going to be ready for desktop/NC
> use until after the year 2000 (slightly exaggerated!), and /power/
> apps like web browsing and Java interpreting are rather let down by
> RISC OS 3... the easiest solution is to do some low-level admin on
> good ol' RISC OS! :-)
No, power apps are let down by the hardware.
Acorn are trying to keep a few gullible, err sorry,
loyal followers with them to the end...
Cheers,
Jon.
Jon Harrop
In article <8de11fed47%walk...@manorsch.demon.co.uk>, Richard Walker
<URL:mailto:runny...@mindless.com> wrote:
> > > > I don't believe they will continue to use RISC OS as it is
> > > > not pre-emptive which makes writing efficient web software
> > > > a pig.
> > >
> > > Oh, I dunno... why do you think Acorn are developing RISC OS 4?
> >
> > I don't think that!
>
There is money to be made elsewhere
> > > They've realised that Galileo isn't going to be ready for desktop/NC
> > > use until after the year 2000 (slightly exaggerated!), and /power/
> > > apps like web browsing and Java interpreting are rather let down by
> > > RISC OS 3... the easiest solution is to do some low-level admin on
> > > good ol' RISC OS! :-)
> >
> > No, power apps are let down by the hardware.
>
> OK. Floating point. There, said it - no new version of RISC OS is
> going to fix that, but for what Acorn are trying to get RISC OS to
> do at the moment (in desktops, and other things) version 4 is just
> what the doctor ordered.
I disagree. The applications in question were specifically Java and
web browsing where a wider bus would result in a far higher speed
increase than any RISC OS update could.
> > Acorn are trying to keep a few gullible, err sorry,
> > loyal followers with them to the end...
>
> So what? At the end of the day, I prefer to use RISC OS to Windows.
> I can get my work done /so/ much faster, and remain calmer at the end
> of it. I don't consider it a miracle when I get a single A4 page
> printed OK, or my 'sound card works with game X'...
No, you consider it a miracle to _have_ game X...
Cheers,
Jon.
Jon Harrop
In article <ant22173...@pixeleye.online.no>, Sveinung W. Tengelsen
> Linux etc.).
Hehe, oh, *them*... ;-)
> > > The only I'm a mite disappointed over is the (alleged)
> > > 33MHz PCI-bus (byebye online video-editing) plus that the system-
> > > bus at 66MHz *may* be a bit slow for slapping on four next-generation
> > > StrongARM's,
> >
> > Not the speed of the CPUs and the lack of multiple
> > processors, graphics acceleration, floating point,
> > etc?
>
> Let's just say that certain developments *could* have been brought
> to the market a bit faster, and then I don't mean the RPC2. I'm
> lowly becoming a fan of the idea of postponing the RPC2-launch,
> in order to wait out the availability of certain nice components.
Acorn seem to agree with you about postponing the Risc PC II,
although I am not convinced it is for the same reason. :-(
> > > totalling a theoretical processing-power (SA core + AMP)
> >
> > And what makes you think that Acorn will be using
> > a processor with an AMP core in the Risc PC II?
>
> Logic. It will not be the SA1500 (for reasons explained in this group).
Sure, but there isn't much demand for a desktop version...
> Hint: do you *really* believe Acorn would work with Dec on the "new
> super-chip" (as PB has mentioned umpteen times lately, latest at AW '97)
> without also doing a "desktop" version? Huh. 8)
I don't believe that Acorn have much of a say in what Intel
manufactures... 8(
Maybe they should give up on the RISC PC II and concentrate
on the A7000++ with the SA1500 at under 1000 quid. Then
everyone would he happier, non?
> > > Though, if we get a reasoneably flexible case, one may be allowed to
> > > upgrade the motherboard sometime late 1998. I hope.
> >
> > I don't care about upgrading the motherboard, I just
> > want a decent spec in the first place.
>
> What one wants and what may be attainable aren't always synonymous.
> If you want the RPC2 *fast* (which I certainly do) this will (probably)
> be a necessary sacrifice. Not Acorn's fault.
But possibly their downfall...
> > Sure, but it is the 'If/when' that matters.
>
> From my point of observation, it looks like a 80/20 chance in Acorn's
> favour. Can't expound the reasons for this belief, since I'm not
> allowed to.
Be careful with those cryptic comments, if the wind changes
you may turn into ... Martin Piper! ;-)
Cheers,
Jon.
Greg Hennessy
On Sat, 22 Nov 1997 11:48:52 GMT, Mechanoid <dmal...@argonet.co.uk>
wrote:
> It doesn't run RiscOS software though does it?
>
So, Is that supposed to be an advantage ? Given minisucle sales, who
in their right mind is going waste breath developing for it.
> Compared to the current RiscPC, the RiscPC 2 would seem to compete
>strongly. Sales of the current RiscPC have not declined.
>
The RPC II as currently defined will not even come close an Alpha on
integer, FP ? ROFLMAO 'nuff said really.
>
> Multiply that figure by 50 and I might agree.
So who are the mystical buyers of these 50000-75000 £2K risc-os boxes.
Schools ? Nada, By the time this platform is available, £1200 will buy
a PC that will substantially outperform it with a lot more readily
available software. The £800 difference buys a lot of extra toys.
Business ? Not likely.
Halls Of Academe ? One or two to enthusiasts maybe, However for real
work research, simulations etc the alphas FP performance will be the
order of the day.
Niche Markets ? It might help sell a few hundred extra copies of
Sibelius along with a machine to run it on.
Enthusiasts ?, Well £2k is a lot of wonga, 50000 enthusiasts ? More
like 500 methinks.
greg
Ian Lynch
cmk...@cix.compulink.co.uk (Greg Hennessy) wrote:
>> And perhaps that is their plan. So what?
> It speaks volumes about the products life expectancy as a desktop
> platform .
Yeah, but Acorn's main business ain't desktop machines any more so I
for one couldn't care less. I dare say that some desktop machines
will still be produced for several years but apart from enthusiasts
and specialist applications who needs anything more than a StrongArm
or P200 anyway? The desktop computer in its current form is at the end
of the road. The future is in reducing price, integration with
broadcasting and telecommunications not in the million megaHz CPU with
Windows 2020 and 100x CD-ROM. So you just go and play with your
desktop and let us get on with discussing misc. things to do with
Acorn.
--
Ian
John Ward
In article <ant19170...@gromit.tquest.org.uk>,
Barry Wickett <Ba...@tquest.org.uk> wrote this tantalising snippet:
> You wouldn't believe some of the ways we use Acorns here.
Sounds intriguing: try me ;-)
> Anyway, I'm getting off topic a little here!
It might be worth starting a new thread (in the Advocacy NG, perhaps?)
--
<*> John M Ward <*>
"A Shining Beacon in Cyberspace, All Alone in the Night.."
Andy McMullon
In article <ant22215...@isiselec.demon.co.uk>, Jon Harrop
<URL:mailto:j...@isiselec.demon.co.uk> wrote:
> In article <ant22173...@pixeleye.online.no>, Sveinung W. Tengelsen
> <URL:mailto:pixeleye@remove_this.sn.no> wrote:
> > >
> > > Not the speed of the CPUs and the lack of multiple
> > > processors, graphics acceleration, floating point,
> > > etc?
> >
> > Let's just say that certain developments *could* have been brought
> > to the market a bit faster, and then I don't mean the RPC2. I'm
> > lowly becoming a fan of the idea of postponing the RPC2-launch,
> > in order to wait out the availability of certain nice components.
>
> Acorn seem to agree with you about postponing the Risc PC II,
> although I am not convinced it is for the same reason. :-(
>
> > > > totalling a theoretical processing-power (SA core + AMP)
> > >
> > > And what makes you think that Acorn will be using
> > > a processor with an AMP core in the Risc PC II?
> >
> > Logic. It will not be the SA1500 (for reasons explained in this group).
>
> Sure, but there isn't much demand for a desktop version...
>
> Maybe they should give up on the RISC PC II and concentrate
> on the A7000++ with the SA1500 at under 1000 quid. Then
> everyone would he happier, non?
>
> > > > Though, if we get a reasoneably flexible case, one may be allowed to
> > > > upgrade the motherboard sometime late 1998. I hope.
> > >
> > > I don't care about upgrading the motherboard, I just
> > > want a decent spec in the first place.
> >
> > What one wants and what may be attainable aren't always synonymous.
> > If you want the RPC2 *fast* (which I certainly do) this will (probably)
> > be a necessary sacrifice. Not Acorn's fault.
>
> But possibly their downfall...
What worries me is that there is nothing around which needs all this
processing power. I've had my StrongArm more than a year and I am
still doing all the things I did with my lowly Arm 610 and using
exactly the same software. In fact nothing much has changed since I
sold my A5000 in 1994 - apart from Net access!
In other words we need much more than better hardware - we need a
reason to use it!
Andy
--
Andy: skyp...@ANTISPAM.bigfoot.com / http://www.mcfamily.demon.co.uk
I'm nobody! Who are you?
Are you nobody, too?
Then there's a pair of us - don't tell!
They'd banish us, you know.
Sveinung W. Tengelsen
In article <nED4...@ianlynch.demon.co.uk>, Ian Lynch
<URL:mailto:i...@ianlynch.demon.co.uk> wrote:
> cmk...@cix.compulink.co.uk (Greg Hennessy) wrote:
>
> >> And perhaps that is their plan. So what?
>
> > It speaks volumes about the products life expectancy as a desktop
> > platform .
>
> Yeah, but Acorn's main business ain't desktop machines any more so I
> for one couldn't care less. I dare say that some desktop machines
> will still be produced for several years but apart from enthusiasts
> and specialist applications who needs anything more than a StrongArm
> or P200 anyway?
"Who will ever need more than 640k RAM?" 8)
> The desktop computer in its current form is at the end
> of the road. The future is in reducing price, integration with
> broadcasting and telecommunications not in the million megaHz CPU with
> Windows 2020 and 100x CD-ROM.
The lines between technologies are blurring, sure. Most interesting
is the decreasing gap between desktop/portable/embedded.
> So you just go and play with your desktop and let us get on with
> discussing misc. things to do with Acorn.
I've found them to be popular among cats for resting on top of. 8)
Sveinung W. Tengelsen
In article <ant22215...@isiselec.demon.co.uk>, Jon Harrop
<URL:mailto:j...@isiselec.demon.co.uk> wrote:
> In article <ant22173...@pixeleye.online.no>, Sveinung W. Tengelsen
> <URL:mailto:pixeleye@remove_this.sn.no> wrote:
<snip>
> > Let's just say that certain developments *could* have been brought
> > to the market a bit faster, and then I don't mean the RPC2. I'm
> > lowly becoming a fan of the idea of postponing the RPC2-launch,
> > in order to wait out the availability of certain nice components.
>
> Acorn seem to agree with you about postponing the Risc PC II,
> although I am not convinced it is for the same reason. :-(
Um, no, maybe not, but the result could be the same (a more
decently spec'ed RPC2 release 1).
> > > > totalling a theoretical processing-power (SA core + AMP)
> > >
> > > And what makes you think that Acorn will be using
> > > a processor with an AMP core in the Risc PC II?
> >
> > Logic. It will not be the SA1500 (for reasons explained in this group).
>
> Sure, but there isn't much demand for a desktop version...
The RPC sells "quite well". :)
> > Hint: do you *really* believe Acorn would work with Dec on the "new
> > super-chip" (as PB has mentioned umpteen times lately, latest at AW '97)
> > without also doing a "desktop" version? Huh. 8)
>
> I don't believe that Acorn have much of a say in what Intel
> manufactures... 8(
Irrelevant. If Intel don't see fit to earn money by manufacturing
SA-processors, there are others who gladly will take on the job.
> Maybe they should give up on the RISC PC II and concentrate
> on the A7000++ with the SA1500 at under 1000 quid. Then
> everyone would he happier, non?
The SA1500 could power an *AWESOME* portable.
> > > > Though, if we get a reasoneably flexible case, one may be allowed to
> > > > upgrade the motherboard sometime late 1998. I hope.
> > >
> > > I don't care about upgrading the motherboard, I just
> > > want a decent spec in the first place.
> >
> > What one wants and what may be attainable aren't always synonymous.
> > If you want the RPC2 *fast* (which I certainly do) this will (probably)
> > be a necessary sacrifice. Not Acorn's fault.
>
> But possibly their downfall...
Why? Is there something blasphemic to swapping motherboards? ;)
> > > Sure, but it is the 'If/when' that matters.
> >
> > From my point of observation, it looks like a 80/20 chance in Acorn's
> > favour. Can't expound the reasons for this belief, since I'm not
> > allowed to.
>
> Be careful with those cryptic comments, if the wind changes
> you may turn into ... Martin Piper! ;-)
I shall have to take that chance, then. 8)
PS - Success is a question of using money, technologies, partnerships and
marketing in creative/innovative ways. If you look closely at the
constellations of these premises around Acorn, maybe you see what
is brewing. :)
Mark Gillman
In article <ant23083...@mcfamily.demon.co.uk>, Andy McMullon
<URL:mailto:skyp...@bigfoot.com> wrote:
> In other words we need much more than better hardware - we need a
> reason to use it!
It seems to me that the better hardware just makes life easier for
programmers. The last time I felt a computer was /really/ being pushed
to the limits was playing Elite/Revs on a BBC B.
Mind you, playing 9 Replays as a backdrop image and still being able to
work (just) at the 'front' of the window stack came pretty close :-)
--
Mark Gillman
Jon Harrop
> In other words we need much more than better hardware - we need a
> reason to use it!
Programming... :-)
Cheers,
Jon.
Rainer M Duffner
In article <347b504a...@nntp.netcomuk.co.uk>, Greg Hennessy
<URL:mailto:cmk...@cix.compulink.co.uk> wrote:
> On Sat, 22 Nov 1997 11:48:52 GMT, Mechanoid <dmal...@argonet.co.uk>
> The RPC II as currently defined will not even come close an Alpha on
> integer, FP ? ROFLMAO 'nuff said really.
A shame that Acorn never ported RISC-OS to the Alpha architecture.
But then, CC/Xara needed several years just to port Artworks to Win32...
> >
> > Multiply that figure by 50 and I might agree.
>
> So who are the mystical buyers of these 50000-75000 £2K risc-os boxes.
Not impossible, but fairly hefty :-)
> Schools ? Nada, By the time this platform is available, £1200 will buy
> a PC that will substantially outperform it with a lot more readily
> available software. The £800 difference buys a lot of extra toys.
'Decent' PCs will always be expensive. If you just want to run Linux or
FreeBSD, the £1200 maybe enough.
But the WIntel-mafia will work hard to produce software that will only work
on the latest, coolest, X'iest hardware.
A PC that could theoretically do the work I normally do unter RISC-OS in
the same time under Windoze would cost about 2K£, too.
(With monitor, though.)
> Business ? Not likely.
coNCord for them...
> Halls Of Academe ? One or two to enthusiasts maybe, However for real
> work research, simulations etc the alphas FP performance will be the
> order of the day.
If there are still Alphas left then.
By 2000 it may well be IA64. And they will be Intel.
And they will *cost* !
> Niche Markets ? It might help sell a few hundred extra copies of
> Sibelius along with a machine to run it on.
Erm. I don't think that Sibelius opened an extra office in Unlce Sam-Land
for just 200 copies. (Or at least I hope so)
> Enthusiasts ?, Well £2k is a lot of wonga, 50000 enthusiasts ? More
> like 500 methinks.
Yes, 2000 pounds *is* tough.
But then, the platform is worth living after all.
And some people spend even more money on even stranger things:
- old wines that they never drink
- stamps (used stamps with only an imaginary value)
- cars, motorbikes, yachts
But generally, your points are valid. And Acorn knows this, like everybody
knows it. But then, what are they supposed to do ?
Jumping the Windoze-train ? - Who needs yet another company writing
software that is either obsolete or so good that a large, American company
steels it. (Again, Xara -> Corel.......)
Or what about Acorn assembling PCs from imported Asian components ? - Not
really and option either, isn't it !
Just say 'Goodbye fellows, 't was nice' - Well, they can do that anyway.
So, Greg - in your inexhaustable wisdom- share your ideas for Acorn with
us inferior populace.
:-))
Aside:
Has anybody considering buying a RPC2 allready considered that one needs
a good 21" screen to get the most out of the 4MB VRAM ?
This might almost double to cost of the system (w/ EIZO F78), but I'm
seriously thinking of it. Large monitors are sooooo great !
cheers,
Rainer
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|Rainer Duffner, E-Mail: duf...@fh-konstanz.de |
| & Rainer....@konstanz.netsurf.de |
|Fachhochschule Konstanz, Germany |
|"What's a Network ?" - Bill Gates, early 1980s |
| WWW:http://www-stud.fh-konstanz.de/~duffner |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Ernst Dinkla
Rainer M Duffner wrote:
> In article <347b504a...@nntp.netcomuk.co.uk>, Greg Hennessy
> <URL:mailto:cmk...@cix.compulink.co.uk> wrote:
> > On Sat, 22 Nov 1997 11:48:52 GMT, Mechanoid <dmal...@argonet.co.uk>
>
> > The RPC II as currently defined will not even come close an Alpha on
> > integer, FP ? ROFLMAO 'nuff said really.
>
> A shame that Acorn never ported RISC-OS to the Alpha architecture.
> But then, CC/Xara needed several years just to port Artworks to Win32...
Well Galileo will be system independent. (Or does this mean it will
never reach a system? :-)) And the Alpha may survive if the information
on comp.sys.dec isn't too biased.
> > Halls Of Academe ? One or two to enthusiasts maybe, However for real
> > work research, simulations etc the alphas FP performance will be the
> > order of the day.
>
> If there are still Alphas left then.
> By 2000 it may well be IA64. And they will be Intel.
> And they will *cost* !
Samsung and others will make Alpha's in 2000.
> > Enthusiasts ?, Well £2k is a lot of wonga, 50000 enthusiasts ? More
> > like 500 methinks.
>
> Yes, 2000 pounds *is* tough.
> But then, the platform is worth living after all.
> And some people spend even more money on even stranger things:
> - old wines that they never drink
> - stamps (used stamps with only an imaginary value)
> - cars, motorbikes, yachts
What will a BBC of '82 do on a Sotheby auction in 2000?
Ernst
--
Even a blind pig finds an acorn sometimes
Ernst Dinkla Serigrafie,Zeefdruk edi...@inter.nl.net
Thomas Boroske
In message <ant23083...@mcfamily.demon.co.uk> Andy McMullon wrote:
> What worries me is that there is nothing around which needs all this
> processing power. I've had my StrongArm more than a year and I am
> still doing all the things I did with my lowly Arm 610 and using
> exactly the same software. In fact nothing much has changed since I
> sold my A5000 in 1994 - apart from Net access!
>
> In other words we need much more than better hardware - we need a
> reason to use it!
BHP ?
Quake ?
Development (make can't be TOO fast) ?
Viewing Postscript files ?
Running RiscBSD with Emacs ?
Raytracing entire movies (OK, not enough software here, but TopModel and
that other package can certainly use ALL the processing power available) ?
Professional graphics design, with really big images ?
In general, just because your ARM610 system is *useable* for most tasks
does not mean you couldn't do with a faster system. By that logic,
you would only ever upgrade your system when you wanted to play 3D
games.
You can do word processing on an A5000, and that probably won't change in
the next few years. That means that unless you need, say Truecolour, you'll
never upgrade to a faster system.
Since you've got a RPC and that supports enough memory as well as highcolour
the conclusion is that for desktop apps there's really no reason for
you to upgrade ever. Unless developers would intentionally cripple their
apps to only run on later systems. Or you would work/user your system in one
of the areas where a faster system definitely helps (see above).
Kind regards,
--
Thomas Boroske
dgs
In article <ant23184...@duffner.konstanz.netsurf.de>, Rainer M Duffner
<Rainer....@konstanz.netsurf.de> wrote:
> In article <347b504a...@nntp.netcomuk.co.uk>, Greg Hennessy
> <URL:mailto:cmk...@cix.compulink.co.uk> wrote:
>
> > So who are the mystical buyers of these 50000-75000 £2K risc-os boxes.
>
> Not impossible, but fairly hefty :-)
...
> > Well £2k is a lot of wonga, 50000 enthusiasts ? More like 500
> > methinks.
Let's add a note of realism. The original Risc PC sold about
fifty thousand, and is still selling. Fifty thousand for the
successor is perhaps optimistic, five hundred is obviously
ridiculous.
But then, Greg is trolling, after all :)
--
d...@argonet.co.uk - accept no imitations
Acorn user near Manchester? - or - PC card owner?
See MAUG and PC card info pages via : http://www.cybernexus.demon.co.uk/dgs/
"Good spelling is welcomed, what is better is actually having a knowledge
of what one is writing." - Barry Wickett
Stephen Burke
In article <19971124....@y0001006.tu-bs.de>, y000...@ws.rz.tu-bs.de (Thomas Boroske) writes:
> You can do word processing on an A5000, and that probably won't change in
> the next few years. That means that unless you need, say Truecolour, you'll
> never upgrade to a faster system.
I have an A310 upgraded to Arm 3, RO 3 and 4Mb RAM. I will probably buy
a new machine next year - not because I need anything more powerful, but
because the software I'd like to use is not available for RISC OS, and
I will therefore, somewhat reluctantly, be buying a PC. Both Acorn and
the majority of the enthusiasts here seem to think that what matters
is to have the fastest possible machine, but that is not what counts
for most users.
--
e----><----p | Stephen Burke | E-mail: (anti-junk mail version)
H H 1 | Gruppe FH1T (Lancaster) | stephen (dot) burke (at)
H H 11 | DESY, Notkestrasse 85 | desy (dot) de (dot) ignore
HHHHH 1 | 22603 Hamburg, Germany | All junk mail deleted on sight!
H H 1 | "It is also a good rule not to put too much confidence in
H H 11111 | experimental results until they have been confirmed by theory"
Ian Lynch
Ernst Dinkla <edi...@inter.nl.net> wrote:
> What will a BBC of '82 do on a Sotheby auction in 2000?
Not a lot because there will still be hundreds in use in schools ;-)
--
Ian
John Ward
In article <nED4...@ianlynch.demon.co.uk>,
> >> And perhaps that is their plan. So what?
> > It speaks volumes about the products life expectancy as a desktop
> > platform .
> Yeah, but Acorn's main business ain't desktop machines any more so
> I for one couldn't care less. I dare say that some desktop machines
> will still be produced for several years
That still seems likely, even if only as a means of demonstration (from
Acorn's own POV) of their most marketable products.
> The future is in reducing price, integration with broadcasting and
> telecommunications not in the million megaHz CPU with Windows 2020
> and 100x CD-ROM.
Aha! Someone *really* looking towards the future -- I like it! :)
It's so easy to go the predictable route of "more of everything", but
that's a cul-de-sac, too narrow in outlook. The next few years look
much more exciting than recent times have been, but that's precisely
because of the new directions this technology can go and take us along
with it.
Whatever is in the RiscPC-II, it's still going to be just an in-between
product, in a way like the BBC-B+ models were in their day. The real
advances will come later with the launch of Galileo and the integration
of a wide range of compatible devices.
John Ward
In article <ant23111...@gillmen.demon.co.uk>,
Mark Gillman <ma...@gillmen.demon.co.uk> wrote:
[snip]
> Mind you, playing 9 Replays as a backdrop image and still being
> able to work (just) at the 'front' of the window stack came pretty
> close :-)
Isn't all that "background activity" rather distracting, though? :)