Jim,
I totally agree with you, todays GPUs are screamers. Did you know that the
core of NVIDIA's design team came from SGI?
Sad to see that company went wintel... I think they were almost de-listed
from the stock excchage at one point, as the stock
dipped below a dollar.
Jim, have you followed the recent threads in the news, to use the GPU
processor fabric as a general purpose engine, not just for pixels? If you
have IEEE access, search GPU.
Due to the PC gamers, we really do have a supercomputer on a card for 40
bucks or less. If anyone else is exploring this topic, please reply or
email me rdawson16 at hotmail dot com.
I think the strenght of the Stardent system was a combination of things from
Gordon More and Seymour Cray, their rules for supercomputer design. The
memory bandwidth was awsome for the time, and with 4 processors and a vector
unit accessing it independently, crunching code from a compiler that knew
how to vectorize and divide tasks.
I had a chance to represent another company (I was a graphics consultant to
NASA) Superset- they had a cool idea too, to process Fortran in near native
instruction of the machine. They used bit slice 2901s from AMD to create a
A (operation) B = C machine, and a compiler to generate this code. So the
machine was sort of a hardware interpreter. If you look at todays DSPs they
are very similar, they can read an operand, write an operand and perform a
computation in a single cycle.
I have Stardent (Kubota) Dore' running today and porting my old app to it.
I wrote the PC roller coaster simulator, COASTER. I want to re-spin it, now
that fast GPU hardware has finaly arrived and in every PC.
My rose colored glasses recall all of these machines booting to prompt and a
screen in seconds. They always ran for months without a crash, and could
read/write 500 mbyte images/data files in seconds too. The PC aint there
yet...
I write this from FreeBSD, I took all the 'Gates' out of my computer, and
guess what, it still works.
Thanks for your coments, Jim.
Randy
From: Jim Battle <frustum at pacbell.net>
Reply-To: "General Discussion: On-Topic and Off-Topic
Posts"<cctalk at classiccmp.org>
To: General Discussion: On-Topic and Off-Topic Posts
<cctalk at classiccmp.org>
Subject: Re: Modern Marvels: Computers ?? no graphics supers
Date: Sat, 10 Mar 2007 13:44:55 -0600
Randy Dawson wrote:
...
Lost, sadly was the machine between then and now,
the Graphics
Supercomputer. In an effort to add computational speed to graphics and
scientific visualazation, two vendors went head to head on this problem,
Ardent and Stellar.
If you were around at the time, and saw one of these I would love to hear
from you. The performance was truly spectacular. I had a chance to use
one for a couple of years, and it still comes pretty close to current GPU
tec in graphics performance. With pipeline vector processor and compiler
to unroll loops it was WOW. Todays Ghz processors cannot beat a vector
machine in computation, Titan had a 16 Mhz 1K floating point vector ALU.
Randy, no doubt you know a lot more about the ardent/stellar/stardent stuff
than me. I was aware of it and I once got hold of the design spec for the
TOE processor (the 4x4 pixel "stamper"). However, I think you aren't aware
of how sophisticated todays GPUs are.
16 MHz * 1K flops = 16 Gflops. A single top end GPU is more like 500
GFLOPS (single prec only, though). Today's GPUs have myriad pixel formats,
including ARGB with an FP32 for each component. Pixel shaders are highly
programmable. A single GPU can have > 80 GB/sec of bandwidth to DRAM (not
cache).
The TOE processor was a fixed point affair with limited, fixed point
precision. There is no comparison. I wish I still had the spec to make a
more concrete comparison.
A google search turned up this quote:
With the Dore' rendering package [Borden89], each processor is capable of
rendering a maximum of 20,000 smoothly shaded small polygons/seconds.
Today's GPUs can render thousands of times more triangle per second,
antialiased, with multiple, high quality texture maps and arbitrary
blending.
Another google search
http://www.ece.cmu.edu/~ece548/handouts/17v_perf.pdf
says that the Titan 1 had a 125 ns clock period and two FPUs, for 16
MFLOP/s peak. Perhaps you recall 1K FPUs, but maybe it was a 1K vector
register length. The same pdf (written by Philip Koopman) says that even
with four processor, and with a large (1000x1000) array size, the titan-1
peaked at 15.7 Mflops. It attributes this to the fact that the aggregate
bus bandwidth of the titan was 256 MB/sec. By rewriting the linpack code
to block the data appropriately, they got it up to 46 MFLOP/s.
So, overall, I think there is no comparison. The rose colored glasses of
time have fooled you.
...
Are there any graphics guys on the list?
Yes, from the hw end of things.
_________________________________________________________________
The average US Credit Score is 675. The cost to see yours: $0 by Experian.
http://www.freecreditreport.com/pm/default.aspx?sc=660600&bcd=EMAILFOOT…