Please visit the sponsor! 11 June 2025 I fondly recall the era when the pinnacle of supercomputing was the Cray 1. Even the shape of this computer was massively different to anything that came before…
 Please visit the sponsor! |
I fondly recall the era when the pinnacle of supercomputing was the Cray 1.
Even the shape of this computer was massively different to anything that came before and it was so futuristic that it could have come straight from a scifi movie.
While almost all other computers of the 1970s were just a collection of huge rectangular cabinets with blinky lights and perhaps a few tape drives, the Cray 1 looked more like a piece of space-age furniture and seemed very small for what it was.
Never the less, the Cray was a pretty hefty bit of kit, weighing in at over 5 tonnes (most of that being the cooling system) and consuming a staggering 115KW of electricity when its 64-bit processor was chugging along at 80MHz and interrogating the 8MB of memory it was attached to.
Those numbers, and the 160MFLOPS of raw performance may have been mind-blowing "back in the day" but when you compare them to what we can buy today you'll likely be even more shocked.
The Cray 1 cost an astounding US$8 million in 1977 which, if you adjust for inflation, is equal to more than US$40 million in today's dollars.
Suffice to say that at that price, only around 100 or so systems were ever sold and they tended to be used for very specific scientific applications rather than as a general purpose computer.
Let's jump forward about half a century to the present day and compare what was once the world's fastest computer to something far more modest -- by today's standards.
The RPi5 is much smaller and lighter than the Cray 1, by many orders of magnitude. In fact we're talking just 50g versus 10 tonnes.
The RPi5 uses far less power at around 12W versus the 115KW of the Cray.
But what about performance? Can this tiny single-board computer match the awesome power of what was once the fastest computer on the planet?
Well, as I mentioned earlier, the Cray had about 160MFLOPS of raw processing power; the Pi has... up to 30GFLOPS. Yes... that's *giga*FLOPS. This makes it almost 200 times faster than the Cray.
So how does the price compare then?
Well that Cray was US$40m in today's currency and an RPi5 will run you just US$120.
You can crunch your own numbers to calculate interesting things like the "bang per buck" ratio of these two machines but it's very, very apparent through this comparison, just how rapidly our computer technology has advanced over recent decades.
As someone who spent countless hours tinkering with early 8-bit microprocessors back in the 1970s (when the Cray was king), it still astonishes me just how much raw processing power, memory and storage we now consider to be "average".
That we can fit a terabyte onto a microSD card smaller than your thumbnail also boggles my mind -- especially when I remember all those years ago that a regular 16 pin DIP integrated circuit was doing well to store just a single Kbyte of data.
I find it hard to believe that this rate of progress will continue unchecked for the next 50 years -- but then again if you'd showed me an RPi5 back in 1977 I would have said "nah, impossible" so who knows?
If AI systems continue to improve at the current rate and we combine that with improvements in hardware that are measured in orders of magnitude every 15 years or so then it stands to reason that we'll get that "super-intelligent GAI" system any day now.
Once that happens I wonder if we'll be relegated to the role of "curious pets" by that system?
Carpe Diem folks!
|
Please visit the sponsor! |
Here is a PERMANENT link to this column
Read the original article
Comments
My former boss (Steve Parker, RIP) shared a story of Turner Whitted making predictions about how much compute would be needed to achieve real-time ray tracing, some time around when his seminal paper was published (~1980). As the story goes, Turner went through some calculations and came to the conclusion that it’d take 1 Cray per pixel. Because of the space each Cray takes, they’d be too far apart and he thought they wouldn’t be able to link it to a monitor and get the results in real time, so instead you’d probably have to put the array of Crays in the desert, each one attached to an RGB light, and fly over it in an airplane to see the image.
Another comparison that is equally astonishing to the RPi is that modern GPUs have exceeded Whitted’s prediction. Turner’s paper used 640x480 images. At that resolution, extrapolating the 160 Mflops number, 1 Cray per pixel would be 49 Tera flops. A 4080 GPU has just shy of 50 Tflops peak performance, so it has surpassed what Turner thought we’d need.
Think about that - not just faster than a Cray for a lot less money, but one cheap consumer device is faster than 300,000 Crays.(!) Faster than a whole Cray per pixel. We really have come a long, long way.
The 5090 has over 300 Tflops of ray tracing perf, and the Tensor cores are now in the Petaflops range (with lower precision math), so we’re now exceeding the compute needed for 1 Cray per pixel at 1080p. 1 GPU faster than 2M Crays. Mind blowing.
By msgodel 2025-06-1620:05 Well that's the way parallelism goes.
By magicalhippo 2025-06-1415:411 reply > 1 Cray per pixel would be 49 Tera flops. A 4080 GPU has just shy of 50 Tflops peak performance
Interesting, wonder how it compares in terms of transistors. How many transistors combined did one Cray have in compute and cache chips?
By dahart 2025-06-153:49 The Wikipedia article says the Cray-1 has 200k gates. I assume that would mean something slightly north of 2x the number of transistors? https://en.wikipedia.org/wiki/Cray-1#Description
200k * 300k Cray-1s would be 60B gates, whereas the 4080 actually has 46B transistors. Seems like we’re totally in the right ballpark.
By phendrenad2 2025-06-160:03 Whitted mentioned! Cofounder of the first 3d game engine company.
But the Cray had a general purpose CPU while the GPUs have specialized hardware. Not exactly apples to apples.
By monocasa 2025-06-1514:54 The main part of the Cray was a compute offload engine that asynchronously executed job lists submitted by front end general purpose computers that ran OSes like Unix.
It was actually pretty close to the model of a GPU.
Back in 2020, someone built a working model of a Cray-1.[1] Not only is it instruction compatible, using an FPGA, it's built into a 1/10 scale case that looks like a Cray-1.
The Cray-1 is really a very simple machine, with a small instruction set. It just has 64 of everything. It was built from discrete components, almost the last CPU built that way.
[1] https://www.cpushack.com/2010/09/15/homebrew-cray-1a-1976-vs...
By Animats 2025-06-1619:18 (2010, actually.)
By qingcharles 2025-06-158:50 In 2013 I'd just built a new top-spec PC. I looked up the performance and then back-calculated using the TOP500† and I believe it would have been the most powerful supercomputer in the world in about 1993. If you back-calculated further, I think around 1980 it became more powerful than every computer on the planet combined.
