NVIDIA Announces Tesla Personal Supercomputer

http://www.hpcwire.com/offthewire/NVIDIA_Announces_Tesla_Personal_Supercomputer.html

by jaydub • 18 years ago

21 28 18 years ago

jm4• 18 years ago

This sounds cool, but the page linked to is only a standard press release and the link to the interesting stuff is buried in the marketing-speak. The fun stuff is here: http://www.nvidia.com/object/personal_computing.html

bootload• 18 years ago

I took a look at the Tesla after viewing an Nvidia demo with the Mythbusters Adam & Jamie doing a simple demo of CPU v's GPU which you can see here ~ http://www.youtube.com/watch?v=fKK933KK6Gg

Firstly you can run the thing as either a card (cheaper, slower) or standalone machine (expensive, nobody lists the price). The CUDA toolkit which is C based is Win/Lin 32/64 bit compatible and available for most mainstream disto's ~ http://www.nvidia.com/object/cuda_learn.html

"... The Tesla architecture is built around a scalable array of multithreaded Streaming Multiprocessors (SMs) ..."

If you think this is a similar to mainstream development think again. Programming with this machine is right down to the metal and reminds me of programming the PS2 and other specialised consoles. You need to spend some time reading the Hardware manuals, understand the architecture to use the machine to it's capability.

SirWart• 18 years ago

I had some friends who tried to port the Linpack benchmark to a small cluster of computers each with 8 NVIDIA GPUs using CUDA, and they found that the biggest bottleneck was the bandwidth to and from the GPUs. It's just hard to keep the GPUs fed with enough data. They confirmed that they are both hard to program and incredibly powerful.

DarkShikari4• 18 years ago

I tried CUDA programming as well and found that the threading model and such is an utter nightmare. For certain tasks its quite easy, such as upscaling or filtering an image. However, such a task is exactly the kind of task where you'll end up bandwidth-limited anyways. The kind of complex tasks where you actually fully use the GPU processors are often the exact kind of situations where the API will work against you every step of the way.

The "960 cores" moniker is also very misleading, as last I recall there were really only (960/8) cores, with each core being able to run 8 instructions at the same if all the instructions were exactly the same.

brfox• 18 years ago

Protein sequence alignment implemented (open source):

http://www.biomedcentral.com/1471-2105/9/S2/S10

michaelneale2• 18 years ago

you mention ps2 - is the ps3 similar? (although its powerpc - there are 7 cores to use - so is it still as tricky?)

jaydub• 18 years ago

the PS3 uses the Cell processor which is also tricky to program

DarkShikari4• 18 years ago

The Cell is a whole different type of nightmare, with various interesting properties:

1. Bitshifts by variable amounts take ~7 clocks each on the main PPU.

2. The SPUs have no cache whatsoever; all caching has to be done explicitly, and all memory access is DMA'd.

3. The SPUs do scalar math no faster than SIMD (in fact, from what I know, scalar math is just calling an SIMD function on a single value).

4. There is no instruction reordering on SPUs, and everything has to be synced exactly for max performance (certain instructions run on odd clock cycles, others on even cycles).

5. The SPUs are not Altivec chips. The SPU SIMD instruction set is similar to Altivec, but more versatile. Of course, this means you can't just run existing Altivec code on them.

6. Overall, the integer SIMD on SPUs is much slower than that on modern Intel processors. Not sure about float, as I have no experience in that arena.

tsally• 18 years ago

I guess I am a little skeptical about the market. First, I have to believe it is rare that an individual would need a super computer. Besides, any individual that actually needs one is probably going to build his/her own. Businesses might want these on a large scale, but then why market it as a personal computer?

wmf4• 18 years ago

Lots of scientists, engineers, and finance people could use these. It's personal in the sense that it's used by one person, not that people would buy it for personal use.

Also, it doesn't really matter whether you build or buy; either way all the cost is in the Tesla card and NVidia gets their money.

LPTS• 18 years ago

"They'll never sell more than 5 of those pc computers"

jcromartie• 18 years ago

960 cores. Wow. I thought that I was in over my head trying to program 2 at a time!

jodrellblank• 18 years ago

What was it Joel Spolsky said about programmers counting "It must work for 1. Oh, there's more than 1? Then it must work for any number".

;)

riobard• 18 years ago

Reading the title, I initially thought it was something related to the electric car ...

Anyway, I guess the problem is that most programmers have no clue of how to program GPU. Programming multi-core CPU is already very hard, and now comes the GPU stuff ...

noonespecial• 18 years ago

I like Tesla. He was under-appreciated in his time, brilliant, and did things just for the joy of the science.

I just wish they'd stop with the naming of products, rock bands, and breakfast cereals, etc. after him. Its starting to wear thin.

light3• 18 years ago

It'll be interesting to see how much these will cost. 4 Teraflops of single precision calculation is very impressive, however if you need double precision, the speed is not quite as impressive at 400 Gflops - although still very good compared to my laptop with 20 Gflops :)

The main drawback with these is that you have to use CUDA, which certainly takes a while before you wrap your mind across. I played CUDA for a while but considered it is too much effort for something which is very specialised and might not become mainstream. Still there seems to be many people using CUDA and with lots of research roles - see the CUDA forums.

kahseng• 18 years ago

It says "Available from VARs worldwide for under $10,000" in here http://www.nvidia.com/object/personal_computing.html

Now I know how it felt like in the '80s when people were looking at the mainframes/desktops of the time and wondering... can I afford this $10K machine? :)

NVIDIA Announces Tesla Personal Supercomputer

Related Stories