NVIDIA CUDA Technology Dramatically Advances The Pace Of Scientific Research


Distributed Computing Applications use NVIDIA GPUs for Biomedical Research, Space Exploration and Searching for Extra Terrestrial Intelligence

For further information, contact:

Brian Burke
NVIDIA Corporation
(512) 401-4385
bburke@nvidia.com


Using NVIDIA CUDA Technology, a GeForce GTX 280 GPU runs SETI@home is nearly 10 times faster than a AMD Phenom 9950 multi-core consumer CPU.


“NVIDIA CUDA technology opens up processing power for scientific research that was previously unavailable and impossible for researchers to afford,”
Dr. David Anderson, Research Scientist U.C. Berkeley Space Sciences Laboratory and founder of BOINC


"Running GPUGRID on NVIDIA GPUs innovates volunteer computing by delivering supercomputing class applications on a cost effective infrastructure which will greatly impact the way biomedical research is performed."
Dr. Gianni De Fabritiis, researcher at the Research Unit on Biomedical Informatics at the Municipal Institute of Medical Research and Pompeu Fabra University in Barcelona
FOR IMMEDIATE RELEASE:

SANTA CLARA, CA—December 17, 2008— Once thought of as a technology used only for computer games, NVIDIA® GeForce® graphics processing units (GPUs) with CUDA™ technology are now being used for the serious business of scientific computation. Berkeley’s Open Infrastructure for Network Computing (BOINC), one of the leading distributed computing platforms in the world, is using CUDA technology to tap the massively parallel processing power of NVIDIA GPUs with astounding results that could change the pace of scientific discovery through projects like GPUGRID and Einstein@home. The latest breakthrough came with the release of an optimized client that will allow SETI@home to analyze SETI (Search for Extraterrestrial Intelligence) data in about one-tenth of the time it previously took using CPUsi.

“NVIDIA CUDA technology opens up processing power for scientific research that was previously unavailable and impossible for researchers to afford,” said Dr. David Anderson, Research Scientist U.C. Berkeley Space Sciences Laboratory and founder of BOINC. “CUDA technology makes it easy for scientists and researchers to optimize BOINC projects for NVIDIA GPUs and they are already using it for applications in molecular dynamics, protein structure prediction, climate and weather modeling, medical imaging, and many other areas.”

BOINC is a unique approach to supercomputing in which multiple consumer computers are joined together over the Internet and their combined computing power is used to tackle very large computational tasks. BOINC provides the distributed computing grid layer for a wide variety of scientific projects that work to help cure diseases, study global warming, discover pulsars, and do many other types of scientific research on home PCs.

SETI@Home
Researchers in the scientific field of SETI received a massive increase in computing power today, when NVIDIA and BOINC released an optimized client that will allow SETI@home to be accelerated on GeForce GPUs. SETI@home, the largest BOINC project with nearly 200,000 active users, searches for extra terrestrial intelligence by using radio telescopes to listen for narrow-bandwidth radio signals from space. The performance of a GeForce GTX 280 GPU running SETI@ is nearly 2 times faster than the fastest consumer multicore CPU (3.2GHz Intel Core i7 965) and almost 10 times faster than an average dual core consumer CPU (2.66 GHz AMD Phenom 9950)ii.

GPUGRID
GPUGRID, the first BOINC project to use NVIDIA GeForce GPUs with CUDA technology for computing, utilizes NVIDIA-based graphics cards in participating PCs to compute high-performance biomolecular simulations for scientific research. Adding support for NVIDIA GPUs resulted in 1000 active GPUs delivering the same amount of computing power as 20,000 CPUs in similar projects, delivering an average speed-up of 20 times.

"The molecular simulations performed by our volunteer computing project are some of the most common types performed by scientists, but they are also some of the most computationally demanding and usually require a supercomputer," stated Dr. Gianni De Fabritiis, researcher at the Research Unit on Biomedical Informatics at the Municipal Institute of Medical Research and Pompeu Fabra University in Barcelona. "Running GPUGRID on NVIDIA GPUs innovates volunteer computing by delivering supercomputing class applications on a cost effective infrastructure which will greatly impact the way biomedical research is performed."

Einstein@Home
NVIDA CUDA technology will soon be powering the third most widely used BOINC project, Einstein@Home, which uses distributing computing to search for spinning neutron stars (also called pulsars) using data from gravitational wave detectors.

“We expect that porting Einstein@Home to GPUs will increase the throughput of our computing by an order of magnitude,” said Bruce Allen, director of the Max Plank Institute for Gravitational Physics and Einstein@Home Leader for the LIGO Scientific Collaboration. “This would permit deeper and more sensitive searches for continuous-wave sources of gravitational waves.”

“Parallel processing is the key to enabling visual computing, whether in the home, office or research lab, and the CUDA-accelerated GPU is the leading engine behind this trend. Distributed computing is an ideal application for parallel processing, so it’s no surprise that these amazing applications are taking advantage of the GPU’s unprecedented computational power” said Michael Steele, General Manager of Visual Consumer Solutions at NVIDIA. “NVIDIA GPUs are transforming the way we work, play, live and discover.”

To download the NVIDIA SETI@home client visit http://setiathome.berkeley.edu/cuda.php. For more information on BOINC visit http://boinc.berkeley.edu/. For more information on the Einstein@Home visit http://einstein.phys.uwm.edu. For more information on GPUGRID visit http://www.gpugrid.net/.

About NVIDIA
NVIDIA (Nasdaq: NVDA) is the world leader in visual computing technologies and the inventor of the GPU, a high-performance processor which generates breathtaking, interactive graphics on workstations, personal computers, game consoles, and mobile devices. NVIDIA serves the entertainment and consumer market with its GeForce graphics products, the professional design and visualization market with its Quadro graphics products, and the high-performance computing market with its Tesla™ computing solutions products. NVIDIA is headquartered in Santa Clara, Calif. and has offices throughout Asia, Europe, and the Americas. For more information, visit www.nvidia.com.

Certain statements in this press release including, but not limited to, statements as to the benefits, impact, performance, power and capabilities of NVIDIA GeForce GPUs with CUDA technology; the partnership between NVIDIA and BOINC and its projects; and the impact of BOINC, SETI@home, GPUGRID and Einstein@Home in their respective fields of study are forward-looking statements that are subject to risks and uncertainties that could cause results to be materially different than expectations. Important factors that could cause actual results to differ materially include: development of faster or more efficient technology; the impact of technological development and competition; design, manufacturing or software defects; changes in consumer preferences or demands; changes in industry standards and interfaces; unexpected loss of performance of our products or technologies when integrated into systems; as well as other factors detailed from time to time in the reports NVIDIA files with the Securities and Exchange Commission including its Form 10-Q for the fiscal period ended October 26, 2008. Copies of reports filed with the SEC are posted on our website and are available from NVIDIA without charge. These forward-looking statements are not guarantees of future performance and speak only as of the date hereof, and, except as required by law, NVIDIA disclaims any obligation to update these forward-looking statements to reflect future events or circumstances.



###

© 2008 NVIDIA Corporation. All rights reserved. NVIDIA, the NVIDIA logo, GeForce, Tesla, and Quadro are trademarks or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated. Features, pricing, availability, and specifications are subject to change without notice.

Note to editors: If you are interested in viewing additional information on NVIDIA, please visit the NVIDIA Press Room at http://www.nvidia.com/page/press_room.html

___________________________
iBased on a consistent and reproducible SETI@home workload. Time-to-compute is measured and lower time is better. NVIDIA® GeForce® GTX 280-based system processes workload on the NVIDIA GPU and is based on an NVIDIA nForce® 780i SLI™-based motherboard, NVIDIA GTX 280 GPU, Intel Core i7 965 CPU, 2GB DDR2 DRAM and processes the workload in 391 seconds. “Fastest consumer multicore CPU-based system” processes the entire workload on CPU and is based on an ATI Radeon HD4870 GPU, Intel x58-based motherboard, Intel Core i7 965, 3GB DDR3 DRAM and processes the workload in 670 seconds. “Average dual core CPU-based system” processes the entire workload on CPU and is based on an ATI Radeon HD4870 GPU, AMD Phenom 9950 CPU (Dual Core 2.66GHz) 2GB DDR2 DRAM and processes the workload in 3,777 seconds

ii Based on a consistent and reproducible SETI@home workload. Time-to-compute is measured and lower time is better. NVIDIA® GeForce® GTX 280-based system processes workload on the NVIDIA GPU and is based on an NVIDIA nForce® 780i SLI™-based motherboard, NVIDIA GTX 280 GPU, Intel Core i7 965 CPU, 2GB DDR2 DRAM and processes the workload in 391 seconds. “Fastest consumer multicore CPU-based system” processes the entire workload on CPU and is based on an ATI Radeon HD4870 GPU, Intel x58-based motherboard, Intel Core i7 965, 3GB DDR3 DRAM and processes the workload in 670 seconds. “Average dual core CPU-based system” processes the entire workload on CPU and is based on an ATI Radeon HD4870 GPU, AMD Phenom 9950 CPU (Dual Core 2.66GHz) 2GB DDR2 DRAM and processes the workload in 3,777 seconds

Out of the beta site and on to the main Seti site for the nvidia cards. I have run a few hundred workunits without a problem. They have only released the short 1 hour cpu units on the main site for the gpu's. They take about 6 min to run on a gtx 260.

Trying to test some Einstein units. Maybe Rosetta will beat em to the punch.

As you may know, several projects now use GPU application ! It is a really good thing ! So if you have graphic cards you might be interested in !

There are :
- Milkyway@home, Collatz@home for ATI. (Milkyway accept only ATI HD 38XX 48XX 47XX 57XX 58XX graphic cards, but collatz accept lower end GPU) This can be used under Windows. Not for Linux unfortunately.
- Seti@home, GpuGrid@home, Milkyway (very recent), Collatz@home (very recent too) and Einstein (In test) for Nvidia. Every cards over 8xxx are good. But if you want to use your computer in the same time, it is better to use these projects with a card higher than 8600.
These projects can be used on Windows AND Linux ! (Yahouu!!!)

You can have more information on the respective website or even by asking here xD

Recently, Primegrid has made a Nvidia app for Linux 64b relating to their AP26 research.
Into 1 month they want to improve it and built an app for windows as well.

I have a [GeForce 8600 GTS] 6.6.20][CUDA (1) driver: 255MB. I need to use that computer, it's not an idle old box.

I guess I can't do anything good with that, right?

Actually you can !
Even if it is a "low end" GC, you can still run some projects on it.
When I've tried my 8600GTS, it was as powerful as a 1.8Ghz dualcore. (So it worthy I guess)

But you gonna have some problem with your display...it will be quite slow...
So I suggest you to try...but may be, use your card only when you don't use the computer...

So, you should try SETI@home.
And also Collatz (Not sure if 255mb is enough with this project, but you will get a higher RAC with it. It is the best projet for NVIDIA card)

If you need any help, do not hesitate to contact me.

Well actually my computer runs at 3.00 GHZ (dual-core) so I guess I'd have to stop running Boinc on the CPU if I wanna run it on my GPU, so it wouldn't be worth it!

No no, you can do both.
Seti or collatz will use less than 5% of your CPU.
I meant , with a 8600GTS running, the graphic interface will be quite slow (I guess), so when you will use your computer, the interface will be slow, but during idle time it's perfect !

So you should have a try !

I will try that right tomorrow, when I'll leave for some hours I'll start it, and I will stop it when I get back. Let's see if that helps! Thanks for the advice!

Update: about 6 hours for one task on SETI, I don't know how many credits this is, but it's over 0 I'm pretty sure Nice addition to the crunching process, not to mention I never crunched SETI before.

Great !!
And can you use your computer in the same time ?
If not, you should uncheck the option "Use GPU while computer is in use"

Bon crunch ! :p

Oh well, Actually, I can use my cpu I think. I mean, nothing seems to be that wrong, maybe a little bit more laggy.

If I have a video to watch, though, I suspend SETI to make sure everything's allright.

So, that's perfect !
You gonna increase your RAC with that card !!

Only 90 credits per task (7 hours long).

Still, it's free credit, so I don't complain.

Sounds about right for 32 shader cores. I have a 9600 gt(64 shaders) that does a seti unit in around 3 hours and a gtx280(240 shaders) that does one in around 15 min. Good to see more people using the gpu stuff.

Someone should get santa to bring us all a ati 5890(1600 shaders).

Now, you don't need anymore to worry about the Einstein app.
They finally automatically deliver CUDA work & applications.
But you need to respect these points :
- enabled NVIDIA GPU work in Einstein@home preferences
- NVidia GPU with at least 450MB of free memory
- Display Driver version 190.38 (&up), i.e. CUDA 2.3 capability
- BOINC Core Client version 6.10 (&up)

Furthermore, the app still use a full core.

Ya I just did a few, takes about 4 hours of 1 cpu and gpu to do one task, it's not very fast at all. Looking at the gpu heat, and throttle the gpu is just above idle. I think they may be doing what gpugrid did at the start in using the cpu too much not allowing the gpu to throttle up to speed.

I might wait a while till they get it working a bit better. It is good to see more projects jumping in to the gpu stuff.

Yeah! you're right.
They are going to improve their app.
For now, it used both CPU and GPU. unfortunatly, the GPU is used only for a small part of the calcul...(sigh)

A PrimeGrid CUDA23 application has been released for 32 bit Windows (will run on 64 bit) and for 64 bit Linux !
"Sieving/primality testing strikes at the GPU weaknesses. Therefore, the application, while the fastest at PrimeGrid, does not compare to some of the other projects where the code is better suited for the GPU's. As such, credit will be much lower than what you might experience at other projects. Nevertheless, you now have one more "device" to crunch on at PrimeGrid. "

Enjoy it !!

On December 18th (in one week), there's the winter solstice challenge on PrimeGrid. The application is AP26.

I'll give it a try then, let's see if my old GPU could do some PrimeGrid!

I'm still crunching SETI while I'm not at home, approaching 10 000 credits, love it! I'll let you know if I catch a Martian.

After 10y or so running it, id settle for a UotD... for now anyways

I don't think it will works on a 8600GT. You need at least a 8800GTS 512 (320 and 640mo won't works neither)

Oh well, I thought so...

I'll still run AP26 for the challenge, I kinda like PrimeGrid! I hope Canada will participate this time (finished 23rd in the last challenge...)

Nice rank indeed !
AP26 works quite well with 64b OS !
PSP and PPS works quite well too !

But for AMD proc, the best is AP26 (For the highest RAC)

Did I mention it was possible to crunch with ATI on Seti ? (Only a part of the calculation can be done by the GPU, then as Einstein, a full core will be allocated to the GPU. But it can be really nice for Seti's and ATI fans !)

And about GpuGrid, there are developping an ATI app too.
But it will certainly be available only for 58xx or 59xx cards.

Yum GPU !!!!

I've tried Primegrid with my GTX275 running on Gentoo 64b.
It takes me a bit more than 500s per unit. I can assume a rac of 6k/day with it. (Which is quite low comparing to my 50k/d on Collatz)

Anyway I'm not a fan of Collatz, especially since they added a bottom "JOIN S.USA" into BOINC Manager !

Btw, you don't need to install any app for Primegrid. It is, as GPUGrid, automatically downloaded.

Edit : You don't need neither to install the app for Milkyway with Nvidia card !

I just did some add on's to a couple of computers. Added liquid cooling to the cpu's and gpu's on 2 machines, and started playing with the overclocking ability's on em both, wow this overclocking is very cool. I was just too scared to do it before with the forced air fan coolers, as some projects really heated up the boxes. I looked around the internet for a few day's pricing it out and eventually picked a place down south that made all their own rad's, and heat plates on a cnc machine, and were about 4X less money for the whole set up ($100 us per machine). The last couple of days I have been running them just about full out on the cpu's and gpu's and the temp's are steady around 60c, and before underclocked they were around 80c. I will try some Primegrid later today, as that was one of the ones that heated things up. I have retired a couple of computers, and am only running 4 boxes, but the RAC is way better now.