News
 

Bookmark and Share

(8) 

On Thursday the National University of Defense Technology (NUDT) in China revealed its improved supercomputer called Tianhe-1A that achieves peak performance of 2.5PetaFLOPS as measured by the LINPACK benchmark. The system is the fastest in China and in the world at the moment. What is noteworthy is that Nvidia Corp.'s Tesla powers the new supercomputer.

Tianhe-1A epitomizes modern heterogeneous computing by coupling massively parallel GPUs with multi-core CPUs, enabling significant achievements in performance, size and power. The system uses 7168 Nvidia Tesla M2050 compute cards and 14 336 central processing units (CPUs); it would require more than 50 000 CPUs and twice as much floor space to deliver the same performance using microprocessors alone.

According to Nvidia, a 2.507PetaFLOPS system built entirely with CPUs would consume more than 12 megawatts (MWs). Thanks to the use of GPUs in a heterogeneous computing environment, Tianhe-1A consumes only 4.04MW, making it 3 times more power efficient.

"The performance and efficiency of Tianhe-1A was simply not possible without GPUs. The scientific research that is now possible with a system of this scale is almost without limits; we could not be more pleased with the results," said Guangming Liu, chief of National Supercomputer Center in Tianjin.

Tianhe-1A was designed by the National University of Defense Technology (NUDT) in China. The system is housed at National Supercomputer Center in Tianjin and is already fully operational. The Tianhe-1A supercomputer will be operated as an open access system to use for large scale scientific computations.

"GPUs are redefining high performance computing. With the Tianhe-1A, GPUs now power two of the top three fastest computers in the world today. These GPU supercomputers are essential tools for scientists looking to turbocharge their rate of discovery," said Jen-Hsun Huang, president and chief executive officer of Nvidia.

Nvidia Tesla C2050 and C2070 computing processor boards are single-chip cards with 3GB and 6GB (respectively) of on-board GDDR5 memory (with ECC enabled, user available memory will be 2.625GB for a C2050 and to 5.25GB for a C2070) with 384-bit interface operating at 1.5GHz. Nvidia Tesla 2000-series deliver up to 515GFLOPS of double-precision peak performance and are based on the GF100/T20 chips with 448 stream processors operating at 1.15GHz.

Nvidia Tesla 2000-series cards and compute modules are now available from all major suppliers of systems for high-performance computing, including Cray, Dell, HP, IBM, SGI and others.

Tags: Nvidia, Fermi, Tesla, GPGPU

Discussion

Comments currently: 8
Discussion started: 10/29/10 05:31:16 PM
Latest comment: 11/02/10 07:43:43 AM
Expand all threads | Collapse all threads

[1-5]

1. 
AMD should play catch up fast ! Besides, its good for nVIDIA that their GPGPU initiative get recognition.
0 0 [Posted by: East17  | Date: 10/29/10 05:31:16 PM]
Reply

2. 
I don't get this yet: Why AMD usin 256-bit inteface instead of 384-bit of NVIDIA's?
That is a huge effect in my opinion, anyway.
0 0 [Posted by: Pouria  | Date: 10/30/10 01:20:47 AM]
Reply
- collapse thread

 
Answer is simple. There is not so much difference between an 256, 384 or even 512 bit Bus, because not even with a 256-bit bus , it's not fully saturated yet. You need a resolution like 3*** x 2*** with 24xSSAA to to really make an increase of Bus worth. Why do you think nvidia has a 48xAA option on their professional cards???
0 0 [Posted by: TAViX  | Date: 10/31/10 02:28:58 AM]
Reply

3. 
The system uses 7168 Nvidia Tesla M2050 compute cards and 14 336 central processing units (CPUs);


Read that out loud! Read it twice, if needed, and then tell me what all is wrong with how this - so called - "news article" is spinning things?

Hint: start with the title
0 0 [Posted by: MyK  | Date: 10/30/10 09:16:24 AM]
Reply
- collapse thread

 
Originally, Tianhe-1 was powered by 4,096 Intel Xeon E5540 processors and 1,024 Intel Xeon E5450 processors, with 5,120 AMD GPUs which were made up of 2,560 dual-GPU ATI Radeon HD 4870 X2 graphics cards.


I think they already had experience with using GPUs so now, they say they have ANOTHER DIFFERENT supercomputer using nVIDIA :


In October 2010, Tianhe-1A, a separate supercomputer, was unveiled at HPC 2010 China. It is now equipped with 14,336 Xeon X5670 processors and 7,168 Nvidia Tesla M2050 general purpose GPUs. 2,048 NUDT FT1000 heterogeneous processors are also installed in the system, but their computing power was not counted into the machine's official Linpack statistics as of October 2010.
0 0 [Posted by: East17  | Date: 11/01/10 05:39:02 AM]
Reply
 
Thanks for the clarification East17

Now, doing a quick Google search would probably yield similar results, and that's fine, but still doesn't change the fact the original article was more than a little misleading in how conveniently it dropped out any info on anything but Tesla. LOL

Can't find much on the NUDT's FT1000 processor, tho. But that's hardly surprising - "New Undetermined Defense Threat"?
0 0 [Posted by: MyK  | Date: 11/01/10 08:02:34 AM]
Reply

4. 
Which CPUs were used? AMD or Intel? There are also twice as many CPUs in this supercomputer than Nvidia Tesla cards and you know they provide alot of the processing power of the SC. Not that these Tesla cards are slouches but my point is that it is a combined effort.

Good to see though, that is massive performance. Do want.
0 0 [Posted by: iLLz  | Date: 10/30/10 11:32:30 AM]
Reply

5. 
In reply to MyK and iLLz, please read the article again:

"The system uses 7168 Nvidia Tesla M2050 compute cards and 14 336 central processing units (CPUs); it would require more than 50 000 CPUs and twice as much floor space to deliver the same performance using microprocessors alone."

The whole point is that the Tesla cards are much more powerful than the CPUs. The above implies that the 7168 M2050 cards have the equivalent processing power to around 36000 CPUs, therefore providing most of the performance.
0 0 [Posted by: Skel  | Date: 11/02/10 07:43:43 AM]
Reply

[1-5]

Add your Comment




Related news

Latest News

Monday, July 28, 2014

6:02 pm | Microsoft’s Mobile Strategy Seem to Fail: Sales of Lumia and Surface Remain Low. Microsoft Still Cannot Make Windows a Popular Mobile Platform

12:11 pm | Intel Core i7-5960X “Haswell-E” De-Lidded: Twelve Cores and Alloy-Based Thermal Interface. Intel Core i7-5960X Uses “Haswell-EP” Die, Promises Good Overclocking Potential

Tuesday, July 22, 2014

10:40 pm | ARM Preps Second-Generation “Artemis” and “Maya” 64-Bit ARMv8-A Offerings. ARM Readies 64-Bit Cores for Non-Traditional Applications

7:38 pm | AMD Vows to Introduce 20nm Products Next Year. AMD’s 20nm APUs, GPUs and Embedded Chips to Arrive in 2015

4:08 am | Microsoft to Unify All Windows Operating Systems for Client PCs. One Windows OS will Power PCs, Tablets and Smartphones