Tag Archives: NVIDIA A100

NVIDIA Offers A800 GPU To Bypass US Ban On China!

How NVIDIA A800 Bypasses US Chip Ban On China!

Find out how NVIDIA created the new A800 GPU to bypass the US ban on sale of advanced chips to China!

 

NVIDIA Offers A800 GPU To Bypass US Ban On China!

Two months after it was banned by the US government from selling high-performance AI chips to China, NVIDIA introduced a new A800 GPU designed to bypass those restrictions.

The new NVIDIA A800 is based on the same Ampere microarchitecture as the A100, which was used as the performance baseline by the US government.

Despite its numerically larger model number (the lucky number 8 was probably picked to appeal to the Chinese), this is a detuned part, with slightly reduced performance to meet export control limitations.

The NVIDIA A800 GPU, which went into production in Q3, is another alternative product to the NVIDIA A100 GPU for customers in China.

The A800 meets the U.S. government’s clear test for reduced export control and cannot be programmed to exceed it.

NVIDIA is probably hoping that the slightly slower NVIDIA A800 GPU will allow it to continue supplying China with A100-level chips that are used to power supercomputers and high-performance datacenters for artificial intelligence applications.

As I will show you in the next section, except in very high-end applications, there won’t be truly significant performance difference between the A800 and the A100. So NVIDIA customers who want or need the A100 will have no issue opting for the A800 instead.

However, this can only be a stopgap fix, as NVIDIA is stuck selling A100-level chips to China until and unless the US government changes its mind.

Read more : AMD, NVIDIA Banned From Selling AI Chips To China!

 

How Fast Is The NVIDIA A800 GPU?

The US government considers the NVIDIA A100 as the performance baseline for its export control restrictions on China.

Any chip equal or faster to that Ampere-based chip, which was launched on May 14, 2020, is forbidden to be sold or exported to China. But as they say, the devil is in the details.

The US government didn’t specify just how much slower chips must be, to qualify for export to China. So NVIDIA could technically get away by slightly detuning the A100, while offering almost the same performance level.

And that was what NVIDIA did with the A800 – it is basically the A100 with a 33% slower NVLink interconnect speed. NVIDIA also limited the maximum number of GPUs supported in a single server to 8.

That only slightly reduces the performance of A800 servers, compare to A100 servers, while offering the same amount of GPU compute performance. Most users will not notice the difference.

The only significant impediment is on the very high-end – Chinese companies are now restricted to a maximum of eight GPUs per server, instead of up to sixteen.

To show you what I mean, I dug into the A800 specifications, and compared them to the A100 below:

NVIDIA A100 vs A800 : 80GB PCIe Version

Specifications A100
80GB PCIe
A800
80GB PCIe
FP64 9.7 TFLOPS
FP64 Tensor Core 19.5 TFLOPS
FP32 19.5 TFLOPS
Tensor Float 32 156 TFLOPS
BFLOAT 16 Tensor Core 312 TFLOPS
FP16 Tensor Core 312 TFLOPS
INT8 Tensor Core 624 TOPS
GPU Memory 80 GB HBM2
GPU Memory Bandwifth 1,935 GB/s
TDP 300 W
Multi-Instance GPU Up to 7 MIGs @ 10 GB
Interconnect NVLink : 600 GB/s
PCIe Gen4 : 64 GB/s
NVLink : 400 GB/s
PCIe Gen4 : 64 GB/s
Server Options 1-8 GPUs

NVIDIA A100 vs A800 : 80GB SXM Version

Specifications A100
80GB SXM
A800
80GB SXM
FP64 9.7 TFLOPS
FP64 Tensor Core 19.5 TFLOPS
FP32 19.5 TFLOPS
Tensor Float 32 156 TFLOPS
BFLOAT 16 Tensor Core 312 TFLOPS
FP16 Tensor Core 312 TFLOPS
INT8 Tensor Core 624 TOPS
GPU Memory 80 GB HBM2
GPU Memory Bandwifth 2,039 GB/s
TDP 400 W
Multi-Instance GPU Up to 7 MIGs @ 10 GB
Interconnect NVLink : 600 GB/s
PCIe Gen4 : 64 GB/s
NVLink : 400 GB/s
PCIe Gen4 : 64 GB/s
Server Options 4/ 8 / 16 GPUs 4 / 8 GPUs

NVIDIA A100 vs A800 : 40GB PCIe Version

Specifications A100
40GB PCIe
A800
40GB PCIe
FP64 9.7 TFLOPS
FP64 Tensor Core 19.5 TFLOPS
FP32 19.5 TFLOPS
Tensor Float 32 156 TFLOPS
BFLOAT 16 Tensor Core 312 TFLOPS
FP16 Tensor Core 312 TFLOPS
INT8 Tensor Core 624 TOPS
GPU Memory 40 GB HBM2
GPU Memory Bandwifth 1,555 GB/s
TDP 250 W
Multi-Instance GPU Up to 7 MIGs @ 10 GB
Interconnect NVLink : 600 GB/s
PCIe Gen4 : 64 GB/s
NVLink : 400 GB/s
PCIe Gen4 : 64 GB/s
Server Options 1-8 GPUs

 

Please Support My Work!

Support my work through a bank transfer /  PayPal / credit card!

Name : Adrian Wong
Bank Transfer : CIMB 7064555917 (Swift Code : CIBBMYKL)
Credit Card / Paypal : https://paypal.me/techarp

Dr. Adrian Wong has been writing about tech and science since 1997, even publishing a book with Prentice Hall called Breaking Through The BIOS Barrier (ISBN 978-0131455368) while in medical school.

He continues to devote countless hours every day writing about tech, medicine and science, in his pursuit of facts in a post-truth world.

 

Recommended Reading

Go Back To > Business | ComputerTech ARP

 

Support Tech ARP!

Please support us by visiting our sponsors, participating in the Tech ARP Forums, or donating to our fund. Thank you!

How Biren Got Its Own AI Chips Banned At TSMC!

TSMC stopped making artificial intelligence chips for China’s Biren Technology, and it was all Biren’s own fault!

 

TSMC Stops Making Biren AI Chips Over US Sanctions

TSMC – Taiwan Semiconductor Manufacturing Company – has suspended production of advanced artificial intelligence (AI) chips for China’s Biren Technology.

TSMC was forced to make this decision after public domain information revealed that the Biren BR100 and BR104 chips outperformed the NVIDIA A100 chip, which was used as the baseline of US sanctions.

While TSMC has not reached a conclusion on whether the top-of-the-line Biren BR100 or the slower BR104 meet or exceed US government threshold on advanced AI chip technology restrictions, it decided to stop production and supply of the Biren chips for now.

For TSMC to continue producing BR100 or BR104 chips, Biren must now prove that their chips do not offer “peak performance and chip-to-chip I/O performance equal to or greater than thresholds that are roughly equivalent to the [NVIDIA] A100“, or get an export licence from the US Department of Commerce.

And believe it or not – it was Biren Technology that created this mess for itself!

 

How Biren Screwed Up Its Own BR100 AI Chips

Biren, which is one of China’s most promising semiconductor design firms, earlier claimed that its AI chips that were being produced by TSMC are not covered by the latest US export restrictions.

However, its own website touts that the BR100 family of chips offers “world-class performance“, and has “improved by more than 3X” compared to mainstream rivals.

On top of that, Biren actually released a press statement on September 9, 2022, declaring that the slower BR104 was proven by the MLPerf to beat the NVIDIA A100!

Releasing such a statement less than 2 weeks after the US government ordered both AMD and NVIDIA to stop exporting their MI250 and A100 and faster AI chips to China is either amazing chutzpah, or a combination of hubris and idiocy.

Either way, the US government took notice, and TSMC came under pressure to comply with American export restrictions. Awesome PR, but stupid move, Biren…

Take a look at the benchmark results that Biren itself released into public domain, showing that the slower BR104 chip was between 27% and 58% faster than the NVIDIA A100.

With such results, the BR104 would certainly fall under the latest US tech export restrictions. No wonder TSMC quickly stopped making and supplying Biren BR100 series chips.

As powerful as the BR100 and BR104 GPGPU chips may be, they are now dead in the water as TSMC will not manufacture them anymore, and Biren Technology has no plausible alternatives for 7nm fabrication.

Read more : US Targets Chinese Military With New Chip Export Ban!

 

Biren BR100 AI Chips That TSMC Stopped Producing

The Biren BR100 and slower BR104 are General Purpose GPU (GPGPU) chips, which are targeted at artificial intelligence applications.

They are both fabricated on the TSMC 7nm process technology, and use chipset and 2.5D Chip-on-Wafer-on-Substrate (CoWoS) packaging technologies to achieve high yield, and high performance.

The Biren BR100 family of GPGPU chips supports up to eight independent virtual instances (SVI) – each physically isolated with their own hardware resources, for improved security.

Their chips are designed with a proprietary Blink high-speed GPU interconnect bus offering bandwidth of up to 448 GB/s, with the ability to connect up to 8 cards in a single node, using state-of-the-art PCI Express 5.0.

Biren Technology offers two BR100-based products – the Bili 100P OCP Accelerator Module (OAM), and the Bill 104P PCI Express accelerator card.

 

Please Support My Work!

Support my work through a bank transfer /  PayPal / credit card!

Name : Adrian Wong
Bank Transfer : CIMB 7064555917 (Swift Code : CIBBMYKL)
Credit Card / Paypal : https://paypal.me/techarp

Dr. Adrian Wong has been writing about tech and science since 1997, even publishing a book with Prentice Hall called Breaking Through The BIOS Barrier (ISBN 978-0131455368) while in medical school.

He continues to devote countless hours every day writing about tech, medicine and science, in his pursuit of facts in a post-truth world.

 

Recommended Reading

Go Back To > Computer | BusinessTech ARP

 

Support Tech ARP!

Please support us by visiting our sponsors, participating in the Tech ARP Forums, or donating to our fund. Thank you!

AMD, NVIDIA Banned From Selling AI Chips To China!

Both AMD and NVIDIA have just been banned from selling high-performance AI chips to both China and Russia!

Here is what you need to know…

 

AMD, NVIDIA Banned From Selling AI Chips To China + Russia!

On Friday, 26 August 2022, NVIDIA and AMD were both ordered by the US government to stop exporting high-performance AI chips to both China and Russia.

In its regulatory filing to the SEC, NVIDIA stated that the US government introduced this ban to prevent them from being used for military purposes, or by the Chinese or Russian military.

The ban uses the A100 Tensor Core GPU as the baseline for NVIDIA, and the 3rd Gen Instinct MI250 accelerator chip for AMD.

Effective immediately, anything equal to, or faster than, the NVIDIA A100 or the AMD Instinct MI250 can no longer be exported to either China or Russia. That includes the upcoming H100 chip.

On August 26, 2022, the U.S. government, or USG, informed NVIDIA Corporation, or the Company, that the USG has imposed a new license requirement, effective immediately, for any future export to China (including Hong Kong) and Russia of the Company’s A100 and forthcoming H100 integrated circuits.

DGX or any other systems which incorporate A100 or H100 integrated circuits and the A100X are also covered by the new license requirement.

The license requirement also includes any future NVIDIA integrated circuit achieving both peak performance and chip-to-chip I/O performance equal to or greater than thresholds that are roughly equivalent to the A100, as well as any system that includes those circuits. A license is required to export technology to support or develop covered products.

The USG indicated that the new license requirement will address the risk that the covered products may be used in, or diverted to, a ‘military end use’ or ‘military end user’ in China and Russia. The Company does not sell products to customers in Russia.

I should point out that the NVIDIA A100 was launched two years ago – on June 22, 2020, while the AMD Instinct MI250 was introduced on November 8, 2021.

So it appears that the US government wants to maintain at least a 1-year advantage in high-performance AI chips.

 

AMD, NVIDIA Set To Lose Billions From AI Chip Sales To China!

Both AMD and NVIDIA stopped selling chips to Russia, after the invasion of Ukraine, so there is no material loss from such a ban on sales to Russia.

However, they both stand to lose billions of dollars worth of sales to China. NVIDIA alone estimates that the ban will affect US$400 million of sales, just in the third fiscal quarter.

Even more worrying was the likelihood that this ban may delay the H100 chip’s launch, and force NVIDIA to move “certain operations” out of China.

The new license requirement may impact the Company’s ability to complete its development of H100 in a timely manner or support existing customers of A100 and may require the Company to transition certain operations out of China.

AMD however claims that the ban will not have a material impact on its business, because shipments of its less powerful MI100 chips are not affected.

At this time, we do not believe that shipments of MI100 integrated circuits are impacted by the new requirements.

We do not currently believe it is a material impact on our business.

Both AMD and NVIDIA will now have to discuss with their customers in China on using alternative (less powerful) chips, or obtaining special licenses from the US government.

 

Please Support My Work!

Support my work through a bank transfer /  PayPal / credit card!

Name : Adrian Wong
Bank Transfer : CIMB 7064555917 (Swift Code : CIBBMYKL)
Credit Card / Paypal : https://paypal.me/techarp

Dr. Adrian Wong has been writing about tech and science since 1997, even publishing a book with Prentice Hall called Breaking Through The BIOS Barrier (ISBN 978-0131455368) while in medical school.

He continues to devote countless hours every day writing about tech, medicine and science, in his pursuit of facts in a post-truth world.

 

Recommended Reading

Go Back To > Enterprise | ComputerTech ARP

 

Support Tech ARP!

Please support us by visiting our sponsors, participating in the Tech ARP Forums, or donating to our fund. Thank you!