What's new

Huawei now take on Nvidia with her own super AI chip (2x faster than Nv's best)

On 10 Oct 2018, Huawei launched her own AI platform and AI chips called Shengteng (Ascend) series.

The flagship of this AI series, Shengteng 910, can deliever over 256Tflops performance and 512Tiops (int8) at 350Watt, 2X better than Nvidia's best offering in terms of performance and performance per watt.

Huawei also launched her own complete toolchain and framework, including complier, core maths libraries and machine learning libraries and AI framework/cloud etc.

This is very serious stuff, and judging by Huawei's track record, Nvidia, and its stockholders should be extremely worried, to the world of AI computing hardware, it is like a shark come to a pond.:lol::D


f76f2d96-864e-4c65-b480-f32b36476816.png


6cf0d765-905a-4125-89a5-327c5f725960.png


Source (in Chinese):
http://news.mydrivers.com/1/598/598312.htm

You have just given the US regime more reason to whine about fair trade. :p:
 
.
This is the reason of the Trade War.

I guess USA is already know this from economic espionage, that is why they try to prevent it to happening.
 
.
Tesla V100 support much better floating computing performance: FP64/FP32/FP16, and its design is more towards computing (register/cache/programmable cache/memory bandwidth etc).

Whilst Turing like RTX2080Ti or Quadro RTX8000, simply design for game/rendering and some AI computation, but lack the high precious floating point power the Tesla V100 have.

For HPC, machine learning thats involve anything beyond simply GEMM, V100 is much better, of cause, if price is not a factor.
Just like i thought , so tesla v100 is more focus toward computing such as Handle load of data and Cloud server? While 2080 TI more to graphic processing.

It seems for me, huawei want to build full ecosystem. From cloud service to chip ( software to hardware )
 
Last edited:
.
Just like i thought , so tesla v100 is more focus toward computing such as Handle load of data and Cloud server? While 2080 TI more to graphic processing.

It seems for me, huawei want to build full ecosystem. From cloud service to chip ( software to hardware )
not more focused look at Tesla v100
37b49a8a7f.jpg

there is no video output connection , it only can be used for computing
 
.
LeGenD is a US lackey, despite US drone violating Pakistan Airspace, killing pakistanis at will. He will try lick the boots for American. Blindly defend anything from US. He is good at selective post and facts to try pull off his points.
Illiterate people like you should not comment on matters which they do not understand.

I am speaking with experience and exposure (IT background); raw power means squat unless it translate into "meaningful ends."

NVIDIA Tesla V100 is a specialized product, intended for computation and machine learning (HPC). Essentially, Tesla V100 have little utility for personal use, and suitable for use in workstations and supercomputers.

NVIDIA RTX 2080 Ti is the endgame - it can perform all functions of Tesla V100 and much more; the former is incredibly efficient per $ even in strictly HPC role:

fp32-performance-per-dollar-2080ti-v100-titanv-1080ti-1.png


- and a graphics processing powerhouse with no peer in the world, by far. Graphics processing is the domain where the true competition exist, and the most challenging endeavor for any company to push boundaries in.

You can buy two or more RTX 2080 Ti for less price than a single Tesla V100, SLI them, and enjoy vastly superior HPC output on your PC, let alone a workstation. And not to forget, vastly superior graphics processing.

---

Now, shall I remind you from where Huawei source components for its hardware? Two American firms involved. This is common practice in China - a Chinese company imports products from US, assemble them, label them as "Made in China" and voila - consumer base is happy.

While Huawei might have developed a more powerful HPC-oriented product than Tesla V100, the comparison shown in the original article is ambiguous in the absence of real-world computations in various applications. Huawei should hand over its product to neutral sources for proper evaluation. If it is good, it will shine. And NVIDIA will get the memo. Otherwise, hot air.

---

On a side note: irrespective of your nationality, if you do not respect NVIDIA and its contributions to computing industry since its birth - you are not fit to partake in computer-related discussions. Stay out and shove your political biases elsewhere.
 
Last edited:
.
Illiterate people like you should not comment on matters which they do not understand.

I am speaking with experience and exposure (IT background); raw power means squat unless it translate into "meaningful ends."

NVIDIA Tesla V100 is a specialized product, intended for computation and machine learning (HPC). Essentially, Tesla V100 have little utility for personal use, and suitable for use in workstations and supercomputers.

NVIDIA RTX 2080 Ti is the endgame - it can perform all functions of Tesla V100 and much more; the former is incredibly efficient per $ even in strictly HPC role:

fp32-performance-per-dollar-2080ti-v100-titanv-1080ti-1.png


- and a graphics processing powerhouse with no peer in the world, by far. Graphics processing is the domain where the true competition exist, and the most challenging endeavor for any company to push boundaries in.

You can buy two or more RTX 2080 Ti for less price than a single Tesla V100, SLI them, and enjoy vastly superior HPC output on your PC, let alone a workstation. And not to forget, vastly superior graphics processing.

---

Now, shall I remind you from where Huawei source components for its hardware? Two American firms involved. This is common practice in China - a Chinese company imports products from US, assemble them, label them as "Made in China" and voila - consumer base is happy.

While Huawei might have developed a more powerful HPC-oriented product than Tesla V100, the comparison shown in the original article is ambiguous in the absence of real-world computations in various applications. Huawei should hand over its product to neutral sources for proper evaluation. If it is good, it will shine. And NVIDIA will get the memo. Otherwise, hot air.

---

On a side note: irrespective of your nationality, if you do not respect NVIDIA and its contributions to computing industry since its birth - you are not fit to partake in computer-related discussions. Stay out and shove your political biases elsewhere.
Im not aware two american company involved, can you tell me which company involved?
https://defence.pk/pdf/threads/more...rst-7nm-arm-server-chip.581220/#post-10857872
 
.
not more focused look at Tesla v100
37b49a8a7f.jpg

there is no video output connection , it only can be used for computing
Good post.

RTX 2080 Ti connectivity options:

Zotac-Zotac-GeForce-RTX-2080-Ti-AMP-displays.png


@Offshore

RTX 2080 Ti is the graphics processing powerhouse, with no peer in the world.



Im not aware two american company involved, can you tell me which company involved?
https://defence.pk/pdf/threads/more...rst-7nm-arm-server-chip.581220/#post-10857872
I can give you some pointers:-
  • Lumentum Holdings Inc
  • Oclaro
  • NeoPhotonics Corp
  • Finisar Corp
  • Inphi Corp
  • Analog Devices Inc
  • Xilinx Inc
  • Google
Lesser known fact.

ZTE, Huawei, BATX (Baidu, Alibaba, Tencent, Xiaomi), Didi (China’s Uber), ICBC, Bank of China, China Mobile, China Telecom, Petro China, SAIC Motor and many others rely American technologies (components, software and intellectual properties) to function. Notable American/Western suppliers include Apple, 3M, AMD, Applied Materials, Cisco, Corning, Google, Intel, Micron, Microsoft, Qualcomm, Seagate and Western Digital.

---

Huawei have a diversified supplier base in comparison to ZTE actually. Huawei is better prepared for potential contigencies. I know about this brand before the era of smartphones. Any IT specialist would.

I laugh on the face of those who try to boast about Huawei on my face. Hello, I know better.
 
Last edited:
.
Good post.

RTX 2080 Ti connectivity options:

Zotac-Zotac-GeForce-RTX-2080-Ti-AMP-displays.png


@Offshore

RTX 2080 Ti is the graphics processing powerhouse, with no peer in the world.




I can give you some pointers:-
  • Lumentum Holdings Inc
  • Oclaro
  • NeoPhotonics Corp
  • Finisar Corp
  • Inphi Corp
  • Analog Devices Inc
  • Xilinx Inc
  • Google
Lesser known fact.

ZTE, Huawei, BATX (Baidu, Alibaba, Tencent, Xiaomi), Didi (China’s Uber), ICBC, Bank of China, China Mobile, China Telecom, Petro China, SAIC Motor and many others rely American technologies (components, software and intellectual properties) to function. Notable American/Western suppliers include Apple, 3M, AMD, Applied Materials, Cisco, Corning, Google, Intel, Micron, Microsoft, Qualcomm, Seagate and Western Digital.

---

Huawei have a diversified supplier base in comparison to ZTE actually. Huawei is better prepared for potential contigencies. I know about this brand before the era of smartphones. Any IT specialist would.

I laugh on the face of those who try to boast about Huawei on my face. Hello, I know better.
What IT spesialis are you? Here we have people working oN Real IT, not wannabe IT
 
.
What IT spesialis are you? Here we have people working oN Real IT, not wannabe IT
I am Bachelors in Computer Science and have experience in programming. I was into this stuff since 3DFx days.

Some of you are way too arrogant and judgemental, by the way. No substance in arguments.

Wannabe IT? Seriously?

What is your expertise? Masters in silly rants?
 
Last edited:
.
I am Bachelors in Computer Science and have experience in programming. I was into this stuff since 3DFx days.

Some of you are way too arrogant and judgemental, by the way. No substance in arguments.

Wannabe IT? Seriously?

What is your expertise? Masters in silly rants?
Im asking you which two american company involved in huawei new chip, yet you gave me beating around the bus answer.

@xunzi can you gave me the answer? Is there any american company involved ?
 
.
Im asking you which two american company involved in huawei new chip, yet you gave me beating around the bus answer.
I provided you a list of companies (bullet format) which have a history of supplying 'components' to Huawei.

It is up to Huawei in how it utilize imported components to establish its products; Research & Development stuff (and collaboration programs).

Not my fault, if you have comprehension problems. Design of Huawei new chip - is not my point. Go back and re-read my statements.

By the way, British computing expert Sophie Wilson is the founder of ARM processor type.
 
Last edited:
. .
Back
Top Bottom