Chinese semiconductor thread II

Maikeru

Major
Registered Member
They don’t have any real weight in DC in foreign policy matters. Even the whole Silicon Valley tech lobby has very limited influence relative to the Natsec and defense lobbies in that domain. They’re not swinging the rest of Congress with them and the only potential way their views will have an impact is if they get some inside track influence on cabinet appointments or a direct line to the President’s ear.
"We urge you to use all forms of leverage available to the U.S. government to bring our allies along in aligning their export controls with ours," they wrote.

Not like they want less sanctions, they just want allies leaned on more to go along with them.
 

tokenanalyst

Brigadier
Registered Member
Please, Log in or Register to view URLs content!

I’ve personally met Alex Padilla before when I was in D.C. This letter from him and Zoe Lofgren carry substantial weight IMO. I think they’re getting pretty concerned the sanctions on China are backfiring.
The problem is that now is not just the average Stooge, like in 2014 with 5G. Now there is China-Hawk stooge, National Security Pentagon-Think Tank stooge and now there is a new breed of stooges, we can call them the Silicon Valley AI doomsayer crypto bro stooge who think ChatGPT is the atomic bomb is this generation, they think they are modern age Crypto-Oppenheimer. Basically like zombies in a game, you have the basic one, the one who run faster and the tank. All different but all mindless zombies.

They are ideological motivated, they see US semiconductors companies concerns as treason and unpatriotic, kinda ironic because Eric Schmidt, doesn't seem to have a problem with google making billions in China but for somehow he has a problem with ASML making a living there. Only untold amount of destruction could change these guys minds and I personally don't think even that, they will likely to see US semi companies revenue collapsing and job losses as " effective decoupling" rather than very bad policy.
 

tokenanalyst

Brigadier
Registered Member
"We urge you to use all forms of leverage available to the U.S. government to bring our allies along in aligning their export controls with ours," they wrote.

Not like they want less sanctions, they just want allies leaned on more to go along with them.
I think what they are trying to say is that "we should all sink in the same ship or escape in the same boat", like if US cannot bring Japan and Holland export controls to be as stringent as the US then US export controls should mirror more those of Japan and Holland.
 

latenlazy

Brigadier
"We urge you to use all forms of leverage available to the U.S. government to bring our allies along in aligning their export controls with ours," they wrote.

Not like they want less sanctions, they just want allies leaned on more to go along with them.
Lesson in reading the link even when you don’t want to :eek::eek:
 

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
I mean, I think it's extremely unlikely that 910C will have worse power efficiency scaling than 910 since it's an iteration, but okay.
I'm referring to total power consumption. 1 TFLOP/W for 910B really isn't great. If Ascend-910C is around 1000 TFLOPS, then 1000W is a lot.

Another thing that's underrated is system-level power efficiency (i.e., including power for CPU, networking, etc.). After all, data center power draw is more than just the cards, you also need to consider the entire system and the data center efficiency (PUE) itself.

We can calculate the
Please, Log in or Register to view URLs content!
. Not pictured is the H100 SuperPOD, but the answer to that is
Please, Log in or Register to view URLs content!
.

A proper calculation also factors in PUE, which is how many excess watts the data center draws to generate one watt. For instance, a PUE of 1.1 is excellent, which means that for every 1 watt used to generate compute, 1.1 watts are pulled from the grid. H100 and A100 systems generally have a PUE of 1.25, while Atlas 900/Ascend 910 is rated at the much better 1.15.

Finally, I'll point out that energy efficiency gains aren't purely tied to node scaling. NVIDIA's B200 and H100 are both using TSMC 4N, and yet the energy efficiency of the B200 is ~65% better than the H100. At a card level (i.e., not system level as calculated above) B200 generates 2.3 FP16 TFLOPS/W vs. H100's 1.4 FP16 TFLOPS/W. So we can't write off further efficiency improvements for Huawei's next accelerator either, even if they stick with SMIC's N+2.
i would stop thinking about things just in terms of TFLOPS/W, but rather how many W your data center needs and what that does to power requirement in the region and also cooling.
 

tokenanalyst

Brigadier
Registered Member

Yuwei Semiconductor successfully shipped the first batch of i12-F300 ultra-precision wafer defect inspection products​


Recently, the first batch of i12-F300 ultra-precision wafer defect inspection equipment of Yuwei Semiconductor was successfully shipped from Yuwei Hefei and successfully delivered to important customers in the domestic integrated circuit field.

The i12-F300 equipment independently developed by Yuwei Semiconductor is the company's new generation of wafer inspection products. Relying on Yuwei's rich experience and technical reserves accumulated in the field of integrated circuit wafer inspection over the years, it is based on multi-mode high-precision optical imaging, high-stability imaging control, high-efficiency data processing system, and multi-type material compatibility. Combined with Yuwei's proprietary intelligent detection + classification algorithm and software system, compared with existing integrated circuit wafer defect detection products, it can provide more powerful defect detection capabilities and higher detection efficiency, and meet the subtle defect control needs of different integrated circuit customers in the process links.​

1723736862689.png


In the integrated circuit wafer manufacturing process, the i12-F300 device can be used for defect detection after processes such as lithography, etching, chemical grinding, and thin film deposition, to achieve rapid process monitoring and problem feedback, and reduce customer yield losses.

Please, Log in or Register to view URLs content!
 

tokenanalyst

Brigadier
Registered Member

Qingyi Optoelectronics plans to raise 1.2 billion yuan to build a high-precision, high-end semiconductor mask base​

According to the Semiconductor Industry Network, Qingyi Optoelectronics recently issued an announcement stating that the company plans to issue stocks to specific objects to raise a total amount of no more than 1.2 billion yuan (including principal). The funds raised after deducting relevant issuance expenses will be used to invest in the first phase of the high-precision mask production base construction project and the first phase of the high-end semiconductor mask production base construction project.

The announcement shows that the implementation entity of the first phase of the high-precision mask production base construction project is the subsidiary Foshan Qingyi Optoelectronics, whose main product is flat panel display mask, which is one of the core materials in the upstream of the flat panel display industry chain. The total planned investment of this project is about RMB 800 million, and it is planned to invest RMB 600 million in raised funds. It mainly produces 8.6-generation and below high-precision mask plates, which are applied to a-Si, LTPS, AMOLED, LTPO, MicroLED and other flat panel display products. After the project is completed and put into production, it will increase the company's flat panel display mask production capacity, improve the localization degree and supporting level of high-precision flat panel display masks, and fill the domestic industry gap.

The implementation entity of the first phase of the high-end semiconductor mask production base construction project is the subsidiary Foshan Qingyi Micro, whose main products are high-end semiconductor masks covering 250nm-65nm processes. The total planned investment of this project is about RMB 605 million, and it is planned to invest RMB 600 million in raised funds. The project can improve the technical capabilities and production capacity of Qingyi Optoelectronics' semiconductor mask products, optimize its semiconductor mask product structure, improve its technology and process level, and increase its market share in the semiconductor mask market.

Please, Log in or Register to view URLs content!
 
Top