UncleSam (@unclesam303) 's Twitter Profile
UncleSam

@unclesam303

ID: 1693221063637766144

calendar_today20-08-2023 11:19:05

7,7K Tweet

277 Followers

328 Following

Andrea Junker (@strandjunker) 's Twitter Profile Photo

The 7 largest Dow Jones drops in American history: 1. Trump, 3/16/20 -2,997.10 2. Trump, 3/12/20 -2,352.60 3. Trump, 3/9/20 -2,013.76 4. Trump, 6/11/20 -1,861.82 5. Trump, 4/3/25 -1,679.39 6. Trump, 3/11/20 -1,464.94 7. Trump, 03/18/20 -1,338.46 Tired of winning yet?

UncleSam (@unclesam303) 's Twitter Profile Photo

Dear Trump, can you please ask Xi to dial in a zoom call this weekend so that no one is panicking to call the other and the ppl across the world can have a good weekend and a brighter economic outlook to enjoy their lives?

UncleSam (@unclesam303) 's Twitter Profile Photo

Tired of winning. Donald J. Trump better call Xi to say sorry and you did not mean to harm the relationship. Xis got more cards in play. Remember China own significant amount of semiconductor upstream material? Xi can shutoff Japan and Korea supplier anytime he desires.

UncleSam (@unclesam303) 's Twitter Profile Photo

Elon better remediate his rhetoric attack to the liberals. Otherwise it’s shareholder who suffered from his stupid and unprofessional CEOship

UncleSam (@unclesam303) 's Twitter Profile Photo

Tired of winning yet? Michigan MAGA girls might not afford even 1 Barbie doll now. Maybe you can ask your granddaughter to donate to the innocent MAGA girls who just love dolls.

UncleSam (@unclesam303) 's Twitter Profile Photo

J10C first kill of Rafale marked a new era of super power. People would be interested to know what is J20+ PL21 can deliver today and beyond. youtu.be/mn-nr2l40WU?si…

tphuang (@tphuang) 's Twitter Profile Photo

HW posting Ascend inference cluster details on gitcode Here is Atlas 800I A2 (each containing 8 NPUs) connected using Fullmesh architecture. + CloudMatrix384 (CM384) which connects together multiple A2 LLM inference consists of Prefill + Decode Diagram below of A2 Prefill MLA 👇

HW posting Ascend inference cluster details on gitcode
Here is Atlas 800I A2 (each containing 8 NPUs) connected using Fullmesh architecture.
+ CloudMatrix384 (CM384) which connects together multiple A2
LLM inference consists of Prefill + Decode
Diagram below of A2 Prefill MLA 👇
Rihard Jarc (@rihardjarc) 's Twitter Profile Photo

$AMD might be starting to catch up to $NVDA! I HIGHLY recommend this interview with a current high-ranking employee at $DELL, talking about $AMD vs $NVDA and the client demand: 1. According to him, $AMD's main problem is ROCm, which is immature compared to $NVDA's CUDA, and the

$AMD might be starting to catch up to $NVDA! I HIGHLY recommend this interview with a current high-ranking employee at $DELL, talking about $AMD vs $NVDA and the client demand:

1. According to him, $AMD's main problem is ROCm, which is immature compared to $NVDA's CUDA, and the
UncleSam (@unclesam303) 's Twitter Profile Photo

Vulcan is a good example of how an industrial and open source API can work to standardize hardware for the ease of software developers. By the same virtue, the AI revolution will benefit from open source standard technology/API as well.

UncleSam (@unclesam303) 's Twitter Profile Photo

$intc kept delivery bad news for its data center customers. But this time their clients can buy $amd epyc instead. Throw away the intel trash.