You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Hi Cliff, I knew a little bit about Analog Rails back in my Synopsys days. I guess you are leveraging your EDA design technology into actual chips ? I’ve heard a little about compute-in-memory tech from Micron and their FWDNXT technology, but I think they have spun that out into a new company.
Daughter CS grad, Kayla, has her dream job doing software and data engineering at Riot Games in LA. (She‘s a big time gamer)
OpenTitan and in-memory computing sound important and interesting, but not sure it falls into either one’s highest interest zones. I can introduce you to either, with a...
His team just put together the models and associated backend for the NVIDIA AI Playground, a place to try different generative LLM models using NVIDIA data centers and software.
https://catalog.ngc.nvidia.com/orgs/nvidia/teams/playground/models/sdxl
Hi Cliff,
Thanks for the interest. I have 2 Berkeley EECS CS kids. Son, Kyle Kranen, works for NVIDIA in the machine learning research group. I can’t speak for him, but I think he’s having the time of his life right now in the generative and LLM AI whirlwind.
From what I have seen at TSMC Tech symposiums and OIP events, TSMC measures performance/density/power of new processes using representative designs taken through P&R, using whatever rev of the PDK and characterized library they have at the time. TSMC will generally spell out what kind of...
I lived and went to high school near Mt. Pleasant, Wisconsin. Pretty much all the residents, except the most crazed Trump supporters, have now recognized the 2017 Terry Gou/Trump/Paul Ryan/Scott Walker "deal" (yes, the 4 guys in the hyped-up photo) as the huge photo-op and political scam it was...
That's kind of what got me thinking about the different use cases for AI chips. There used to be plenty of different architectures of microprocessors and microcontrollers. But the market eventually whittled it down to just a few that addressed all of the then existing market sweet spots (HPC...
@blueone and @jwall , great assessments, plus follow-up. One component of the assessment that seems to be missing is comparable strength and market fit for training vs. inference data center vs inference at the edge. I think one of NVIDIA's big strengths is that the tackles the whole AI stack...
Man, the world has changed. C++ was the hot new thing when I started working. The software world for both of my kids today is Python, microservices stacks (including Airflow, Arrow, Spark), CUDA and various huge data storage systems / SQLs.
My take is different - you won’t find the best and the brightest in most electrical engineering programs. The best in STEM are all vectoring to computer science. I have two kids who graduated from the EECS program at Berkeley. They did the mandatory EE classes, but no more. One did one...
Thanks for the link, Fred ! Seemed really sketchy that the news was traveling by social media rather than through reviewed technical papers.
Not clear whether your earlier statement (below) was entirely correct (website vs company) , but they might as well shut down.
AFAIK, LK-99 comes from a very different branch of research and involves different materials than the disputed University of Rochester work.
https://www.nytimes.com/2023/07/26/science/ranga-dias-retraction-physics.html?smid=url-share
LK-99 results aren’t tied to one company or set of...
It’s an interesting race. Seems like China is also buying plenty from ASML.
ASML ups full-year sales forecast as China demand stays strong
https://www.reuters.com/technology/asml-reports-q2-earnings-19-billion-euros-beating-expectations-2023-07-19/
Like Arm, RISC-V’s path to volume and success probably lies in working their way up from the bottom leveraging licensing flexibility and freedom from royalties. A lot is going to depend on whether RISC-V finds a killer app/platform like mobile, PC or server, that fosters creation of a large...
Should be interesting to watch. Hard to substitute everything all at once while targeting a node that is full of performance and yield compromises, just to achieve a geometry. I’m also a bit confused - does that mean they will compete against the Honor smart phone business they spun out to...
@Paul2 , depends on how you measure “braininess”. If you measure in terms of number of scattered microcontrollers, I might agree. But if you measure in terms of compute power or number of transistors, I disagree. Between increased communication (plus OTA updates) and new safety systems with...
@Paul2 , number of chips might be falling but % of car component cost continues to rise, at least through 2030. About 40% of the total BOM cost today growing to 45% in 2030.
https://economistwritingeveryday.com/2022/01/04/how-many-semiconductor-chips-are-there-in-a-car/
And Huawei spun out their Honor smartphone division, and has moved to Qualcomm app processors and 5G, instead of HiSilicon Kirin chips.
https://www.forbes.com/sites/bensin/2021/05/24/honors-smartphones-will-use-qualcomms-chip/?sh=7e66b4f4ec36
Seems like HiSilicon is moving to far lower margin...
Looks like this is a summary from a purchasable market report. Share looks to be base on nominal value of AP-SoCs. And the notes seem to show Apple‘s rise is due to the fall of others via a softening market and inventory readjustment...
