From somebody I don't know (a stranger is just a friend you haven't yet met - wrongly attr. to W.B. Yeats):
There's been a lot of low quality GPT-4 speculation recently. So, here's a relatively informed GPT-4 speculation thread from an outsider who still doesn't know that much. 🧵
— Matthew Barnett (@MatthewJBar) December 20, 2022
At the time, the top supercomputer listed by TOP500 was Summit from IBM, though the much faster Supercomputer Fugaku would be listed in the June 2020 TOP500 release one month later.https://t.co/nuAV223RFc
— Matthew Barnett (@MatthewJBar) December 20, 2022
From NVIDIA, "The theoretical peak performance of the Tensor Cores on the V100 is approximately 120 TFLOPS. This is about an order of magnitude (10x) faster than double precision (FP64)"https://t.co/3m0pbbIPyj
— Matthew Barnett (@MatthewJBar) December 20, 2022
....MUCH MORE (Thread)
HT on the Twitter thread: Marginal Revolution
My ears sort of perk up when Nvidia is mentioned. Kind of like an old dog when it thinks it hears the car keys. Though recently, rousing oneself at the sound of NVDA has been more like realizing those are just the keys to the room where they store the cat furniture. 'Effin' cat.
But back in the day, 2015, and adjusted for the 4:1split (either $6.50 to the current $152 or $26 to $608), glory days. Maybe 300 posts:
....One of the reasons we were able to glom onto Nvidia at $21 was stuff like 2014's "Deep Learning is VC Worthy" pointing toward A.I and other areas faster chips were allowing 'puters to tackle. Our first post on NVDA was May 2015's "Nvidia Wants to Be the Brains Of Your Autonomous Car (NVDA)" which was followed by a hundred more, including 2018's play on that first post title "UPDATED—NVIDIA Wants to Be the Brains Behind the Surveillance State (NVDA)" many with this introduction from May 2016's "NVIDIA Sets New All Time High On Pretty Good Numbers, "Sweeping Artificial Intelligence Adoption" (NVDA)":
We are fans.It worked out.
Before we go any further, our NVIDIA boilerplate: we make very few calls on individual names on the blog but this one is special.
They are positioned to be the brains in autonomous vehicles, they will drive virtual reality should it ever catch on, the current businesses include gaming graphics, deep learning/artificial intelligence, and supercharging the world's fastest supercomputers including what will be the world's fastest at Oak Ridge next year.
Not just another pretty face.
Or food delivery app.
After hours the stock is changing hands at $38.31 up 7.70% which, if it holds through tomorrow's regular session, beats the old highs from 2007.
Possibly also of interest, last summer's "There Is A New World's Fastest Supercomputer (maybe) NVDA":
....Two European (LUMI and France's Adastra), two Chinese and Japan's Fugaku add diversity to the U.S. national laboratories machines with NVIDIA's Selene being the only corporate machine in the top10. Germany and Italy have machines in the 11th and 12th fastest positions with Microsoft's Azure in the 13th spot. Here are the first 100 fastest.
The highlights press release (above) takes pains to point out how deeply NVDA accelerators have penetrated supercomputer architecture:
Highlights from the List
A total of 170 systems on the list are using accelerator/co-processor technology, up from 151 six months ago. 84 of these use NVIDIA Volta chips, 54 use NVIDIA Ampere, and 8 systems with NVIDIA Pascal....
And just for grins and giggles, April 18, 2022:
"Nvidia’s Next GPU Shows That Transformers Are Transforming AI" (NVDA)I don't think NVDA has seen its lows:
And I'm not sure their plan to be the engine that powers the metaverse will ever amount to anything.
And I know it is not as interesting as it was in 2015 - 2016 when we started touting it at $26 to $36 ($6.50 to $9 accounting for last year's 4:1 split)*
But it still has some of the most amazing tech out there.
From IEEE Spectrum, April 8:....
