- 254comments
- 39comments
- 116comments
- 123comments
- 20comments
- 39comments
- 183comments
- 60comments
- 10comments
- 11comments
- 35comments
- 7comments
- 7comments
- 103comments
- 26comments
- 9comments
- 259comments
- 249comments
- —discuss
- 29comments
- 40comments
- 304comments
- 140comments
- 176comments
- 20comments
- 10comments
- —discuss
- 9comments
- 45comments
- 78comments
Depending on Application 144 Core is a bit low considering both Intel and Ampere has 256+ Core in the pipeline. Hopefully they have one with 8 Chiplet design where we get 288 Core per socket.
Stupid question: why doesn't fujitsu commoditize this for deep-learning stuff and steal some of nvda's bullrun ?
I came here to ask this. I don't agree it is a stupid question ;)
Fujitsu seems to be weirdly stubborn about global commercialisation of their CPUs. I know of at least a couple of HPC customers who wanted to build systems based on A64FX and the answer was basically "...nah we're good".
(Apologies for necroing this thread)
Fujitsu actually does sell their A64FX processors. In fact, they are running in one or two European super computers[1][2]. Still, that's not to say that buying them Fujitsu is easy and/or straightforward.
[1]: https://www.hpcwire.com/off-the-wire/fujitsu-selected-to-del...
[2]: https://www.youtube.com/watch?v=yInhWFztvXs
Fujitsu's Monaka CPU = Armv9-A ( The CPU Architecture for the Future of AI )
https://www.arm.com/architecture/cpu/a-profile/armv9 --> "The CPU Architecture for the Future of AI" --> "Inference on Arm CPUs"
"The CPU Architecture for the Future of AI : The Armv9-A architecture gives AI developers the programming tools and environment necessary to innovate at pace for the rapidly expanding AI market. AI applications process vast datasets, perform complex calculations, and interpret data collected in real-time to benefit both commercial and everyday tasks. As these programs compute more data, the need for both performance and security becomes paramount. Armv9-A is built with both needs in mind. It offers faster compute for high-performance use cases with a suite of security features that help ensure both data and operational integrity. These benefits are delivered on the device or server without compromising power consumption."