Updated:01:06 AM CEST Jun,06
(new)
66 lottery login
91 club
okwin
bdg game
55 club
(c) 1998-2026 Gameguru Mania
Privacy Policy statement
|
7nm Intel Xe GPUs Codenamed 'Ponte Vecchio' - tech|
| (hx) 10:00 AM CET - Nov,14 2019 | | Intel's first Xe GPU built on the company's 7 nm silicon fabrication process will be codenamed "Ponte Vecchio," according to a VideoCardz report. These are not gaming GPUs, but rather compute accelerators designed for exascale computing, which leverage the company's CXL (Compute Express Link) interconnect that has bandwidth comparable to PCIe gen 4.0, but with scalability features slated to come out with future generations of PCIe. Intel is preparing its first enterprise compute platform featuring these accelerators codenamed "Project Aurora," in which the company will exert end-to-end control over not just the hardware stack, but also the software.
"Project Aurora" combines up to six "Ponte Vecchio" Xe accelerators with up to two Xeon multi-core processors based on the 7 nm "Sapphire Rapids" microarchitecture, and OneAPI, a unifying API that lets a single kind of machine code address both the CPU and GPU. With Intel owning the x86 machine architecture, it's likely that Xe GPUs will feature, among other things, the ability to process x86 instructions. The API will be able to push scalar workloads to the CPU, and and the GPU's scalar units, and vector workloads to the GPU's vector-optimized SIMD units. Intel's main pitch to the compute market could be significantly lowered software costs from API and machine-code unification between the CPU and GPU. |
|
last 10 comments: | Csimbi | (10:09 AM CET - Nov,14 2019 ) | quote:
These are not gaming GPUs
Of course it's not. You failed over and over again on GPUs.
I am guessing it's intended more of a competition for CUDA to balance the low core count on Intel CPUs.
This way, they can keep making low core count CPU and say: if you need high performance parallelization, buy this shit instead.
Then, they could include a target in Intel Parallel Studio and optimize for that shit.
Could work because CUDA is terrible at this. |
|
| gx-x | (12:24 AM CET - Nov,15 2019 ) | it's not for CUDA it's for AI. And we (well, ones that followed the development of this project) knew these will be for computation and AI. For desktop, intel offers newer IGP that is equally as bad as AMD APU Vega. Sooo...considering the intel's business scale and the amount of R&D, is it so hard to see them investing into future instead of the present?
You can still go desktop and tell me AMD is now better, just make sure you show me core for core, not 6/6 vs 6/12 and we fine.
PS. Anyway, intel has been working with nVidia to deliver Deep learning platforms. But they want to do everything and not have to include nvidia in the servers like they did until now.
you can start here: https://newsroom.intel.com/editorials/intel-deliver-leading-platform-artificial-intelligence/
continue here: https://software.intel.com/en-us/articles/accelerate-deep-learning-inference-with-integrated-intel-processor-graphics-rev-2-0
and then there are some newer papers on youtube, but I am too lazy to search my history for those. Sorry :)
This is where the money is in the future, it's a huge growing business that ain't slowing down. It is the future.
PPS. What low core count? In server segment, intel had 32 cores for 10 years almost. They have over 40 now (80 + threads) on desktop, well, I don't think i9900 is low core count, and you can (thanks AMD) get 6 and 8 core cheap and for peanuts now you can get and 4/8 core i3... |
|
All comments
|
|
Add your comment (free registration required)