Wednesday, November 13th 2019
7nm Intel Xe GPUs Codenamed "Ponte Vecchio"
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.
Image Courtesy: Jan Drewes
Source:
VideoCardz
"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.
50 Comments on 7nm Intel Xe GPUs Codenamed "Ponte Vecchio"
Clearly, they went balls deep on it, so this tells you how confident they are about 7nm, too.
AMD should start taunting them and make fun of them for gluing everything.
videocardz.com/newz/intels-first-xe-graphics-processor-is-called-ponte-vecchio
The lineup will be for business users. The gaming lineup will come in 2020.
How?
Yeah, I'm kinda skeptical, too.
("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.")
This to me means Intel is up to it's usual dirty tricks to leverage things in their favor.
While it may not mean such directly it could be used to block others (AMD?) from being able to "Do" certain things AND to make developers dance to their tune, including the gaming industry - At the very least have the ability to manipulate things in their favor.
One possible example of such abuse/dirty tricks:
You don't favor Intel products by slowing/crippling gaming performance if the software sees an AMD CPU in use you get restricted/cutoff from being able to use the software for your gaming platform development or for it to work with titles already released.
And don't try to convince me it's not possible for them (Or anyone else for that matter TBF) to try it.
Intel has been caught red handed in many schemes over time as fact, not fiction and certainly isn't above doing it again as a company, esp one that's now becoming desperate and you know what desperate folks will do as a rule......
In either instance it is up to them to speak up and do something about it should it occur.
The other thing is, since Intel is not trying to compete with NV and AMD in desktop with their GPU, gives an impression that Intel feels weak in that department and cant compete with the two companies. So this Xe is kind of an attempt to exist and deliver some sort of GPU since it's been told Intel is working on one. More like to be true to its word.
Intel is trying to make money. Margins in compute, especially AI, are far higher than desktop. If they lock down an architecture that works there, it is easier to bring that down to desktop instead of starting here. Even with Intel having been working on their iGPU for years, remember that even here the common assumption is that Intel cannot create a discrete GPU at all.
Edit:
Oh, right. That x86 part is in the original post/news blurb. I call bullshit.
Making money is obvious thing. Not sure why you bring that one up.
We know enough about Xe to say it is not doing x86.I brought up money because there is a lot of that in compute/AI market. And you said Xe not coming to desktop at first is a sign Intel feels weak. When you are designing a new GPU, that is the obvious market to go for. Desktop GPUs are dirt cheap and require a lot of twitchy games-specific software work.