Comments on: Intel Pushes Out Hybrid CPU-GPU Compute Beyond 2025 https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/ In-depth coverage of high-end computing at large enterprises, supercomputing centers, hyperscale data centers, and public clouds. Thu, 16 Nov 2023 13:26:27 +0000 hourly 1 https://wordpress.org/?v=6.7.1 By: John https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-216358 Thu, 16 Nov 2023 13:26:27 +0000 https://www.nextplatform.com/?p=142003#comment-216358 Apple dumped Intel because their chips were just too inefficient for Apple’s MacBook’s. But honestly, I am not sure Intel has made that much progress in efficiency with their Big/Little design. I know they had to do something but clearly at least in mobile Intel has not made significant progress in my opinion. Having bought two different Adler Lake laptops both with U series I don’t find them any better at performance or efficiency then with 10th gen models I had previous. I think in general Intel with its hybrid solution has just managed to try and use E cores to offset power consumption and it just has not worked out that well. I mostly wish I had bought Ryzen low powered laptops instead given that same models with AMD CPU’s had better performance and battery life.

]]> By: EC https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205831 Mon, 13 Mar 2023 20:20:06 +0000 https://www.nextplatform.com/?p=142003#comment-205831 In reply to OGeneral.

TSMC increased wafer prices by 60% between A100 (7nm) and H100 (5nm). (https://www.techpowerup.com/301393/tsmc-3-nm-wafer-pricing-to-reach-usd-20-000-next-gen-cpus-gpus-to-be-more-expensive)

]]> By: OGeneral https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205748 Sat, 11 Mar 2023 14:42:47 +0000 https://www.nextplatform.com/?p=142003#comment-205748 We so desperately need a competitor in the DL training market. Unfortunately AMD still is a long way off to threaten nVidia’s quasi monopoly here. And nVidia is taking the piss by effectively doubling the price from one generation to the next (compare H100 to A100 prices)

]]> By: emerth https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205637 Thu, 09 Mar 2023 08:05:48 +0000 https://www.nextplatform.com/?p=142003#comment-205637 In reply to ST018.

Gotta wonder how in 2019 or so anyone can patent what is more or less the concept of a shared bus.

]]> By: emerth https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205636 Thu, 09 Mar 2023 08:02:35 +0000 https://www.nextplatform.com/?p=142003#comment-205636 I am so happy to see that Intel’s ability to come up with so many names remains unimpaired /s. Halfway thru the article and I’m having to make a list.

Once, in the distant past, we indicated product iterations with integers… any fool could keep up.

I wonder if Intel has naming schemes for it’s naming schemes?

]]> By: Michael Alan Bruzzone https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205579 Tue, 07 Mar 2023 19:15:28 +0000 https://www.nextplatform.com/?p=142003#comment-205579 The 1987 introduction of IBM PC AT 3D add-in-card occurred at PC Expo in Chicago based on IBM merchant silicon known as Professional Graphics Adapter, implemented by Orchid Technology, was introduced demoing AutoCAD walk throughs and 3D model shading. While the card was advertised to AutoCAD users, the primary use target was government. Mike Bruzzone, Camp Marketing

]]> By: ST018 https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205574 Tue, 07 Mar 2023 14:59:38 +0000 https://www.nextplatform.com/?p=142003#comment-205574 “Perhaps for there will indeed be five different Falcon Shores configurations, two more than the three shown above”
I guess the 4-chiplet-MCM-scheme was more a visual aid on Intel’s marketing material rather than an actual partitioning structure for Falcon Shores.
Take a look on the patent US10909652B2 “Enabling product SKUs based on chiplet configurations” originally filled by Intel in 2019.
There are as much as 7 dies (6 compute + 1 I/O) with some sort of attached HBM.

https://patents.google.com/patent/US10909652B2

]]> By: Timothy Prickett Morgan https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205558 Tue, 07 Mar 2023 02:56:08 +0000 https://www.nextplatform.com/?p=142003#comment-205558 In reply to Eric Olson.

Could not agree more.

]]> By: Eric Olson https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205554 Mon, 06 Mar 2023 23:26:28 +0000 https://www.nextplatform.com/?p=142003#comment-205554 Not too long ago the Fugaku supercomputer placed first on the top500 without any GPUs. What’s interesting is Fugaku was explicitly engineered to perform well on a variety of relevant HPC workloads rather than the largely irrelevant Linpack.

It’s possible HBM in the Xeon Max will hae a similar type of performance. Moreover, optimising code for the fast-slow linear address model of the Max is likely easier (and even possible) in cases where GPU offload is difficult.

]]> By: EC https://www.nextplatform.com/2023/03/06/intel-pushes-out-hybrid-cpu-gpu-compute-beyond-2025/#comment-205549 Mon, 06 Mar 2023 19:30:09 +0000 https://www.nextplatform.com/?p=142003#comment-205549 Great article Mr. Morgan, been refeshing TNP all weekend anticipating your thoughts on this.

A nit or two on history if I may: I think the idea “modern” is missing from your recitation of TI and IBM parts, but I appreciate the walk down the lane with 8514. And 3D acceleration, for the PC anyway, didn’t really begin in 1995. Microsoft hadn’t released the first DirectX API until Sept of that year, so the applications these chips would’ve wanted to run didn’t exist yet (and then one could argue 3DFX had the first commercially viable purpose built 3D accelerator in the initial VooDoo release at the end of 96). But what Nvidia claims is the first “programmable shader” (eg modern) in their NV10 part or GeForce256 that released in Oct 99, anyway . . .

>> takes away a means for Intel to build credibility in the HPC and AI training communities.<<

To dig a little deeper on the key point of this story: OneAPI progress seems dead in the water with no viable hardware and a pushed out roadmap. Intel will have trouble advancing their software stack and that allows further entrenchment by competitors in this segment. The way it’s going for Intel, this self-own may prove existential for their DC accelerator aspirations.

]]>