Thursday, February 23rd 2023
Intel Xeon W-3400/2400 "Sapphire Rapids" Processors Run First Benchmarks
Thanks to the attribution of Puget Systems, we have a preview of Intel's latest Xeon W-3400 and Xeon W-2400 workstation processors based on Sapphire Rapids core technology. Delivering up to 56 cores and 112 threads, these CPUs are paired with up to eight TeraBytes of eight-channel DDR5-4800 memory. For expansion, they offer up to 112 PCIe 5.0 lanes come with up to 350 Watt TDP; some models are unlocked for overclocking. This interesting HEDT family for workstation usage comes at a premium with an MSRP of $5,889 for the top-end SKU, and motherboard prices are also on the pricey side. However, all of this should come as no surprise given the expected performance professionals expect from these chips. Puget Systems has published test results that include: Photoshop, After Effects, Premiere Pro, DaVinci Resolve, Unreal Engine, Cinebench R23.2, Blender, and V-Ray. Note that Puget Systems said that: "While this post has been an interesting preview of the new Xeon processors, there is still a TON of testing we want to do. The optimizations Intel is working on is of course at the top, but there are several other topics we are highly interested in." So we expect better numbers in the future.
Below, you can see the comparison with AMD's competing Threadripper Pro HEDT SKUs, along with power usage using different Windows OS power profiles:
Power usage:
Source:
Puget Systems
46 Comments on Intel Xeon W-3400/2400 "Sapphire Rapids" Processors Run First Benchmarks
These CPUs are meant for Server type workloads, it needs wide instruction support, and heavily virtualized environments.
E cores lacks some instruction sets , and virtualization software don't like Hybrid architectures.
So P core only or E core only.
Just don't mix them in a single CPU.
M2 Max is about on par with AMD's 7700x, also known as 1 Zen4 CCD, Genoa has... 12.
Ampere on the other hand... is competative, and Qualcomm with Nuvia cores will be as well.
www.servethehome.com/wp-content/uploads/2022/11/AMD-EPYC-9004-Genoa-c-ray-8K-rendering-Performance.jpg
amperecomputing.com/processors/ampere-altra
I expect for audio work in particular SR will do quite well. Intel's main advantage is also it's main disadvantage, owning its own fabs. It is lagging behind on efficiency, but it can sure make a lot of them.
It is very clearly not the top dog in rendering though.
To be fair to Intel, maybe they can still compete on price and availability. It won't be good for their already slim margins, but it will be good for the users.
So the e cores are weaker but come at much higher numbers. 4e are faster than 1p so I speculate that 160 threds of mix p+e will match and go above the 128 of top TR.
But if AVX and high IO is needed, nix the e cores. No questions ask.
Cinebench (Cinema4D) was always done for Intel (be it Mac or PC) and newest and greatest has nothing to show there. Zen3 TRP stomps SR-X, like its not even funny. (single core is only for benchmark junkies)
Unless you have workflow which will specifically benefits from Intel architecture or firmware updates do some extra magic internally, this release is DOA. No matter how Intel PR will try to spin it or rename it. SR-X can't go toe to toe with old Zen3 and it sucks 50% more power to be in 2nd place. Like whaaat? I was hyped, but I'm fully cured already. Worst part is that AMD may completely ignore new TRP now.
Maybe some day we can have 56p vs 32+96e.
To have have better multi-threading performance ... p-cores are faster then e-cores, but for the same die area, e-cores add more multi-threading performance (at least I think that's how it was).
When when only using p-cores, the same die area would have resulted in a 12c/24t CPU ... they 'sacrificed' 4 p-cores to add 16 e-cores, resulting in a 24c(8p+16e)/32t(16p+16e) CPU.
So the assumption is that 12p cores have lower multi-threading performance then 8p+16e.
Would be interesting to see some benchmarks in software like abaqus explicit or LSdyna or other software where the large cost is the licence per core…. I think per core performance is in advantage of intel is it not ?