|
These processors will have very decent performance for many applications, very similar to that of AWS Graviton5, but with more cores per socket. However, the claim made by Arm: "the Arm AGI CPU, for agentic AI infrastructure, delivering more than 2x performance per rack compared with x86 platforms" is obviously false. The new Intel Clearwater Forest Xeon processors use Darkmont cores, which have approximately the same performance per core, the same die area per core and the same power consumption per core as the Neoverse V3, but Intel offers 288 cores per socket and 576 cores per board, in comparison with only 136 cores per socket for Arm. Therefore there is no chance that these new Arm processors can provide more performance per rack than Intel Clearwater Forest. For applications that benefit from array operations, the AMD Zen 5 compact cores have much more performance per core than Neoverse V3 and AMD has provided 192 cores per socket for a long time. There is no chance for the new processors to exceed the performance per rack of Zen 5, but for those applications that do not benefit from array operations, these new Arm CPUs should have better performance per watt than Zen 5. But by the end of the year AMD should have Zen 6 Epyc CPUs, with more cores per socket, enhanced performance per core and improved performance per watt, so then there would be even less opportunities for these Arm CPUs to be better at something. The only way how the claim of Arm can be true is if they have compared their new CPUs with antiquated CPUs like the Intel Granite Rapids Xeon CPUs, instead of comparing with state-of-the-art Intel Clearwater Forest and AMD Zen 5. |
In no world will Darkmont perform like Neoverse V3 / Cortex-X4. Darkmont is much slower.
SPECint2017
Darkmont @ 3.5 GHz boost = 7.13 points
Cortex-X4 @ 3.2 GHz = 8.20 points (+15% faster)
Source: David Huang, https://blog.hjc.im/spec-cpu-2017
Arm's Neoverse V3 136-core CPU has a 3.2 GHz base clock, so the exact same as this Cortex-X4. Your real problem arises when a 288C Clearwater Forest CPU at the highest 500W TDP means a maximum of 1.7W per core (generous, as we're excluding uncore, fabric, cache, etc.). It's probably closer to 1.5W, but let's be generous and toss in +200mW.
Darkmont will be *nowhere* near 3.5 GHz at a mere 1.7W / core power budget. It'll be much closer to 2 GHz. Sierra Forest (6780E) is 144 cores @ 350W (2.2W / core) → a pitiful 2.2 GHz base clock. Let's go crazy and assume Darkmont magically achieves +13% higher clocks (2.2 → 2.5 GHz) at 22% less power (2.2W per core → 1.7W per core) and much higher IPC.
Darkmont @ hypothetical 2.5 GHz = ~5.09 points
Neoverse-V3 @ 3.2 GHz would be 61% faster.
>The only way how the claim of Arm can be true is if they have compared their new CPUs with antiquated CPUs like the Intel Granite Rapids Xeon CPUs, instead of comparing with state-of-the-art Intel Clearwater Forest and AMD Zen 5.
Intel had a paper announcement of Clearwater Forest this month. They have not revealed SKUs: no clocks, no model numbers, nothing exists. Nobody—including Arm—will be benchmarking against a CPU that doesn't exist on the market yet.