CPU Performance Benchmark — NVIDIA Vera vs x86
Performance Benchmark

NVIDIA Vera CPU vs x86 — Who Runs AI Infrastructure

Single-socket configuration · TDP 450W · Linux kernel source compile & geometric mean
Linux Kernel Compile Time (lower = faster)
NVIDIA Vera 88 Olympus Cores · 450W
20s
128-core x86 baseline Best competing server CPU
40s
Vera's 88 cores outpace a 128-core competitor by 2×. Per-core efficiency advantage: 4×.
Overall Benchmark Geometric Mean (higher = better, Vera = 100%)
NVIDIA Vera 88 Olympus Cores
100%
AMD EPYC 9575F 64 Cores · Zen 5
90.9%
Intel Xeon 6980P 128 Cores
64.5%
Vera leads AMD by +10% and Intel by +55%. Intel's 128-core chip with fewer than half the cores falls behind by a wide margin.
Power Efficiency Breakdown
Best
NVIDIA Vera
450
W compute TDP
Memory bus: <30W
LPDDR5X SOCAMM2
~2× efficiency lead
AMD EPYC
500+
W compute TDP
Memory bus: ~70W+
DDR5 multi-channel
Baseline
Intel Xeon
500+
W compute TDP
Memory bus: 100W+
12-ch DDR5
Below baseline