FLOPS · Floating-Point Operations Per Second
The standard measure of compute throughput; AI clusters are rated in petaFLOPS and exaFLOPS.
Current numbers
37–66%MI300X realized inference performance vs H100/H200 despite ~1.3x paper FLOPS — the ROCm goodput gap
37–66%MI300X realized inference throughput vs H100/H200 (2025), despite ~1.3x paper FLOPS — the historical realized-MFU gap
~90% / ~96%goodput (effective training time): industry average vs best-in-class — the software-and-reliability tax on paper FLOPS