Skip to content
HardwareLadder
AI GPU ladder

NVIDIAAI GPU · SXM4

A100 SXM 40GB

NVIDIA A100 SXM 40GB is a Ampere data-center accelerator with 40 GB HBM2 and 1.56 TB/s of memory bandwidth.

Ladder rank
#16
AI index (H100 SXM = 100)
38.3
Series
Ampere

Specifications

Dense throughput, no sparsity
Architecture Ampere
Form factor SXM4
Process TSMC 7 nm
FP16 / BF16 tensor (dense) 312 TFLOPS
FP8 tensor (dense) —
FP4 tensor (dense) —
Memory 40 GB HBM2
Memory bandwidth 1.56 TB/s
Power 400 W
Launch May 2020

AI index

Tensor × bandwidth

100 × √((312 ÷ 989.4) × (1.56 ÷ 3.35)) = 38.3

The AI index is the geometric mean of dense 16-bit tensor throughput and memory bandwidth, each relative to the H100 SXM. It is a spec-sheet comparison, not a measured benchmark.

Sources

1 references