Flopper.io
Flopper.io GPU Comparison Sheet

AWS Trainium2 vs NVIDIA H100 SXM5 80GB

Side by side datacenter GPU specifications from the Flopper.io database.

Overview

AWS Trainium2NVIDIA H100 SXM5 80GB
FP32 TFLOPS181.067.0
Memory96 GB80 GB
Memory Bandwidth2.9 TB/s3.4 TB/s
TDP—700 W

Performance by Precision

PrecisionAWS Trainium2 peak TFLOPS (dense)NVIDIA H100 SXM5 80GB peak TFLOPS (dense)
FP64 —34.0
FP64_TC Tensor Core—67.0
FP32 181.067.0
TF32 667.0494.0
BF16 667.0990.0
FP16 667.0990.0
FP8 1299.01979.0
FP6 ——
FP4 ——
INT8 —1979.0

Specifications

AWS Trainium2NVIDIA H100 SXM5 80GB
ArchitectureNeuronCore-v3Hopper
Process Node—4nm
Launch Year20242022
Form Factor—SXM
Memory Type——
Interface Width——
InterconnectNeuronLink-v3—
TDP—700 W
FP32 Efficiency—0.10 TFLOPS/W