Compute Comparison
vs
All comparisons →

Ampere

NVIDIA A30

wins

1

The NVIDIA A30 is a data center Ampere GPU with 24GB HBM2 and NVLink support.

InferenceHPCMulti-instance GPU

Hopper

NVIDIA H200 SXM 141GB

wins

4

The H200 SXM is the SXM5 form factor of the H200, with 141 GB HBM3e and 4.8 TB/s bandwidth.

Very large LLMsMemory-bound inferenceMulti-modal models

Performance comparison

Visual bars — winner highlighted

NVIDIA A30

Metric

NVIDIA H200 SXM 141GB

FP16 Tensor TFLOPS (dense)

165 T
989 T

Bandwidth

933 GB/s
4.8k GB/s

VRAM

24 GB
141 GB

TDP (lower=better)

165 W
700 W

TFLOPS/watt

1.00
1.41

GB/s per watt

5.65
6.86

Full specification table

SpecNVIDIA A30NVIDIA H200 SXM 141GB
ArchitectureAmpereHopper
VRAM24 GB HBM2141 GB HBM3e
VRAM typeHBM2HBM3e
Memory bandwidth933 GB/s4,800 GB/s
FP16 Tensor TFLOPS (dense)165 TFLOPS989 TFLOPS
TDP165 W700 W
TFLOPS/watt1.00 T/W1.41 T/W
GB/s per watt5.65 GB/s/W6.86 GB/s/W
NVLinkYesYes
Release year20212024

Live cloud pricing

On-demand pricing across providers — updated in real time

Per hourPer GB VRAM/hr
Loading prices…

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

NVIDIA A30

FP16 TFLOPS165 TFLOPS
TDP165 W
TFLOPS/watt1.000 T/W
GB/s per watt5.65 GB/s/W

NVIDIA H200 SXM 141GB

FP16 TFLOPS989 TFLOPS
TDP700 W
TFLOPS/watt1.413 T/W
GB/s per watt6.86 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

When to choose each GPU

Choose NVIDIA A30 for:

  • Inference
  • HPC
  • Multi-instance GPU

Weaknesses / Limitations

  • 24 GB VRAM limits model size
  • Older Ampere architecture — lower TFLOPS than newer GPUs
  • Being superseded by L4 and A10 on most inference platforms

Choose NVIDIA H200 SXM 141GB for:

  • Very large LLMs
  • Memory-bound inference
  • Multi-modal models

Weaknesses / Limitations

  • Very limited cloud availability — fewer providers than H100
  • Premium pricing over H100 with same FP16 TFLOPS
  • 700 W TDP — same power envelope as H100 with no efficiency gain