NVIDIA V100 PCIe 16GB vs NVIDIA V100 SXM2 32GB

The V100 SXM2 32GB delivers 12% more FP16 throughput than the V100 PCIe 16GB (125 vs 112 TFLOPS dense). The V100 SXM2 32GB also carries 16 GB more memory (32 GB vs 16 GB).

V100 PCIe 16GB: Volta, 2017 V100 SXM2 32GB: Volta, 2018
FP16 dense lead
+12%
V100 SXM2 32GB: 125 vs 112 TFLOPS
VRAM
16 vs 32 GB
16 GB more for the V100 SXM2 32GB
Bandwidth
900 GB/s vs 900 GB/s
Even
TDP
250 W vs 300 W
V100 PCIe 16GB draws 50 W less

Specifications

V100 PCIe 16GB
V100 SXM2 32GB
Architecture
VoltaVolta
Launch Year
20172018
Form Factor
PCIeSXM
VRAM
16 GB32 GB
Memory Bandwidth
900 GB/s900 GB/s
TDP
250 W300 W
Process Node
12nm12nm

Performance (TFLOPS)

V100 PCIe 16GB
V100 SXM2 32GB
FP64
7.0 TFlops 7.8 TFlops
FP32
14.0 TFlops 15.7 TFlops
FP16
112.0 TFlops 125.0 TFlops

FLOPS by Precision

What actually differs

The V100 SXM2 32GB is the newer part: Volta, launched in 2018, against the V100 PCIe 16GB's Volta from 2017. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.

Dense throughput for the V100 PCIe 16GB against the V100 SXM2 32GB: FP64 7 vs 7.8 TFLOPS, FP32 14 vs 16 TFLOPS, FP16 112 vs 125 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 16 GB against 32 GB, fed at 900 GB/s versus 900 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 250 W for the V100 PCIe 16GB and 300 W for the V100 SXM2 32GB. At FP32 that works out to 0.06 against 0.05 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Is the V100 PCIe 16GB faster than the V100 SXM2 32GB?

At FP16 precision the V100 SXM2 32GB reaches 125 TFLOPS dense against 112 TFLOPS for the V100 PCIe 16GB. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the V100 PCIe 16GB or the V100 SXM2 32GB?

The V100 SXM2 32GB carries 32 GB of VRAM versus 16 GB for the V100 PCIe 16GB. Memory bandwidth is 900 GB/s for the V100 PCIe 16GB and 900 GB/s for the V100 SXM2 32GB.

How much power do the V100 PCIe 16GB and the V100 SXM2 32GB draw?

The V100 PCIe 16GB is rated at 250 W TDP and the V100 SXM2 32GB at 300 W. On FP32 throughput per watt, the V100 PCIe 16GB is the more efficient part.

Can I rent the V100 PCIe 16GB or the V100 SXM2 32GB in the cloud?

Yes. Live cloud listings tracked by Flopper start at $0.19 per GPU hour. The rental pricing section on this page lists current providers and rates for both GPUs.

Where to Rent

NVIDIA V100 PCIe 16GB

ProviderConfigurationPrice/GPU-hrChecked
RunPod
1× V100 PCIe Community $0.19 2h ago View →
All V100 PCIe 16GB listings and price history →

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.