NVIDIA A10 vs NVIDIA H100 SXM Full Specs

Summary

9,216 Shaders 1.7GHz	16,896 Shaders 1.98GHz
24GB GDDR6300.3GB/s	96GB HBM33.35TB/s
Peak·FP32·FP16 31.24 TFLOPS	Peak·FP32·FP16 66.91 TFLOPS
Form Factor PCIe Card	Form Factor SXM5
TDP 150W	TDP 700W
Power Connectors 1x 8-Pin	Power Connectors -

Highlights

Peak OPS

Peak OPS FP32 FP16 Tensor FP8-16 FP8-16 Tensor Sparse Tensor FP8-32 FP8-32 Tensor Sparse Tensor FP16-16 FP16-16 Tensor Sparse Tensor FP16-32 FP16-32 Tensor Sparse FP64 Tensor FP64 BF16 Tensor BF16 BF16 Tensor Sparse Tensor TF32 Tensor INT8 Tensor INT4 Memory Bandwidth TDP

A10999.8 TOPSINT4 Tensor Sparse

x1

H100 SXM3.96 POPSINT8 Tensor Sparse

x3.96

Show Top Score

Theoretical Performance

Clock Speed1695MHz UC·Base·Boost·OC	Clock Speed1980MHz UC·Base·Boost·OC
Peak OPS 999.8 TOPSINT4 Tensor Sparse	Peak OPS 3.96 POPSINT8 Tensor Sparse
-	Tensor FP8-16 1.98 PFLOPS FP8-16 Tensor Sparse 3.96 PFLOPS Tensor FP8-32 1.98 PFLOPS FP8-32 Tensor Sparse 3.96 PFLOPS
Tensor FP16-16 125 TFLOPS FP16-16 Tensor Sparse 249.9 TFLOPS Tensor FP16-32 62.48 TFLOPS FP16-32 Tensor Sparse 125 TFLOPS	Tensor FP16-16 989.4 TFLOPS FP16-16 Tensor Sparse 1.98 PFLOPS Tensor FP16-32 989.4 TFLOPS FP16-32 Tensor Sparse 1.98 PFLOPS
BF16 31.24 TFLOPS Tensor BF16 62.48 TFLOPS BF16 Tensor Sparse 125 TFLOPS	BF16 133.8 TFLOPS Tensor BF16 989.4 TFLOPS BF16 Tensor Sparse 1.98 PFLOPS
Tensor TF32 31.24 TFLOPS	Tensor TF32 494.7 TFLOPS
FP32 31.24 TFLOPS	FP32 66.91 TFLOPS
FP64 488.2 GFLOPS Tensor FP64 -	FP64 33.45 TFLOPS Tensor FP64 61.84 TFLOPS
Tensor INT4 499.9 TOPS	Tensor INT4 -
Tensor INT8 249.9 TOPS	Tensor INT8 1.98 POPS
Ray 47.12 TOPS	Ray -
Pixel Rate 162.7 GPixel/s	Pixel Rate 47.5 GPixel/s
Texture Rate 488.2 GTexel/s	Texture Rate 1.04 TTexel/s

Core Config

Shaders 9,216 Shaders	Shaders 16,896 Shaders
TMUs 288 TMUs	TMUs 528 TMUs
ROPs 96 ROPs	ROPs 24 ROPs
Tensor Cores 288 T-Cores	Tensor Cores 528 T-Cores
RT Cores 72 RT-Cores	RT Cores -
SMs 72 SMs	SMs 132 SMs

Clocks

Base Clock 885MHz	Base Clock 1.06GHz
Boost Clock 1.7GHz	Boost Clock 1.98GHz
Tensor Clock -	Tensor Clock 1.83GHz

Cache

L2 Cache 6.1MB shared

L2 Cache 51.2MB shared

Memory

24GB GDDR6	96GB HBM3
Memory Bus 384-bit	Memory Bus 5120-bit
Memory Speed 6.3GT/s	Memory Speed 5.2GT/s
Memory Bandwidth 300.3GB/s	Memory Bandwidth 3.35TB/s

Power & Thermals

TDP 150W

TDP 700W

Encoder

Encoder Model NVENC 7

-

Decoder

Decoder Model NVDEC 5

Decoder Model 7x NVDEC 5

Card

Form Factor PCIe Card	Form Factor SXM5
PCIe 1-Slots	PCIe -
Height 112 mm (4.41") Width 267 mm (10.51") Depth 20 mm (0.79")	-
Cooling Passive	Cooling Passive
Power Connectors 1x 8-Pin	Power Connectors -

Chip

Manufacturer NVIDIA	Manufacturer NVIDIA
Chip Designer NVIDIA	Chip Designer NVIDIA
Architecture Ampere	Architecture Hopper
Family Server	Family Server
Branding	Branding
Codename NV172	Codename NV180
Chip Variant GA102-200-KD-A1	Chip Variant -
Market Segment Server	Market Segment Server
Release Date Apr 12, 2021	Release Date Mar 22, 2022

Fabrication

Foundry Samsung	Foundry TSMC
Fabrication Node 8N	Fabrication Node 4N
Die Size 628mm²	Die Size 814mm²
Transistor Count 28.3 Billion	Transistor Count 80 Billion
Transistor Density 45.04 MTr/mm²	Transistor Density 98.28 MTr/mm²

Images

No images available

No images available

Competitors

NVIDIAA10 vs GPU NVIDIARTX 4000 Ada NVIDIAA10 vs GPU NVIDIARTX 4500 Ada NVIDIAA10 vs GPU AMDRadeon RX 9060 XT 16GB Show 12 more

NVIDIAH100 SXM vs GPU NVIDIAB100 NVIDIAH100 SXM vs GPU NVIDIAB200 NVIDIAH100 SXM vs GPU NVIDIAGB200 Show 12 more