NVIDIA GB200 vs NVIDIA H100 Full Specs

Summary

37,888 Shaders 2.11GHz	14,592 Shaders 1.75GHz
384GB HBM3e16TB/s	80GB HBM2e2.04TB/s
Peak·FP32·FP16 160 TFLOPS	Peak·FP32·FP16 51.22 TFLOPS
Form Factor Superchip	Form Factor PCIe Card
TDP 2700W	TDP 350W
Power Connectors -	Power Connectors 1x 16-Pin 12VHPWR

Theoretical Performance

Clock Speed2112MHz UC·Boost·OC	Clock Speed1755MHz UC·Base·Boost·OC
Peak OPS 40 PFLOPSFP4 Tensor Sparse	Peak OPS 3.03 POPSINT8 Tensor Sparse
Tensor FP4 20 PFLOPS FP4 Tensor Sparse 40 PFLOPS	-
Tensor FP8-16 10 PFLOPS FP8-16 Tensor Sparse 20 PFLOPS Tensor FP8-32 10 PFLOPS FP8-32 Tensor Sparse 20 PFLOPS	Tensor FP8-16 1.51 PFLOPS FP8-16 Tensor Sparse 3.03 PFLOPS Tensor FP8-32 1.51 PFLOPS FP8-32 Tensor Sparse 3.03 PFLOPS
Tensor FP16-16 5 PFLOPS FP16-16 Tensor Sparse 10 PFLOPS Tensor FP16-32 5 PFLOPS FP16-32 Tensor Sparse 10 PFLOPS	Tensor FP16-16 756.4 TFLOPS FP16-16 Tensor Sparse 1.51 PFLOPS Tensor FP16-32 756.4 TFLOPS FP16-32 Tensor Sparse 1.51 PFLOPS
BF16 320.1 TFLOPS Tensor BF16 5 PFLOPS BF16 Tensor Sparse 10 PFLOPS	BF16 102.4 TFLOPS Tensor BF16 756.4 TFLOPS BF16 Tensor Sparse 1.51 PFLOPS
Tensor TF32 2.5 PFLOPS	Tensor TF32 378.2 TFLOPS
FP32 160 TFLOPS	FP32 51.22 TFLOPS
FP64 80.02 TFLOPS Tensor FP64 78.13 TFLOPS	FP64 25.61 TFLOPS Tensor FP64 47.28 TFLOPS
Tensor INT8 10 POPS	Tensor INT8 1.51 POPS
Pixel Rate 67.6 GPixel/s	Pixel Rate 42.1 GPixel/s
Texture Rate 1.25 TTexel/s	Texture Rate 800.3 GTexel/s

Core Config

Shaders 37,888 Shaders	Shaders 14,592 Shaders
TMUs 1,184 TMUs	TMUs 456 TMUs
ROPs 64 ROPs	ROPs 24 ROPs
Tensor Cores 1,184 T-Cores	Tensor Cores 456 T-Cores
SMs 296 SMs	SMs 114 SMs

Clocks

Base Clock -	Base Clock 1.06GHz
Boost Clock 2.11GHz	Boost Clock 1.75GHz
Tensor Clock 2.06GHz	Tensor Clock 1.62GHz

Cache

L2 Cache 64MB shared

L2 Cache 50MB shared

Memory

384GB HBM3e	80GB HBM2e
Memory Bus 8192-bit	Memory Bus 5120-bit
Memory Speed 7.8GT/s	Memory Speed 3.2GT/s
Memory Bandwidth 16TB/s	Memory Bandwidth 2.04TB/s
ECC No	ECC No

Power & Thermals

TDP 2700W

TDP 350W

Decoder

Decoder Model 7x NVDEC 6	Decoder Model 7x NVDEC 5
Codec MPEG-1 MPEG-2 MPEG-4 VC-1 VP8 VP9 AVC (H.264) HEVC (H.265) AV1	Codec MPEG-1 MPEG-2 MPEG-4 VC-1 VP8 VP9 AVC (H.264) HEVC (H.265) AV1

Card

Form Factor Superchip	Form Factor PCIe Card
PCIe -	PCIe 2-Slots
-	Height 111 mm (4.37") Width 268 mm (10.55") Depth 40 mm (1.57")
Cooling Passive	Cooling Passive
Power Connectors -	Power Connectors 1x 16-Pin 12VHPWR

Chip

Manufacturer NVIDIA	Manufacturer NVIDIA
Chip Designer NVIDIA	Chip Designer NVIDIA
Architecture Blackwell	Architecture Hopper
Family Server	Family Server
Branding	Branding
Codename NV190	Codename NV180
Chip Variant Umbriel	Chip Variant -
Market Segment Server	Market Segment Server
Release Date Mar 18, 2024	Release Date Mar 22, 2022

Fabrication

Foundry TSMC	Foundry TSMC
Fabrication Node 4NP	Fabrication Node 4N
Die Size 1620mm²	Die Size 814mm²
Transistor Count 208 Billion	Transistor Count 80 Billion
Transistor Density 128.4 MTr/mm²	Transistor Density 98.28 MTr/mm²

Images