AMD Instinct MI350P vs NVIDIA H200 SXM Full Specs

Summary

8,192 Shaders 2.2GHz	16,896 Shaders 1.98GHz
144GB HBM3e4TB/s	141GB HBM3e4.81TB/s
Peak·FP32·FP16 72.09 TFLOPS	Peak·FP32·FP16 66.91 TFLOPS
Form Factor PCIe Card	Form Factor SXM5
TDP 600W	TDP 700W
Power Connectors 1x 16-Pin 12V-2x6	Power Connectors -

Highlights

Peak OPS

Peak OPS FP32 FP16 Tensor MXFP4 Tensor MXFP6 Tensor MXFP8 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 Tensor FP32 FP64 Tensor FP64 BF16 Tensor BF16 BF16 Tensor Sparse Tensor TF32 TF32 Tensor Sparse Tensor INT8 INT8 Tensor Sparse Memory Bandwidth TDP

Instinct MI350P4.61 PFLOPSMXFP4 Tensor

x1.17

H200 SXM3.96 POPSINT8 Tensor Sparse

x1

Show Top Score

Theoretical Performance

Clock Speed2200MHz UC·Boost·OC	Clock Speed1980MHz UC·Base·Boost·OC
Peak OPS 4.61 PFLOPSMXFP4 Tensor	Peak OPS 3.96 POPSINT8 Tensor Sparse
MXFP4 Tensor 4.61 PFLOPS	-
MXFP6 Tensor 4.61 PFLOPS	-
Tensor FP8-16 - FP8-16 Tensor Sparse - Tensor FP8-32 2.31 PFLOPS FP8-32 Tensor Sparse 4.61 PFLOPS MXFP8 Tensor 2.31 PFLOPS	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 MXFP8 Tensor -
FP16 72.09 TFLOPS Tensor FP16-16 - FP16-16 Tensor Sparse - Tensor FP16-32 1.15 PFLOPS FP16-32 Tensor Sparse 2.31 PFLOPS	FP16 133.8 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 - Tensor BF16 1.15 PFLOPS BF16 Tensor Sparse 2.31 PFLOPS	BF16 133.8 TFLOPS Tensor BF16 989.4 TFLOPS BF16 Tensor Sparse 1.98 PFLOPS
-	Tensor TF32 494.7 TFLOPS TF32 Tensor Sparse 989.4 TFLOPS
FP32 72.09 TFLOPS Tensor FP32 72.09 TFLOPS	FP32 66.91 TFLOPS Tensor FP32 -
FP64 36.04 TFLOPS Tensor FP64 36.04 TFLOPS	FP64 33.45 TFLOPS Tensor FP64 61.84 TFLOPS
Tensor INT8 2.31 POPS INT8 Tensor Sparse 4.61 POPS	Tensor INT8 1.98 POPS INT8 Tensor Sparse 3.96 POPS
Pixel Rate -	Pixel Rate 47.5 GPixel/s
Texture Rate -	Texture Rate 1.04 TTexel/s

Core Config

Shaders 8,192 Shaders	Shaders 16,896 Shaders
TMUs -	TMUs 528 TMUs
ROPs -	ROPs 24 ROPs
Tensor Cores 512 T-Cores	Tensor Cores 528 T-Cores
CUs 128 CUs	SMs 132 SMs

Clocks

Base Clock -	Base Clock 1.06GHz
Boost Clock 2.2GHz	Boost Clock 1.98GHz
Tensor Clock -	Tensor Clock 1.83GHz

Cache

L1 32KB/CU	L1 64KB/SM Tex
L2 Cache 16MB shared	L2 Cache 50MB shared
L3 Cache 128MB shared	L3 Cache -

Memory

144GB HBM3e	141GB HBM3e
Memory Bus 4096-bit	Memory Bus 6144-bit
Memory Speed 7.8GT/s	Memory Speed 6.3GT/s
Memory Bandwidth 4TB/s	Memory Bandwidth 4.81TB/s
ECC Yes	ECC No

Power & Thermals

TDP 600W

TDP 700W

Video Output

Adaptive Sync -	Adaptive Sync G-Sync FreeSync
DSC Not Supported	DSC Not Supported

API Support

OpenCL 3	OpenCL 3
CUDA - GFX 9.5 PureVideo HD - VDPAU -	CUDA 9 GFX - PureVideo HD VP11 VDPAU Feature Set K

Decoder

Decoder 2x VCN 5.0	Decoder 7x NVDEC 5
Codec - - - JPEG - - 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 PCIe Card	Form Factor SXM5
PCIe PCIe 5.0 x16 2-Slots	PCIe PCIe 5.0 x16 -
Width 267mm (10.51")	-
Cooling Passive	Cooling Passive
Power Connectors 1x 16-Pin 12V-2x6	Power Connectors -
Multi-GPU Multi-GPU Support Supported Multi-GPU Type Infinity Fabric	Multi-GPU Multi-GPU Support Supported Multi-GPU Type NVLink

Chip

Manufacturer AMD	Manufacturer NVIDIA
Chip Designer AMD	Chip Designer NVIDIA
Architecture CDNA 4	Architecture Hopper
Family Instinct	Family Server
Branding	Branding
Codename CDNA Next	Codename NV180
Market Segment Server	Market Segment Server
Release Date Jul 23, 2026	Release Date Nov 13, 2023

Fabrication

Foundry TSMC	Foundry TSMC
Other Foundries TSMCIO	Other Foundries -
Fabrication Node N3P N6IO	Fabrication Node 4N -
Die Size 724mm²	Die Size 814mm²
Transistor Count 69.6B 23BIO	Transistor Count 80B -
Transistor Density 96.13 MTr/mm²	Transistor Density 98.28 MTr/mm²

Images