AMD Instinct MI250 vs AMD Radeon PRO W6400 Full Specs

Summary

13,312 Shaders 1.7GHz	768 Shaders 2.33GHz
128GB HBM2e3.28TB/s	4GB GDDR6128GB/s
Peak·FP32·FP16 45.26 TFLOPS	Peak·FP32·FP16 3.58 TFLOPS
Form Factor OAM Module	Form Factor PCIe Card
TDP 500W	TDP 50W

Highlights

Peak OPS

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

Instinct MI250362.1 TOPSINT4 Tensor

x50.56

Radeon PRO W64007.16 TFLOPSFP16

x1

Show Top Score

Theoretical Performance

Clock Speed1700MHz UC·Base·Boost·OC	Clock Speed2331MHz UC·Boost·OC
Peak OPS 362.1 TOPSINT4 Tensor	Peak OPS 7.16 TFLOPSFP16
Tensor FP16-16 362.1 TFLOPS Tensor FP16-32 362.1 TFLOPS	-
Tensor BF16 362.1 TFLOPS	-
FP32 45.26 TFLOPS Tensor FP32 90.52 TFLOPS	FP32 3.58 TFLOPS Tensor FP32 -
FP64 45.26 TFLOPS Tensor FP64 90.52 TFLOPS	FP64 223.8 GFLOPS Tensor FP64 -
Tensor INT4 362.1 TOPS	Tensor INT4 -
Tensor INT8 362.1 TOPS	-
Pixel Rate -	Pixel Rate 74.6 GPixel/s
Texture Rate 1.41 TTexel/s	Texture Rate 111.9 GTexel/s

Core Config

Shaders 13,312 Shaders	Shaders 768 Shaders
TMUs 832 TMUs	TMUs 48 TMUs
ROPs -	ROPs 32 ROPs
Tensor Cores 832 T-Cores	Tensor Cores -
RT Cores -	RT Cores 12 RT-Cores
CUs 208 CUs	CUs 12 CUs

Clocks

Base Clock 1GHz	Base Clock -
Boost Clock 1.7GHz	Boost Clock 2.33GHz

Cache

L0 Cache -	L0 Cache 32KB/WGP
L1 16KB/CU	L1 128KB/Array
L2 Cache 16MB shared	L2 Cache 1MB shared
L3 Cache -	L3 Cache 16MB shared
L3 Bandwidth -	L3 Bandwidth 416GB/s

Memory

128GB HBM2e	4GB GDDR6
Memory Bus 8192-bit	Memory Bus 64-bit
Memory Speed 3.2GT/s	Memory Speed 16GT/s
Memory Bandwidth 3.28TB/s	Memory Bandwidth 128GB/s
ECC Yes	ECC No

Power & Thermals

TDP 500W

TDP 50W

Video Output

Max Resolution -	Max Resolution 7680x4320
Refresh Rate Calculator -	Refresh Rate Calculator 7680x4320 Resolution8K UHD Refresh Rate 120Hz
Multi-Monitor -	Multi-Monitor 2
Adaptive Sync -	Adaptive Sync FreeSync
DSC Not Supported	DSC Supported
HDCP -	HDCP HDCP 2.3

Ports

-	2x DisplayPort 1.4

API Support

Shader Model -	Shader Model 6.6
-	DirectX DirectX 12 Direct3D 12_2
OpenGL - OpenCL 3 Vulkan -	OpenGL 4.6 OpenCL 2.2 Vulkan 1.3
GFX 9.4	GFX 10.3

Encoder

Encoder 2x VCN 2.6	-
Codec AVC (H.264) HEVC (H.265)	Codec -

Decoder

Decoder 2x VCN 2.6	Decoder VCN 3.0
Codec -	Codec MPEG-1 MPEG-2 MPEG-4 JPEG VC-1 VP9 AVC (H.264) HEVC (H.265) AV1

Card

Form Factor OAM Module	Form Factor PCIe Card
PCIe PCIe 4.0 x16 -	PCIe PCIe 4.0 x4 1-Slots
Cooling Passive -	Cooling Blower 1x Fan
Multi-GPU Multi-GPU Support Supported Multi-GPU Type Infinity Fabric	Multi-GPU Multi-GPU Support Supported Multi-GPU Type Bridgeless

Chip

Manufacturer AMD	Manufacturer AMD
Chip Designer AMD	Chip Designer AMD
Architecture CDNA 2	Architecture RDNA 2
Family Instinct	Family Radeon PRO W
Branding	Branding
Codename Aldebaran	Codename Beige Goby
Chip Variant Aldebaran XT	Chip Variant Navi 24 XL
Market Segment Server	Market Segment Workstation
Release Date Nov 8, 2021	Release Date Jan 19, 2022

Fabrication

Foundry TSMC	Foundry TSMC
Fabrication Node N6	Fabrication Node N6
Die Size 1448mm²	Die Size 107mm²
Transistor Count 56B	Transistor Count 5.4B
Transistor Density 38.67 MTr/mm²	Transistor Density 50.47 MTr/mm²

Images