AMD Instinct MI300A vs Intel Gaudi 3 Full Specs

Summary

14,592 Shaders 2.1GHz	16,384 Shaders 438MHz
128GB HBM35.33TB/s	128GB HBM2e3.7TB/s
Peak·FP32·FP16 122.6 TFLOPS	Peak·FP32·FP16 -
Form Factor APU SH5 Socket	Form Factor OAM Module
TDP 560W	TDP 900W

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 Tensor FP32 FP64 Tensor FP64 Tensor BF16 BF16 Tensor Sparse Tensor TF32 Tensor INT8 Tensor INT4 Memory Bandwidth TDP

Instinct MI300A3.92 POPSINT4 Tensor Sparse

x2.14

Gaudi 31.84 PFLOPSFP8-32 Tensor

x1

Show Top Score

Theoretical Performance

Clock Speed2100MHz UC·Base·Boost·OC	Clock Speed438MHz UC·Boost·OC
Peak OPS 3.92 POPSINT4 Tensor Sparse	Peak OPS 1.84 PFLOPSFP8-32 Tensor
Tensor FP8-16 1.96 PFLOPS FP8-16 Tensor Sparse 3.92 PFLOPS Tensor FP8-32 1.96 PFLOPS FP8-32 Tensor Sparse 3.92 PFLOPS	Tensor FP8-16 - FP8-16 Tensor Sparse - Tensor FP8-32 1.84 PFLOPS FP8-32 Tensor Sparse -
Tensor FP16-16 980.6 TFLOPS FP16-16 Tensor Sparse 1.96 PFLOPS Tensor FP16-32 980.6 TFLOPS FP16-32 Tensor Sparse 1.96 PFLOPS	Tensor FP16-16 - FP16-16 Tensor Sparse - Tensor FP16-32 459.3 TFLOPS FP16-32 Tensor Sparse -
Tensor BF16 980.6 TFLOPS BF16 Tensor Sparse 1.96 PFLOPS	Tensor BF16 1.84 PFLOPS BF16 Tensor Sparse -
Tensor TF32 490.3 TFLOPS	Tensor TF32 459.3 TFLOPS
FP32 122.6 TFLOPS Tensor FP32 122.6 TFLOPS	FP32 - Tensor FP32 229.6 TFLOPS
FP64 61.29 TFLOPS Tensor FP64 122.6 TFLOPS	-
Tensor INT4 1.96 POPS	Tensor INT4 -
Tensor INT8 1.96 POPS	-
Texture Rate 1.92 TTexel/s	Texture Rate -

Core Config

Shaders 14,592 Shaders	Shaders 16,384 Shaders
TMUs 912 TMUs	TMUs -
Tensor Cores 912 T-Cores	Tensor Cores 64 T-Cores
CUs 228 CUs	CUs 8 CUs

Clocks

Base Clock 1GHz	Base Clock -
Boost Clock 2.1GHz	Boost Clock 438MHz

Cache

L2 Cache 16.4MB shared	L2 Cache -
L3 Cache 192MB shared	L3 Cache -

Memory

128GB HBM3	128GB HBM2e
Memory Bus 8192-bit	Memory Bus 8192-bit
Memory Speed 5.2GT/s	Memory Speed 3.6GT/s
Memory Bandwidth 5.33TB/s	Memory Bandwidth 3.7TB/s
ECC Yes	ECC Yes
ESRAM -	ESRAM 96MBSize 12800GB/sBandwidth

Power & Thermals

TDP 560W

TDP 900W

Encoder

Encoder Model 2x VCN 2.6

-

Decoder

Decoder Model 2x VCN 2.6

Decoder Model 14x Gaudi

Card

Form Factor APU SH5 Socket	Form Factor OAM Module
Cooling Passive	Cooling Passive

Chip

Manufacturer AMD	Manufacturer Intel
Chip Designer AMD	Chip Designer Intel
Architecture CDNA 3	Architecture Gaudi 3
Family Instinct	Family Gaudi
Branding	Branding
Codename Aqua Vanjaram	Codename Gaudi 3
Chip Variant Aqua Vanjaram XT	Chip Variant -
Market Segment Server	Market Segment Server
Release Date Dec 6, 2023	Release Date Apr 9, 2024

Fabrication

Foundry TSMC	Foundry TSMC
Other Foundries TSMCActive Interposer	Other Foundries -
Fabrication Node N5 N6Active Interposer	Fabrication Node N5 -
Die Size 1110mm²	Die Size -
Transistor Count 109.5 Billion	Transistor Count -
Transistor Density 98.65 MTr/mm²	Transistor Density -

Images

No images available

No images available

Competitors

AMDInstinct MI300A vs GPU AMDInstinct MI300X AMDInstinct MI300A vs GPU AMDInstinct MI250X AMDInstinct MI300A vs GPU AMDInstinct MI325X Show 12 more

IntelGaudi 3 vs GPU AMDInstinct MI250X IntelGaudi 3 vs GPU AMDInstinct MI250 IntelGaudi 3 vs GPU AMDInstinct MI300X Show 12 more