AMD Radeon Instinct MI300X vs NVIDIA GeForce RTX 3080 Mobile

Comparison of AMD Radeon Instinct MI300X with 192 GB HBM3 and 19,456 cores vs NVIDIA GeForce RTX 3080 Mobile with 8 GB GDDR6 and 6,144 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

AMD Radeon Instinct MI300X

94.1

AMD Radeon Instinct MI300X

94.1
MI250 MI250
Instinct MI300X Instinct MI300X
RX 7900 XTX RX 7900 XTX

NVIDIA GeForce RTX 3080 Mobile

NVIDIA GeForce RTX 3080 Mobile

Contents:

Memory ML Performance Compute Power Architecture & Compatibility ML Software Support Clocks & Performance Power Consumption Rendering Benchmarks Additional

Memory

Memory Size

🔥 +2,300% 192 ГБ
8 GB

Memory Type

HBM3 GDDR6

Memory Bandwidth

🔥 10.3 TB/s
448.0 GB/s

Memory Bus Width

8,192 бит 256 бит

ML Performance

FP16 (Half Precision)

🔥 +3,344% 653.7 TFLOPS
18.98 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +331% 81.72 TFLOPS
18.98 TFLOPS

FP64 (Double Precision)

🔥 +27,452% 81.72 TFLOPS
0.2966 TFLOPS

CUDA Cores

🔥 +217% 19,456
6,144

RT Cores

No 48

Architecture & Compatibility

GPU Architecture

CDNA 3.0 Ampere

SM (Streaming Multiprocessor)

No 48

PCIe Version

PCIe 5.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

No 8.6

Clocks & Performance

Base Clock

🔥 1,000
1,110

Boost Clock

🔥 +36% 2,100
1,545

Memory Clock

🔥 +44% 2,525
1,750

Power Consumption

TDP/TGP

750 W
🔥 -85% 115 W

Recommended PSU

1150 W No

Power Connector

None None

Rendering

Texture Units (TMU)

🔥 +533% 1,216
192

ROP

No 48

L2 Cache

🔥 +300% 16 MB
4 MB

Benchmarks

MLPerf, llama2-70b-99.9 (UNSET)

1 983 tokens/s

MLPerf, llama2-70b-99.9 (fp16)

1 740 tokens/s

MLPerf, llama2-70b-99.9 (fp8)

1 057 tokens/s

MLPerf, llama3.1-405b (UNSET)

30.4 tokens/s

MLPerf, llama3.1-405b (fp16)

34.8 tokens/s

llama.cpp, llama-2-7b-Q4_0

232.9 tokens/s

MLPerf, mixtral-8x7b (fp8)

5 975 tokens/s

Additional

Slots

OAM Module No

Release Date

Dec. 6, 2023 Jan. 12, 2021

Display Outputs

No outputs
Portable Device Dependent

Renting is cheaper than buying