AMD Radeon Instinct MI300X vs NVIDIA A10 PCIe

Comparison of AMD Radeon Instinct MI300X with 192 GB HBM3 and 19,456 cores vs NVIDIA A10 PCIe with 24 GB GDDR6 and 9,216 cores.

Loading...

Performance Rating

AMD Radeon Instinct MI300X outperforms NVIDIA A10 PCIe by 455.82% in the overall GPU ARK performance rating

A100 A100
H200 H200
MI325X MI325X

AMD Radeon Instinct MI300X

94.1

AMD Radeon Instinct MI300X

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

NVIDIA A10 PCIe

16.9

NVIDIA A10 PCIe

16.9

Expert Comparison

AMD Radeon Instinct MI300X значительно превосходит NVIDIA A10 PCIe в вычислительной мощности благодаря большему количеству ядер и более высокой пропускной способности памяти. MI300X больше подходит для тяжелых вычислительных задач, таких как машинное обучение и научные вычисления. В то время как A10 больше ориентирован на графические задачи и менее эффективен в вычислительных нагрузках.

Contents:

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

Memory

Memory Size

🔥 +700% 192 ГБ
24 ГБ

Memory Type

HBM3 GDDR6

Memory Bandwidth

🔥 10.3 TB/s
600.2 GB/s

Memory Bus Width

8,192 бит 384 бит

ML Performance

FP16 (Half Precision)

🔥 +1,993% 653.7 TFLOPS
31.24 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +162% 81.72 TFLOPS
31.24 TFLOPS

FP64 (Double Precision)

🔥 +8,270% 81.72 TFLOPS
0.9763 TFLOPS

CUDA Cores

🔥 +111% 19,456
9,216

RT Cores

No
🔥 72

Architecture & Compatibility

GPU Architecture

CDNA 3.0 Ampere

SM (Streaming Multiprocessor)

No
🔥 72

PCIe Version

PCIe 5.0 x16 PCIe 4.0 x16

ML Software Support

CUDA Version

No 8.6

Clocks & Performance

Base Clock

🔥 +13% 1,000
885

Boost Clock

🔥 +24% 2,100
1,695

Memory Clock

🔥 +62% 2,525
1,563

Power Consumption

TDP/TGP

750 W
🔥 -80% 150 W

Recommended PSU

1150 W
🔥 -61% 450 W

Power Connector

None 1x 8-pin

Rendering

Texture Units (TMU)

🔥 +322% 1,216
288

ROP

No
🔥 72

L2 Cache

16 MB
🔥 6 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 Single-slot

Release Date

Dec. 6, 2023 April 12, 2021

Display Outputs

No outputs
No outputs

Renting is cheaper than buying