NVIDIA H200 SXM 141 GB vs NVIDIA ION

Comparison of NVIDIA H200 SXM 141 GB with 141 GB HBM3e and 16,896 cores vs NVIDIA ION and 16 cores.

Loading...

Performance Rating

A100 A100
H200 H200
MI325X MI325X

NVIDIA H200 SXM 141 GB

67.4

NVIDIA H200 SXM 141 GB

67.4
RX 7900 XTX RX 7900 XTX
MI250 MI250
Instinct MI300X Instinct MI300X

NVIDIA ION

NVIDIA ION

Contents:

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

Memory

Memory Size

🔥 141 ГБ
No

Memory Type

HBM3e System Shared

Memory Bandwidth

🔥 4.89 TB/s
System Dependent

Memory Bus Width

6,144 бит No

ML Performance

FP16 (Half Precision)

🔥 267.6 TFLOPS
No

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 +189,985% 66.91 TFLOPS
0.0352 TFLOPS

FP64 (Double Precision)

🔥 33.45 TFLOPS
No

CUDA Cores

🔥 +105,500% 16,896
16

RT Cores

No No

Architecture & Compatibility

GPU Architecture

Hopper Tesla

SM (Streaming Multiprocessor)

🔥 +13,100% 132
1

PCIe Version

PCIe 5.0 x16 PCI

ML Software Support

CUDA Version

9.0 No

Clocks & Performance

Base Clock

🔥 1,500
No

Boost Clock

🔥 1,980
No

Memory Clock

🔥 1,593
No

Power Consumption

TDP/TGP

700 W
🔥 -97% 20 W

Recommended PSU

1100 W No

Power Connector

8-pin EPS No

Rendering

Texture Units (TMU)

🔥 +6,500% 528
8

ROP

No No

L2 Cache

🔥 50 MB
No

Benchmarks

MLPerf, llama2-70b-99.9 (UNSET)

3 534 tokens/s

MLPerf, llama2-70b-99.9 (fp16)

3 553 tokens/s

MLPerf, llama2-70b-99.9 (fp8)

2 444 tokens/s

MLPerf, llama3.1-405b (fp16)

40.8 tokens/s

MLPerf, llama3.1-405b (fp8)

25.3 tokens/s

MLPerf, llama3.1-8b (fp8)

5 161 tokens/s

MLPerf, deepseek-r1 (fp8)

1 113 tokens/s

MLPerf, mixtral-8x7b (fp8)

7 132 tokens/s

Additional

Slots

🔥 SXM Module
IGP

Release Date

Nov. 18, 2024 June 3, 2008

Display Outputs

No outputs
Portable Device Dependent

Renting is cheaper than buying