Intel Data Center GPU Max 1100 vs NVIDIA H200 SXM 141 GB

Comparison of Intel Data Center GPU Max 1100 with 48 GB HBM2e and 7,168 cores vs NVIDIA H200 SXM 141 GB with 141 GB HBM3e and 16,896 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

Intel Data Center GPU Max 1100

24.4

Intel Data Center GPU Max 1100

24.4
RX 7900 XTX RX 7900 XTX
Instinct MI300X Instinct MI300X
MI250 MI250

NVIDIA H200 SXM 141 GB

NVIDIA H200 SXM 141 GB

Contents:

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

Memory

Memory Size

🔥 48 ГБ
141 GB

Memory Type

HBM2e HBM3e

Memory Bandwidth

🔥 1.23 TB/s
4.89 TB/s

Memory Bus Width

8,192 бит 6,144 бит

ML Performance

FP16 (Half Precision)

🔥 22.22 TFLOPS
267.6 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

🔥 22.22 TFLOPS
66.91 TFLOPS

FP64 (Double Precision)

🔥 22.22 TFLOPS
33.45 TFLOPS

CUDA Cores

🔥 7,168
16,896

RT Cores

🔥 56
No

Architecture & Compatibility

GPU Architecture

Generation 12.5 Hopper

SM (Streaming Multiprocessor)

No 132

PCIe Version

PCIe 5.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

No 9.0

Clocks & Performance

Base Clock

🔥 1,000
1,500

Boost Clock

🔥 1,550
1,980

Memory Clock

🔥 600
1,593

Power Consumption

TDP/TGP

🔥 -57% 300 W
700 W

Recommended PSU

🔥 -36% 700 W
1100 W

Power Connector

1x 12-pin 8-pin EPS

Rendering

Texture Units (TMU)

🔥 448
528

ROP

🔥 56
No

L2 Cache

🔥 +308% 204 MB
50 MB

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

Dual-slot SXM Module

Release Date

Jan. 10, 2023 Nov. 18, 2024

Display Outputs

No outputs
No outputs

Renting is cheaper than buying