AMD Radeon RX 7900M vs NVIDIA GeForce RTX 5090

Comparison of AMD Radeon RX 7900M with 16 GB GDDR6 and 4,608 cores vs NVIDIA GeForce RTX 5090 with 32 GB GDDR7 and 21,760 cores.

Loading...

Performance Rating

H200 H200
MI325X MI325X
A100 A100

AMD Radeon RX 7900M

AMD Radeon RX 7900M

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

NVIDIA GeForce RTX 5090

NVIDIA GeForce RTX 5090

Contents:

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

Memory

Memory Size

16 GB 32 GB

Memory Type

GDDR6 GDDR7

Memory Bandwidth

576.0 GB/s 1.79 TB/s

Memory Bus Width

256 бит 512 бит

ML Performance

FP16 (Half Precision)

77.05 TFLOPS 104.8 TFLOPS

BF16 (Brain Float)

No No

TF32 (TensorFloat)

No No

Compute Power

FP32 (Single Precision)

38.52 TFLOPS 104.8 TFLOPS

FP64 (Double Precision)

1.2038 TFLOPS 1.637 TFLOPS

CUDA Cores

4,608 21,760

RT Cores

72 170

Architecture & Compatibility

GPU Architecture

RDNA 3.0 Blackwell 2.0

SM (Streaming Multiprocessor)

No 170

PCIe Version

PCIe 4.0 x16 PCIe 5.0 x16

ML Software Support

CUDA Version

No 12.0

Clocks & Performance

Base Clock

1,825 2,017

Boost Clock

2,090 2,407

Memory Clock

2,250 1,750

Power Consumption

TDP/TGP

🔥 -69% 180 W
575 W

Recommended PSU

No 950 W

Power Connector

None 1x 16-pin

Rendering

Texture Units (TMU)

288 680

ROP

72 170

L2 Cache

6 MB 96 MB

Benchmarks

llama.cpp, llama 7B Q4_0

315.3 tokens/s

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

273.7 tokens/s

Geekbench AI, FP16

69 765 points

Geekbench AI, INT8

34 585 points

Geekbench AI, FP32

46 834 points

Additional

Slots

IGP Dual-slot

Release Date

Oct. 19, 2023 Jan. 30, 2025

Display Outputs

Portable Device Dependent
1x HDMI 2.1b
3x DisplayPort 2.1b

Renting is cheaper than buying