Home NVIDIA GeForce RTX 5090

NVIDIA GeForce RTX 5090

This is a GPU manufactured with TSMC 4nm process, based on Nvidia Blackwell 2.0 architecture and released on 2025. It features 21760 CUDA cores and 32GB GDDR7 memory, with 88MB L2 cache, theoretical performance of 109.7TFLOPS, with total power consumption of 500W.

Graphics Card

[Report Issues]
Release Date
Jan 2025
Generation
GeForce 50
Type
Desktop
Bus Interface
PCIe 5.0 x16

Clock Speeds

[Report Issues]
Base Clock
2235 MHz
Boost Clock
2520 MHz
Memory Clock
1875 MHz

Memory

[Report Issues]
Memory Size
32GB
Memory Type
GDDR7
Memory Bus
512bit
Bandwidth
1.52TB/s

Render Config

[Report Issues]
SM Count
170
Shading Units
21760
TMUs
680
ROPs
192
Tensor Cores
680
RT Cores
170
L1 Cache
128 KB (per SM)
L2 Cache
88 MB

Theoretical Performance

[Report Issues]
Pixel Rate
483.8 GPixel/s
Texture Rate
1714 GTexel/s
FP16 (half)
109.7 TFLOPS
FP32 (float)
109.7 TFLOPS
FP64 (double)
1.714 TFLOPS

Graphics Processor

[Report Issues]
GPU Name
GB202
GPU Variant
GB202-400-A1
Architecture
Blackwell 2.0
Foundry
TSMC
Process Size
4 nm
Transistors
Unknown
Die Size
Unknown

Board Design

[Report Issues]
TDP
500W
Suggested PSU
900 W
Outputs
1x HDMI 2.13x DisplayPort 1.4a
Power Connectors
1x 16-pin

Graphics Features

[Report Issues]
DirectX
12 Ultimate (12_2)
OpenGL
4.6
OpenCL
3.0
Vulkan
1.3
CUDA
10.1
Shader Model
6.8

Rankings

[Report Issues]
FP32 (float)
163.4 TFLOPS
122.6 TFLOPS
109.7 TFLOPS
NVIDIA TITAN Ada 48 GB GDDR6X
92.9 TFLOPS
NVIDIA L40S 48 GB GDDR6
91.61 TFLOPS
91.06 TFLOPS
91.06 TFLOPS

Related Comparisons

Related News

© 2024 - TopCPU.net   Contact Us Privacy Policy