GPU

NVIDIA GeForce RTX 5050

Edit@8 days ago

Intergrated Memory(VRAM)
Capacity

8 GB

(GDDR6 128-bit)

Bandwidth

224 GB/s

32 Token/s

Vector Compute
FP64
201.60 G
FP32
12.90 T
FP16
12.90 T
BF16
12.90 T
INT32
12.90 T
INT8
X

NVIDIA GeForce RTX 5050 General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 201.60 GFLOPS

FP32: 12.90 TFLOPS

FP16: 12.90 TFLOPS

BF16: 12.90 TFLOPS

INT32: 12.90 TOPS

Matirx Compute
FP64
X
FP32
X
FP16
25.80 T
51.61 T
FP8
51.61 T
103.22 T
TF32
12.90 T
25.80 T
BF16
25.80 T
51.61 T
INT16
X
INT8
103.22 T
206.44 T
INT4
X

NVIDIA GeForce RTX 5050 AI performance (Tensor Performance / Matrix Performance)

FP16: 25.80 TFLOPS, with sparsity: 51.61 TFLOPS

FP8: 51.61 TFLOPS, with sparsity: 103.22 TFLOPS

TF32: 12.90 TFLOPS, with sparsity: 25.80 TFLOPS

BF16: 25.80 TFLOPS, with sparsity: 51.61 TFLOPS

INT8: 103.22 TOPS, with sparsity: 206.44 TOPS

Hardware Specs
NVIDIA GeForce RTX 5050 is a 5nm chip, launched by NVIDIA at 2025. It has 8 GB built-in(On-Board/On-Chip) memory with bandwidth up to 224 GB/s. It has 2560 general-purpose ALUs(CUDA cores/Shader cores) and 80 matrix cores(Tensor cores) .
Process Node
5 nm
Launch Year
2025
Vector(CUDA) Cores
2560
Matrix(Tensor) Cores
80
Core Frequency
2235 ~ 2520 MHz
Cache
32MB