GPU

NVIDIA GeForce RTX 4080 SUPER

Edit@1 months ago

Intergrated Memory(VRAM)
Capacity

16 GB

(GDDR6X 256-bit)

Bandwidth

736 GB/s

105 Token/s

Vector Compute
FP64
816 G
FP32
52.22 T
FP16
52.22 T
BF16
52.22 T
INT32
26.11 T
INT8
X

NVIDIA GeForce RTX 4080 SUPER General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 816 GFLOPS

FP32: 52.22 TFLOPS

FP16: 52.22 TFLOPS

BF16: 52.22 TFLOPS

INT32: 26.11 TOPS

Matirx Compute
FP64
X
FP32
X
FP16
104.45 T
208.90 T
FP8
208.90 T
417.79 T
TF32
52.22 T
104.45 T
BF16
104.45 T
208.90 T
INT16
X
INT8
417.79 T
835.58 T
INT4
835.58 T
1671.17 T

NVIDIA GeForce RTX 4080 SUPER AI performance (Tensor Performance / Matrix Performance)

FP16: 104.45 TFLOPS, with sparsity: 208.90 TFLOPS

FP8: 208.90 TFLOPS, with sparsity: 417.79 TFLOPS

TF32: 52.22 TFLOPS, with sparsity: 104.45 TFLOPS

BF16: 104.45 TFLOPS, with sparsity: 208.90 TFLOPS

INT8: 417.79 TOPS, with sparsity: 835.58 TOPS

INT4: 835.58 TOPS, with sparsity: 1671.17 TOPS

Hardware Specs
NVIDIA GeForce RTX 4080 SUPER is a 5nm chip, has 45900 million transistors, launched by NVIDIA at 2024. It has 16 GB built-in(On-Board/On-Chip) memory with bandwidth up to 736 GB/s. It has 10240 general-purpose ALUs(CUDA cores/Shader cores) and 320 matrix cores(Tensor cores) .
Process Node
5 nm
Launch Year
2024
Vector(CUDA) Cores
10240
Matrix(Tensor) Cores
320
Core Frequency
2295 ~ 2550 MHz
Cache
64MB