GPU

NVIDIA GeForce RTX 4060 Ti

Edit@2 months ago

Intergrated Memory(VRAM)
Capacity

16 GB

(GDDR6 128-bit)

Bandwidth

288 GB/s

41 Token/s

Vector Compute
FP64
344.80 G
FP32
22.06 T
FP16
22.06 T
BF16
22.06 T
INT32
11.03 T
INT8
X

NVIDIA GeForce RTX 4060 Ti General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 344.80 GFLOPS

FP32: 22.06 TFLOPS

FP16: 22.06 TFLOPS

BF16: 22.06 TFLOPS

INT32: 11.03 TOPS

Matirx Compute
FP64
X
FP32
X
FP16
44.13 T
88.26 T
FP8
88.26 T
176.52 T
TF32
22.06 T
44.13 T
BF16
44.13 T
88.26 T
INT16
X
INT8
176.52 T
353.03 T
INT4
353.03 T
706.07 T

NVIDIA GeForce RTX 4060 Ti AI performance (Tensor Performance / Matrix Performance)

FP16: 44.13 TFLOPS, with sparsity: 88.26 TFLOPS

FP8: 88.26 TFLOPS, with sparsity: 176.52 TFLOPS

TF32: 22.06 TFLOPS, with sparsity: 44.13 TFLOPS

BF16: 44.13 TFLOPS, with sparsity: 88.26 TFLOPS

INT8: 176.52 TOPS, with sparsity: 353.03 TOPS

INT4: 353.03 TOPS, with sparsity: 706.07 TOPS

Hardware Specs
NVIDIA GeForce RTX 4060 Ti is a 5nm chip, has 22900 million transistors, launched by NVIDIA at 2023. It has 16 GB built-in(On-Board/On-Chip) memory with bandwidth up to 288 GB/s. It has 4352 general-purpose ALUs(CUDA cores/Shader cores) and 136 matrix cores(Tensor cores) .
Process Node
5 nm
Launch Year
2023
Vector(CUDA) Cores
4352
Matrix(Tensor) Cores
136
Core Frequency
2310 ~ 2535 MHz
Cache
32MB