GPU

NVIDIA Tesla M4

Edit@2 months ago

Intergrated Memory(VRAM)
Capacity

4 GB

(GDDR5 128-bit)

Bandwidth

88 GB/s

Out of Memory Token/s

Vector Compute
FP64
68.61 G
FP32
2.19 T
FP16
X
BF16
X
INT32
NI
INT8
X

NVIDIA Tesla M4 General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 68.61 GFLOPS

FP32: 2.19 TFLOPS

Matirx Compute
FP64
X
FP32
X
FP16
X
FP8
X
TF32
X
BF16
X
INT16
X
INT8
X
INT4
X

NVIDIA Tesla M4 AI performance (Tensor Performance / Matrix Performance)

Hardware Specs
NVIDIA Tesla M4 is a 28nm chip, has 2940 million transistors, launched by NVIDIA at 2015. It has 4 GB built-in(On-Board/On-Chip) memory with bandwidth up to 88 GB/s. It has 1024 general-purpose ALUs(CUDA cores/Shader cores).
Process Node
28 nm
Launch Year
2015
Vector(CUDA) Cores
1024
Matrix(Tensor) Cores
Core Frequency
872 ~ 1072 MHz
Cache
1MB