AIU

MLU 270

Edit@3 months ago

Intergrated Memory
No available data
Vector Compute
FP64
X
FP32
X
FP16
X
BF16
X
INT32
X
INT8
X

MLU 270 General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

Matirx Compute
FP64
X
FP32
X
FP16
X
FP8
X
TF32
X
BF16
X
INT16
64 T
INT8
128 T
INT4
256 T

MLU 270 AI performance (Tensor Performance / Matrix Performance)

INT16: 64 TOPS

INT8: 128 TOPS

INT4: 256 TOPS

Hardware Specs
MLU 270 launched by Cambricon at 2021.
Process Node
nm
Launch Year
2021