GPU

Moore Threads MTT S80

Edit@2 months ago

Intergrated Memory(VRAM)
Capacity

16 GB

(GDDR6 256-bit)

Bandwidth

448 GB/s

64 Token/s

Vector Compute
FP64
230.40 G
FP32
14.75 T
FP16
29.49 T
BF16
X
INT32
NI
INT8
X

Moore Threads MTT S80 General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 230.40 GFLOPS

FP32: 14.75 TFLOPS

FP16: 29.49 TFLOPS

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

Moore Threads MTT S80 AI performance (Tensor Performance / Matrix Performance)

Hardware Specs
Moore Threads MTT S80 is a 12nm chip, has 22000 million transistors, launched by Moore Threads at 2022. It has 16 GB built-in(On-Board/On-Chip) memory with bandwidth up to 448 GB/s. It has 4096 general-purpose ALUs(CUDA cores/Shader cores).
Process Node
12 nm
Launch Year
2022
Vector(CUDA) Cores
4096
Matrix(Tensor) Cores
Core Frequency
1800 ~ MHz
Cache
4MB