GPU

AMD Radeon Instinct MI355X

Edit@17 days ago

Intergrated Memory(VRAM)
Capacity

288 GB

(HBM3e 8192-bit)

Bandwidth

8192 GB/s

1170 Token/s

Vector Compute
FP64
78.64 T
FP32
157.28 T
FP16
314.55 T
BF16
X
INT32
X
INT8
X

AMD Radeon Instinct MI355X General-Purpose Float-Point performance (Vector Performance / Scalar Performance)

FP64: 78.64 TFLOPS

FP32: 157.28 TFLOPS

FP16: 314.55 TFLOPS

Matirx Compute
FP64
78.64 T
157.28 T
FP32
157.28 T
314.56 T
FP16
2516.60 T
5033.20 T
FP8
5033.20 T
10.07 P
TF32
X
BF16
2516.60 T
5033.20 T
INT16
X
INT8
5033.20 T
10.07 P
INT4
5033.20 T
10.07 P

AMD Radeon Instinct MI355X AI performance (Tensor Performance / Matrix Performance)

FP64: 78.64 TFLOPS, with sparsity: 157.28 TFLOPS

FP32: 157.28 TFLOPS, with sparsity: 314.56 TFLOPS

FP16: 2516.60 TFLOPS, with sparsity: 5033.20 TFLOPS

FP8: 5033.20 TFLOPS, with sparsity: 10.07 PFLOPS

BF16: 2516.60 TFLOPS, with sparsity: 5033.20 TFLOPS

INT8: 5033.20 TOPS, with sparsity: 10.07 POPS

INT4: 5033.20 TOPS, with sparsity: 10.07 POPS

Hardware Specs
AMD Radeon Instinct MI355X is a 3nm chip, has 185000 million transistors, launched by AMD at 2025. It has 288 GB built-in(On-Board/On-Chip) memory with bandwidth up to 8192 GB/s. It has 16384 general-purpose ALUs(CUDA cores/Shader cores) and 1024 matrix cores(Tensor cores) .
Process Node
3 nm
Launch Year
2025
Vector(CUDA) Cores
16384
Matrix(Tensor) Cores
1024
Core Frequency
1000 ~ 2400 MHz
Cache
256MB