AMD Instinct MI300A Specs
AMD Instinct MI300A
|
14592 Shaders
128GB HBM3
2100MHz
|
|
Peak AI Performance
3.92 POPS
INT8 Tensor Sparse
|
|
FP32
122.57 TFLOPS
|
|
FP16
245.15 TFLOPS
|
|
Form Factor
APU SH5 Socket
|
|
TDP
560W
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
0%
|
|
Foundry
TSMC
TSMC Active Interposer Die
|
|
Fabrication Node
N5
N6 Active Interposer Die
|
|
Die Size
3x 370 mm²
|
|
Transistor Count
3x 36.5 Billion
|
|
Transistor Density
98.65M/mm²
|
|
Form
APU SH5 Socket
|
|
Shading Units
14592 Shaders
|
|
Texture Mapping Units
912 TMUs
|
|
Render Output Units
|
|
Tensor Cores
912 T-Cores
|
|
Compute Units
228 CUs
|
|
1000MHz Base
2100MHz
|
|
Peak AI Performance
3.92 POPS
INT8 Tensor Sparse
|
|
FP8
1.96 PFLOPS Tensor (FP16 Accumulate)
3.92 PFLOPS Tensor (FP16 Accumulate) Sparse
1.96 PFLOPS Tensor (FP32 Accumulate)
3.92 PFLOPS Tensor (FP32 Accumulate) Sparse
|
|
FP16
245.15 TFLOPS
980.58 TFLOPS Tensor (FP16 Accumulate)
1.96 PFLOPS Tensor (FP16 Accumulate) Sparse
980.58 TFLOPS Tensor (FP32 Accumulate)
1.96 PFLOPS Tensor (FP32 Accumulate) Sparse
|
|
FP32
122.57 TFLOPS
122.57 TFLOPS Tensor
245.15 TFLOPS Tensor Sparse
|
|
FP64
61.29 TFLOPS
122.57 TFLOPS Tensor
|
|
BF16
980.58 TFLOPS Tensor
1.96 PFLOPS Tensor Sparse
|
|
TF32
490.29 TFLOPS Tensor
980.58 TFLOPS Tensor Sparse
|
|
INT8
1.96 POPS Tensor
3.92 POPS Tensor Sparse
|
|
Pixel Fillrate
|
|
Texture Fillrate
1915.2 GTexel/s
|
|
L1
16KB/CU
|
|
L2
16MB Shared
|
|
L3
192MB Shared
|
|
128GB
HBM3
ECC
|
|
Bus Width
8192Bit
|
|
Clock
2600MHz
Transfer Rate
5.2GT/s
Bandwidth
5324.8GB/s
|
|
TDP
560W
|
|
Max Resolution
Unknown
|
|
Max Resolution Refresh Rate
|
|
Variable Refresh Rate
|
|
Display Stream Compression (DSC)
Not Supported
|
|
Multi Monitor Support
Unknown
|
|
Model
2x VCN 2.6
|
|
Codec
AVC (H.264)
HEVC (H.265)
|
|
No Decoders
|
|
OpenCL
3.0
|
|
GFX
9.4
|
|
PCIe Version
5.0
PCIe Lanes
16
|
|
Multi GPU Support
Supported
Type
Infinity Fabric
|
Copy Link