AMD Instinct MI250 vs NVIDIA T10

AMD Instinct MI250
NVIDIA T10
13312 Shaders
128GB HBM2e
1700MHz
3072 Shaders
16GB GDDR6
1395MHz
Peak AI Performance
362.09 TOPS
INT4 Tensor
Peak AI Performance
274.27 TOPS
INT4 Tensor
FP32
45.26 TFLOPS
FP32
8.57 TFLOPS
FP16
90.52 TFLOPS
FP16
17.14 TFLOPS
Form Factor
OAM Module
-
Form Factor
PCIe Card
1.0-Slots
TDP
500W
TDP
150W
-
-
-
-
-
Power Connectors
-
1x 8-Pin
-
-
GB6 OpenCL N/A
0%
GB6 Metal N/A
0%
GB6 Metal N/A
0%
GB6 Vulkan N/A
0%
GB6 Vulkan N/A
0%
GB5 OpenCL N/A
0%
GB5 OpenCL N/A
0%
GB5 CUDA N/A
0%
GB5 CUDA N/A
0%
GB5 Metal N/A
0%
GB5 Metal N/A
0%
GB5 Vulkan N/A
0%
GB5 Vulkan N/A
0%
OCT 2020.1 N/A
0%
OCT 2020.1 N/A
0%
OCT Metal N/A
0%
OCT Metal N/A
0%
Peak AI Performance
362.09 TOPS
INT4 Tensor
Peak AI Performance
274.27 TOPS
INT4 Tensor
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
FP16
90.52 TFLOPS
362.09 TFLOPS Tensor (FP16 Accumulate)
-
362.09 TFLOPS Tensor (FP32 Accumulate)
-
FP16
17.14 TFLOPS
68.57 TFLOPS Tensor (FP16 Accumulate)
-
68.57 TFLOPS Tensor (FP32 Accumulate)
-
FP32
45.26 TFLOPS
90.52 TFLOPS Tensor
-
FP32
8.57 TFLOPS
-
-
FP64
45.26 TFLOPS
90.52 TFLOPS Tensor
FP64
270 GFLOPS
-
BF16
-
362.09 TFLOPS Tensor
-
-
-
-
-
-
-
-
-
-
-
INT4
362.09 TOPS Tensor
-
INT4
274.27 TOPS Tensor
-
INT8
-
362.09 TOPS Tensor
-
INT8
-
137.13 TOPS Tensor
-
-
-
-
-
-
-
Ray Tracing
33.5 TOPS
Pixel Fillrate
-
Pixel Fillrate
133.92 GPixel/s
-
-
-
-
Texture Fillrate
1414.4 GTexel/s
Texture Fillrate
267.84 GTexel/s
Manufacturer
AMD
Manufacturer
NVIDIA
Chip Designer
AMD
Chip Designer
NVIDIA
Architecture
CDNA 2
Architecture
Turing
Family
Instinct
Family
Server
Codename
Aldebaran
Aldebaran XT
Variant
Aldebaran XT
Codename
NV162
TU102
Variant
TU102-890-KCD-A1
Market Segment
Server
Market Segment
Server
Release Date
11/8/2021
Release Date
9/13/2018
Foundry
TSMC
-
Foundry
TSMC
-
Fabrication Node
N6
-
Fabrication Node
12FFN
-
Die Size
2x 724 mm²
-
Die Size
754 mm²
-
Transistor Count
2x 28 Billion
-
Transistor Count
18.6 Billion
-
Transistor Density
38.67M/mm²
-
Transistor Density
24.67M/mm²
-
Form
OAM Module
Form
PCIe Card
Shading Units
13312 Shaders
-
Shading Units
3072 Shaders
-
Texture Mapping Units
832 TMUs
Texture Mapping Units
192 TMUs
Render Output Units
-
Render Output Units
96 ROPs
Tensor Cores
832 T-Cores
Tensor Cores
384 T-Cores
-
-
Ray-Tracing Cores
48 RT-Cores
-
-
Streaming Multiprocessors
48 SMs
Compute Units
208 CUs
-
-
-
-
-
-
-
-
-
-
-
-
1000MHz Base
1700MHz
-
-
1065MHz Base
1395MHz
-
-
-
-
L1
-
-
16KB/CU
-
L1
32KB/SM Tex
64KB/SM
-
-
L2
16MB Shared
L2
6MB Shared
-
-
-
-
-
-
128GB
HBM2e
ECC
16GB
GDDR6
-
Bus Width
8192Bit
Bus Width
256Bit
Clock
1600MHz
Transfer Rate
3.2GT/s
Bandwidth
3276.8GB/s
Clock
1575MHz
Transfer Rate
12.6GT/s
Bandwidth
403.2GB/s
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
TDP
500W
TDP
150W
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
No Ports
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
No Ports
Max Resolution
Unknown
Max Resolution
Unknown
Max Resolution Refresh Rate
-
Max Resolution Refresh Rate
-
Variable Refresh Rate
-
-
-
Variable Refresh Rate
G-Sync
FreeSync
-
Display Stream Compression (DSC)
Not Supported
Display Stream Compression (DSC)
Not Supported
Multi Monitor Support
Unknown
Multi Monitor Support
Unknown
-
-
-
-
Model
2x VCN 2.6
Model
NVENC 7
Codec
-
-
-
-
-
-
-
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
Codec
-
-
-
-
-
-
-
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
No Decoders
Model
NVDEC 4
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Codec
MPEG-1
MPEG-2
MPEG-4
-
VC-1
VP8
VP9
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
-
-
-
-
Direct X
12
Direct 3D
12_2
-
-
OpenCL
3.0
-
-
OpenGL
4.6
OpenCL
3.0
Vulkan
1.2
-
-
-
-
GFX
9.4
-
-
-
-
Shader Model
6.6
CUDA
7.5
-
-
PureVideo HD
VP10
VDPAU
Feature Set J
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Power Connectors
-
-
-
1x 8-Pin
-
-
-
-
-
PCIe Version
4.0
PCIe Lanes
16
Slots Required
1.0
PCIe Version
3.0
PCIe Lanes
16
Multi GPU Support
Supported
Type
Infinity Fabric
Multi GPU Support
Supported
Type
NVLink
-
-
-
-
-
-
Height
111 mm (4.37 in)
Width
267 mm (10.51 in)
Depth
20 mm (0.79 in)
Change Comparison