AMD Instinct MI350X vs NVIDIA V100 FHHL

AMD Instinct MI350X
NVIDIA V100 FHHL
16384 Shaders
288GB HBM3e
2200MHz
5120 Shaders
16GB HBM2
1290MHz
Peak AI Performance
18.45 PFLOPS
FP4 Tensor Sparse
Peak AI Performance
105.68 TFLOPS
FP16 Tensor (FP16 Accumulate)
FP32
144.18 TFLOPS
FP32
13.21 TFLOPS
FP16
144.18 TFLOPS
FP16
26.42 TFLOPS
Form Factor
OAM Module
-
Form Factor
PCIe Card
1.0-Slots
TDP
1000W
TDP
250W
-
-
-
-
-
Power Connectors
-
1x 8-Pin
-
-
GB6 OpenCL N/A
0%
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
18.45 PFLOPS
FP4 Tensor Sparse
Peak AI Performance
105.68 TFLOPS
FP16 Tensor (FP16 Accumulate)
FP4
9.23 PFLOPS Tensor
18.45 PFLOPS Tensor Sparse
-
-
-
FP8
-
4.61 PFLOPS Tensor (FP16 Accumulate)
9.23 PFLOPS Tensor (FP16 Accumulate) Sparse
4.61 PFLOPS Tensor (FP32 Accumulate)
9.23 PFLOPS Tensor (FP32 Accumulate) Sparse
-
-
-
-
-
-
FP16
144.18 TFLOPS
2.31 PFLOPS Tensor (FP16 Accumulate)
4.61 PFLOPS Tensor (FP16 Accumulate) Sparse
2.31 PFLOPS Tensor (FP32 Accumulate)
4.61 PFLOPS Tensor (FP32 Accumulate) Sparse
FP16
26.42 TFLOPS
105.68 TFLOPS Tensor (FP16 Accumulate)
-
-
-
FP32
144.18 TFLOPS
144.18 TFLOPS Tensor
288.36 TFLOPS Tensor Sparse
FP32
13.21 TFLOPS
-
-
FP64
72.09 TFLOPS
72.09 TFLOPS Tensor
FP64
6.61 TFLOPS
-
BF16
-
2.31 PFLOPS Tensor
4.61 PFLOPS Tensor Sparse
-
-
-
-
-
-
-
-
-
-
INT4
4.61 POPS Tensor
9.23 POPS Tensor Sparse
-
-
-
INT8
-
4.61 POPS Tensor
9.23 POPS Tensor Sparse
INT8
52.84 TOPS
-
-
-
-
INT32
13.21 TOPS
-
-
-
-
Pixel Fillrate
-
Pixel Fillrate
165.12 GPixel/s
-
-
-
-
Texture Fillrate
-
Texture Fillrate
412.8 GTexel/s
Manufacturer
AMD
Manufacturer
NVIDIA
Chip Designer
AMD
Chip Designer
NVIDIA
Architecture
CDNA 4
Architecture
Volta
Family
Instinct
Family
Server
Codename
CDNA Next
-
-
-
Codename
NV140
GV100
-
-
Market Segment
Server
Market Segment
Server
Release Date
6/12/2025
Release Date
3/27/2018
Foundry
TSMC
TSMC IO Die
Foundry
TSMC
-
Fabrication Node
N3P
N6 IO Die
Fabrication Node
12FFN
-
Die Size
8x 181 mm²
-
Die Size
815 mm²
-
Transistor Count
8x 17.4 Billion
2x 23 Billion IO Die
Transistor Count
21.1 Billion
-
Transistor Density
96.13M/mm²
-
Transistor Density
25.89M/mm²
-
Form
OAM Module
Form
PCIe Card
Shading Units
16384 Shaders
-
Shading Units
5120 Shaders
-
Texture Mapping Units
-
Texture Mapping Units
320 TMUs
Render Output Units
-
Render Output Units
128 ROPs
Tensor Cores
1024 T-Cores
Tensor Cores
640 T-Cores
-
-
-
-
-
-
Streaming Multiprocessors
80 SMs
Compute Units
256 CUs
-
-
-
-
-
-
-
-
Graphics Processing Clusters
6 GPCs
-
-
-
2200MHz
-
-
937MHz Base
1290MHz
-
-
-
-
L1
-
-
16KB/CU
-
L1
-
128KB/SM
-
-
L2
16MB Shared
L2
6MB Shared
L3
256MB Shared
-
-
-
-
288GB
HBM3e
ECC
16GB
HBM2
-
Bus Width
8192Bit
Bus Width
4096Bit
Clock
3906MHz
Transfer Rate
7.8GT/s
Bandwidth
8000.1GB/s
Clock
810MHz
Transfer Rate
1.6GT/s
Bandwidth
829.4GB/s
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
TDP
1000W
TDP
250W
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
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
4x VCN 4.0
Model
3x NVENC 5
Codec
-
-
-
-
-
-
VP9
-
AVC (H.264)
HEVC (H.265)
-
AV1
-
-
Codec
-
-
-
-
-
-
-
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
Model
4x VCN 4.0
Model
NVDEC 3
Codec
MPEG-1
MPEG-2
MPEG-4
JPEG
VC-1
-
VP9
-
AVC (H.264)
HEVC (H.265)
-
AV1
-
-
Codec
MPEG-1
MPEG-2
MPEG-4
-
VC-1
-
VP9
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
-
-
-
-
Direct X
12
Direct 3D
12_1
-
-
OpenCL
3.0
-
-
OpenGL
4.6
OpenCL
3.0
-
-
-
-
-
-
GFX
9.4
-
-
-
-
Shader Model
6.7
CUDA
7.0
-
-
PureVideo HD
VP9
VDPAU
Feature Set I
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Power Connectors
-
-
-
1x 8-Pin
-
-
-
-
-
PCIe Version
5.0
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
168 mm (6.61 in)
Depth
20 mm (0.79 in)
Change Comparison