Intel Gaudi 3 vs NVIDIA A100 SXM 40GB

Intel Gaudi 3
NVIDIA A100 SXM 40GB
16384 Shaders
128GB HBM2e
438MHz
6912 Shaders
40GB HBM2e
1410MHz
Peak AI Performance
1.84 PFLOPS
FP8 Tensor (FP32 Accumulate)
Peak AI Performance
2.49 POPS
INT4 Tensor Sparse
FP32
14.35 TFLOPS
FP32
19.49 TFLOPS
FP16
28.7 TFLOPS
FP16
77.97 TFLOPS
Form Factor
OAM Module
-
Form Factor
SXM4
-
TDP
900W
TDP
400W
-
-
-
-
-
-
-
-
-
-
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
1.84 PFLOPS
FP8 Tensor (FP32 Accumulate)
Peak AI Performance
2.49 POPS
INT4 Tensor Sparse
-
-
-
-
-
-
FP8
-
-
-
1.84 PFLOPS Tensor (FP32 Accumulate)
-
-
-
-
-
-
-
FP16
28.7 TFLOPS
-
-
458.75 TFLOPS Tensor (FP32 Accumulate)
-
FP16
77.97 TFLOPS
311.87 TFLOPS Tensor (FP16 Accumulate)
623.74 TFLOPS Tensor (FP16 Accumulate) Sparse
311.87 TFLOPS Tensor (FP32 Accumulate)
623.74 TFLOPS Tensor (FP32 Accumulate) Sparse
FP32
14.35 TFLOPS
229.38 TFLOPS Tensor
-
FP32
19.49 TFLOPS
-
-
-
-
-
FP64
9.75 TFLOPS
19.49 TFLOPS Tensor
BF16
28.7 TFLOPS
1.84 PFLOPS Tensor
-
BF16
38.98 TFLOPS
311.87 TFLOPS Tensor
623.74 TFLOPS Tensor Sparse
TF32
458.75 TFLOPS Tensor
-
TF32
155.94 TFLOPS Tensor
311.87 TFLOPS Tensor Sparse
-
-
-
INT4
1.25 POPS Tensor
2.49 POPS Tensor Sparse
-
-
-
-
INT8
-
623.74 TOPS Tensor
1.25 POPS Tensor Sparse
-
-
INT32
19.49 TOPS
-
-
-
-
Pixel Fillrate
-
Pixel Fillrate
225.6 GPixel/s
-
-
-
-
Texture Fillrate
-
Texture Fillrate
609.12 GTexel/s
Manufacturer
Intel
Manufacturer
NVIDIA
Chip Designer
Intel
Chip Designer
NVIDIA
Architecture
Gaudi 3
Architecture
Ampere
Family
Gaudi
Family
Server
Codename
Gaudi 3
Gaudi 3
-
-
Codename
NV170
GA100
-
-
Market Segment
Server
Market Segment
Server
Release Date
4/9/2024
Release Date
6/28/2021
Foundry
TSMC
-
Foundry
TSMC
-
Fabrication Node
N5
-
Fabrication Node
7N
-
-
-
-
Die Size
826 mm²
-
-
-
Transistor Count
54.2 Billion
-
-
-
-
Transistor Density
65.62M/mm²
-
Form
OAM Module
Form
SXM4
Shading Units
16384 Shaders
-
Shading Units
6912 Shaders
-
Texture Mapping Units
-
Texture Mapping Units
432 TMUs
Render Output Units
-
Render Output Units
160 ROPs
Tensor Cores
64 T-Cores
Tensor Cores
432 T-Cores
-
-
-
-
-
-
Streaming Multiprocessors
108 SMs
Compute Units
8 CUs
-
-
-
-
-
-
-
-
-
-
-
-
-
438MHz
-
-
1095MHz Base
1410MHz
-
-
-
-
L1
-
-
-
Unknown
L1
64KB/SM Tex
192KB/SM
-
-
L2
Unknown
L2
40MB Shared
-
-
-
-
-
-
128GB
HBM2e
ECC
40GB
HBM2e
-
Bus Width
8192Bit
Bus Width
5120Bit
Clock
1808MHz
Transfer Rate
3.6GT/s
Bandwidth
3702.8GB/s
Clock
1215MHz
Transfer Rate
2.4GT/s
Bandwidth
1555.2GB/s
-
-
-
-
eSRAM
96MB
12800GB/s
-
-
-
-
-
-
-
-
-
-
-
TDP
900W
TDP
400W
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
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
-
-
-
-
No Encoders
-
No Encoders
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
No Decoders
Model
5x NVDEC 4
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Codec
MPEG-1
MPEG-2
MPEG-4
-
VC-1
VP8
VP9
-
AVC (H.264)
HEVC (H.265)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
OpenCL
3.0
-
-
-
-
OpenCL
3.0
Vulkan
1.2
-
-
-
-
-
-
-
-
-
-
-
-
CUDA
8.0
-
-
PureVideo HD
VP10
VDPAU
Feature Set J
Not a Card
-
-
-
Not a Card
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
PCIe Version
5.0
PCIe Lanes
16
-
-
PCIe Version
4.0
PCIe Lanes
16
Multi GPU Support
Supported
Type
RoCE
Multi GPU Support
Supported
Type
NVLink
-
-
-
-
-
-
-
-
-
-
-
-
NVIDIA A100 SXM 40GB NVIDIA A100 SXM
NVIDIA A100 SXM 40GB vs NVIDIA A100 SXM
Change Comparison