GPU Comparison
RADEON
AMD Instinct MI100
CORE STATE
Arcturus
VRAM
32 GB
CLOCK SPEED
1502 MHz
TDP
300 W
BUS WIDTH
4096 bit
ARCHITECTURE
CDNA 1.0
PROCESS
7 nm
LAUNCH DATE
2020
VS
GEFORCE
H200 NVL
CORE STATE
GH100
VRAM
141 GB
CLOCK SPEED
1785 MHz
TDP
600 W
BUS WIDTH
6144 bit
ARCHITECTURE
Hopper
PROCESS
5 nm
LAUNCH DATE
2024
PERFORMANCE BENCHMARKS
geekbench_opencl
DETAILED SPECIFICATIONS
SPECIFICATION
Instinct MI100
H200 NVL
Core Specs
Shading Units
7,680
16,896
+120.0%
Shaders
7,680
16,896
+120.0%
TMUs
480
528
+10.0%
ROPs
64
24
-62.5%
Compute Units
120
—
SM Count
—
132
Clocks
Base Clock
1000 MHz
1365 MHz
Boost Clock
1502 MHz
1785 MHz
Memory Clock
1200 MHz 2.4 Gbps effective
1593 MHz 6.4 Gbps effective
Memory
Memory Size
32 GB
141 GB
VRAM (MB)
32,768
144,384
+340.6%
Memory Type
HBM2
HBM3e
Memory Bus
4096 bit
6144 bit
Bandwidth
1.23 TB/s
4.89 TB/s
Cache
L1 Cache
16 KB (per CU)
256 KB (per SM)
L2 Cache
8 MB
50 MB
Performance
Pixel Rate
96.13 GPixel/s
42.84 GPixel/s
Texture Rate
721.0 GTexel/s
942.5 GTexel/s
FP32 (TFLOPS)
23.07 TFLOPS
60.32 TFLOPS
FP64 (TFLOPS)
11.54 TFLOPS (1:2)
30.16 TFLOPS (1:2)
FP16 (TFLOPS)
46.14 TFLOPS (2:1)
120.6 TFLOPS (2:1)
AI/RT
Tensor Cores
—
528
Power
TDP
300 W
600 W
TDP (W)
300
600
+100.0%
Suggested PSU
700 W
1000 W
Power Connectors
2x 8-pin
8-pin EPS
Architecture
Architecture
CDNA 1.0
Hopper
GPU Name
Arcturus
GH100
Generation
Instinct (MIx)
Server Hopper (Hxx)
Process Size
7 nm
5 nm
Transistors
25,600 million
80,000 million
Die Size
750 mm²
814 mm²
Foundry
TSMC
TSMC
Density
34.1M / mm²
98.3M / mm²
API Support
OpenCL
2.1
3.0
CUDA
—
9.0
Physical
Slot Width
Dual-slot
Dual-slot
Length
267 mm 10.5 inches
267 mm 10.5 inches
Height
111 mm 4.4 inches
111 mm 4.4 inches
Outputs
No outputs
No outputs
Bus Interface
PCIe 4.0 x16
PCIe 5.0 x16
Other
Production
End-of-life
Active
Predecessor
Radeon Instinct
Server Ada
Successor
—
Server Blackwell