GPU Comparison
GEFORCE
NVIDIA 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
VS
GEFORCE
RTX A5000
CORE STATE
GA102
VRAM
24 GB
CLOCK SPEED
1695 MHz
TDP
230 W
BUS WIDTH
384 bit
ARCHITECTURE
Ampere
PROCESS
8 nm
LAUNCH DATE
2021
PERFORMANCE BENCHMARKS
geekbench_opencl
3dmark_3dmark_steel_nomad_dx12
geekbench_vulkan
passmark_directx_10
passmark_directx_11
passmark_directx_12
passmark_directx_9
passmark_g2d
passmark_g3d
passmark_gpu_compute
DETAILED SPECIFICATIONS
SPECIFICATION
H200 NVL
RTX A5000
Core Specs
Shading Units
16,896
8,192
-51.5%
Shaders
16,896
8,192
-51.5%
TMUs
528
256
-51.5%
ROPs
24
96
+300.0%
SM Count
132
64
-51.5%
Clocks
Base Clock
1365 MHz
1170 MHz
Boost Clock
1785 MHz
1695 MHz
Memory Clock
1593 MHz
6.4 Gbps effective
2000 MHz
16 Gbps effective
Memory
Memory Size
141 GB
24 GB
VRAM (MB)
144,384
24,576
-83.0%
Memory Type
HBM3e
GDDR6
Memory Bus
6144 bit
384 bit
Bandwidth
4.89 TB/s
768.0 GB/s
Cache
L1 Cache
256 KB (per SM)
128 KB (per SM)
L2 Cache
50 MB
6 MB
Performance
Pixel Rate
42.84 GPixel/s
162.7 GPixel/s
Texture Rate
942.5 GTexel/s
433.9 GTexel/s
FP32 (TFLOPS)
60.32 TFLOPS
27.77 TFLOPS
FP64 (TFLOPS)
30.16 TFLOPS (1:2)
433.9 GFLOPS (1:64)
FP16 (TFLOPS)
241.3 TFLOPS (4:1)
27.77 TFLOPS (1:1)
AI/RT
RT Cores
—
64
Tensor Cores
528
256
-51.5%
Power
TDP
600 W
230 W
TDP (W)
600
230
-61.7%
Suggested PSU
1000 W
550 W
Power Connectors
8-pin EPS
1x 8-pin
Architecture
Architecture
Hopper
Ampere
GPU Name
GH100
GA102
Generation
Server Hopper
(Hxx)
Workstation Ampere
(Ax000)
Process Size
5 nm
8 nm
Transistors
80,000 million
28,300 million
Die Size
814 mm²
628 mm²
Foundry
TSMC
Samsung
Density
98.3M / mm²
45.1M / mm²
API Support
DirectX
—
12 Ultimate (12_2)
OpenGL
—
4.6
Vulkan
—
1.4
OpenCL
3.0
3.0
CUDA
9.0
8.6
Shader Model
—
6.8
Physical
Slot Width
Dual-slot
Dual-slot
Length
267 mm
10.5 inches
267 mm
10.5 inches
Height
111 mm
4.4 inches
112 mm
4.4 inches
Outputs
No outputs
4x DisplayPort 1.4a
Bus Interface
PCIe 5.0 x16
PCIe 4.0 x16
Other
Production
Active
End-of-life
Predecessor
Server Ada
Quadro Turing
Successor
Server Blackwell
Workstation Ada