GPU Comparison
GEFORCE
NVIDIA GRID K2
CORE STATE
GK104
VRAM
4 GB
CLOCK SPEED
—
TDP
225 W
BUS WIDTH
256 bit
ARCHITECTURE
Kepler
PROCESS
28 nm
LAUNCH DATE
2013
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_metal
geekbench_opencl
DETAILED SPECIFICATIONS
SPECIFICATION
GRID K2
H200 NVL
Core Specs
Shading Units
1,536
16,896
+1000.0%
Shaders
1,536
16,896
+1000.0%
TMUs
128
528
+312.5%
ROPs
32
24
-25.0%
SM Count
—
132
Clocks
Base Clock
—
1365 MHz
Boost Clock
—
1785 MHz
GPU Clock
745 MHz
—
Memory Clock
1250 MHz
5 Gbps effective
1593 MHz
6.4 Gbps effective
Memory
Memory Size
4 GB
141 GB
VRAM (MB)
4,096
144,384
+3425.0%
Memory Type
GDDR5
HBM3e
Memory Bus
256 bit
6144 bit
Bandwidth
160.0 GB/s
4.89 TB/s
Cache
L1 Cache
16 KB (per SMX)
256 KB (per SM)
L2 Cache
512 KB
50 MB
Performance
Pixel Rate
23.84 GPixel/s
42.84 GPixel/s
Texture Rate
95.36 GTexel/s
942.5 GTexel/s
FP32 (TFLOPS)
2.289 TFLOPS
60.32 TFLOPS
FP64 (TFLOPS)
95.36 GFLOPS (1:24)
30.16 TFLOPS (1:2)
FP16 (TFLOPS)
—
241.3 TFLOPS (4:1)
AI/RT
Tensor Cores
—
528
Power
TDP
225 W
600 W
TDP (W)
225
600
+166.7%
Suggested PSU
550 W
1000 W
Power Connectors
1x 6-pin + 1x 8-pin
8-pin EPS
Architecture
Architecture
Kepler
Hopper
GPU Name
GK104
GH100
Generation
GRID
(K2)
Server Hopper
(Hxx)
Process Size
28 nm
5 nm
Transistors
3,540 million
80,000 million
Die Size
294 mm²
814 mm²
Foundry
TSMC
TSMC
Density
12.0M / mm²
98.3M / mm²
API Support
DirectX
12 (11_0)
—
OpenGL
4.6
—
Vulkan
1.2.175
—
OpenCL
3.0
3.0
CUDA
3.0
9.0
Shader Model
6.5 (5.1)
—
Physical
Slot Width
Dual-slot
Dual-slot
Length
267 mm
10.5 inches
267 mm
10.5 inches
Height
—
111 mm
4.4 inches
Outputs
No outputs
No outputs
Bus Interface
PCIe 3.0 x16
PCIe 5.0 x16
Other
Launch Price
5,199 USD
—
Production
End-of-life
Active
Predecessor
—
Server Ada
Successor
—
Server Blackwell