GPU Comparison
GEFORCE
NVIDIA CMP 40HX
CORE STATE
TU106
VRAM
8 GB
CLOCK SPEED
1650 MHz
TDP
185 W
BUS WIDTH
256 bit
ARCHITECTURE
Turing
PROCESS
12 nm
LAUNCH DATE
2021
VS
GEFORCE
L4
CORE STATE
AD104
VRAM
24 GB
CLOCK SPEED
2040 MHz
TDP
72 W
BUS WIDTH
192 bit
ARCHITECTURE
Ada Lovelace
PROCESS
5 nm
LAUNCH DATE
2023
PERFORMANCE BENCHMARKS
geekbench_opencl
geekbench_vulkan
DETAILED SPECIFICATIONS
SPECIFICATION
CMP 40HX
L4
Core Specs
Shading Units
2,304
7,424
+222.2%
Shaders
2,304
7,424
+222.2%
TMUs
144
240
+66.7%
ROPs
64
80
+25.0%
SM Count
36
60
+66.7%
Clocks
Base Clock
1470 MHz
795 MHz
Boost Clock
1650 MHz
2040 MHz
Memory Clock
1750 MHz
14 Gbps effective
1563 MHz
12.5 Gbps effective
Memory
Memory Size
8 GB
24 GB
VRAM (MB)
8,192
24,576
+200.0%
Memory Type
GDDR6
GDDR6
Memory Bus
256 bit
192 bit
Bandwidth
448.0 GB/s
300.1 GB/s
Cache
L1 Cache
64 KB (per SM)
128 KB (per SM)
L2 Cache
4 MB
48 MB
Performance
Pixel Rate
105.6 GPixel/s
163.2 GPixel/s
Texture Rate
237.6 GTexel/s
489.6 GTexel/s
FP32 (TFLOPS)
7.603 TFLOPS
30.29 TFLOPS
FP64 (TFLOPS)
237.6 GFLOPS (1:32)
473.3 GFLOPS (1:64)
FP16 (TFLOPS)
15.21 TFLOPS (2:1)
30.29 TFLOPS (1:1)
AI/RT
RT Cores
36
60
+66.7%
Tensor Cores
288
240
-16.7%
Power
TDP
185 W
72 W
TDP (W)
185
72
-61.1%
Suggested PSU
450 W
250 W
Power Connectors
1x 8-pin
None
Architecture
Architecture
Turing
Ada Lovelace
GPU Name
TU106
AD104
Generation
Mining GPUs
Server Ada
(Lxx)
Process Size
12 nm
5 nm
Transistors
10,800 million
35,800 million
Die Size
445 mm²
294 mm²
Foundry
TSMC
TSMC
Density
24.3M / mm²
121.8M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
7.5
8.9
Shader Model
6.8
6.8
Physical
Slot Width
Dual-slot
Single-slot
Length
229 mm
9 inches
169 mm
6.7 inches
Height
111 mm
4.4 inches
56 mm
2.2 inches
Outputs
No outputs
No outputs
Bus Interface
PCIe 1.0 x4
PCIe 4.0 x16
Other
Launch Price
699 USD
—
Production
End-of-life
Active
Predecessor
—
Server Ampere
Successor
—
Server Hopper