NVIDIA CMP 40HX vs Moore Threads MTT S70
We compared two Desktop platform GPUs: 8GB VRAM CMP 40HX and 7GB VRAM MTT S70 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA CMP 40HX 's Advantages
Boost Clock1650MHz
More VRAM (8GB vs 7GB)
Larger VRAM bandwidth (448.0GB/s vs 392.0GB/s)
Lower TDP (185W vs 220W)
Moore Threads MTT S70 's Advantages
Released 2 years and 8 months late
1280 additional rendering cores
Score
Benchmark
FP32 (float)
CMP 40HX
7.603 TFLOPS
MTT S70
+50%
11.47 TFLOPS
Graphics Card
Feb 2021
Release Date
Oct 2023
Mining GPUs
Generation
MUSA-Chunxiao
Desktop
Type
Desktop
PCIe 1.0 x4
Bus Interface
PCIe 4.0 x16
Clock Speeds
1470 MHz
Base Clock
1600 MHz
1650 MHz
Boost Clock
0 MHz
1750 MHz
Memory Clock
1750 MHz
Memory
8GB
Memory Size
7GB
GDDR6
Memory Type
GDDR6
256bit
Memory Bus
224bit
448.0GB/s
Bandwidth
392.0GB/s
Render Config
-
Compute Units
28
36
SM Count
2304
Shading Units
3584
144
TMUs
224
64
ROPs
224
288
Tensor Cores
36
RT Cores
64 KB (per SM)
L1 Cache
4 MB
L2 Cache
0 MB
-
-
-
Theoretical Performance
105.6 GPixel/s
Pixel Rate
358.4 GPixel/s
237.6 GTexel/s
Texture Rate
358.4 GTexel/s
15.21 TFLOPS
FP16 (half)
22.94 TFLOPS
7.603 TFLOPS
FP32 (float)
11.47 TFLOPS
237.6 GFLOPS
FP64 (double)
179.2 GFLOPS
Board Design
185W
TDP
220W
450 W
Suggested PSU
550 W
No outputs
Outputs
1x HDMI 2.1
3x DisplayPort 1.4a
1x 8-pin
Power Connectors
1x 8-pin
Graphics Processor
TU106
GPU Name
MTT Gen 2
TU106-100-A1
GPU Variant
Turing
Architecture
MUSA-Chunxiao
TSMC
Foundry
TSMC
12 nm
Process Size
12 nm
10.8 billion
Transistors
22 billion
445 mm²
Die Size
Unknown
Graphics Features
12 Ultimate (12_2)
DirectX
12 Ultimate (12_2)
4.6
OpenGL
4.6
3.0
OpenCL
3.0
1.3
Vulkan
1.4
7.5
CUDA
6.8
Shader Model
6.5
