NVIDIA CMP 90HX
Graphics processor
- Video memory
- 10 GB*
- Memory bus
- 320 bit*
- Rated memory bandwidth
- 760.30 GB/s*
Graphics & compute benchmarks
GFXBench, GLBenchmark and CompuBench results retain the tested system, API, operating system and resolution. Offscreen tests fix the render resolution; onscreen results also reflect the display workload. These historical submissions do not establish performance in current games.
Catmull-Clark Subdivision Level 3Windows / cu / compute · mTriangles/s
Higher is better.
NVIDIA CMP 90HX
334.30* mTriangles/s
NVIDIA CMP 90HX
Host configuration not reported.
Catmull-Clark Subdivision Level 5Windows / cu / compute · mTriangles/s
Higher is better.
NVIDIA CMP 90HX
672.05* mTriangles/s
NVIDIA CMP 90HX
Host configuration not reported.
Level Set Segmentation – 128Windows / cu / compute · mVoxels/s
Higher is better.
NVIDIA CMP 90HX
3,527.92* mVoxels/s
NVIDIA CMP 90HX
Host configuration not reported.
Level Set Segmentation – 256Windows / cu / compute · mVoxels/s
Higher is better.
NVIDIA CMP 90HX
3,586.58* mVoxels/s
NVIDIA CMP 90HX
Host configuration not reported.
N-Body simulation – 1024kWindows / cu / compute · Iterations/s
Higher is better.
NVIDIA CMP 90HX
49.98* Iterations/s
NVIDIA CMP 90HX
Host configuration not reported.
N-Body simulation – 128kWindows / cu / compute · Iterations/s
Higher is better.
NVIDIA CMP 90HX
324.22* Iterations/s
NVIDIA CMP 90HX
Host configuration not reported.
Ocean Surface SimulationWindows / cu / compute · Iterations/s
Higher is better.
NVIDIA CMP 90HX
3,219.22* Iterations/s
NVIDIA CMP 90HX
Host configuration not reported.
Vertex Connection and MergingWindows / cu / compute · mPixels/s
Higher is better.
NVIDIA CMP 90HX
8.58* mPixels/s
NVIDIA CMP 90HX
Host configuration not reported.
Median results, not the best run. Sample counts and individual test dates are unreported. Different systems using the same GPU remain separate configurations.
Processing & platform
- Architecture
- Ampere*
- Process node
- 8 nm*
- GPU cores
- 6,400*
- Reported GPU frequency
- 1,440 MHz*
- Graphics interface
- PCIe 1.0 x4*
- Reported chip die area
- 628 mm²*
- Chip/package TDP
- 320 W*
- Transistors
- 28,300 million*
- GPU L2 cache
- 5 MB*
- Graphics processor
- GA102*
- Texture mapping units
- 200*
- Raster output units
- 80*
- Streaming multiprocessors
- 50*
- Tensor cores
- 200*
- Ray tracing cores
- 50*
- Chip foundry
- Samsung*
- Card or module format
- Dual-slot*
- Declared GPU TDP
- 320 W*
- Suggested system PSU
- 700 W*
- External power connectors
- 2x 8-pin*
- DirectX / feature level
- 12 Ultimate (12_2)*
- OpenGL
- 4.6*
- OpenCL
- 3.0*
- Vulkan
- 1.4*
- CUDA compute capability
- 8.6*
- Shader model
- 6.9*
Memory & storage
- Memory options
- 10*
- Video memory
- 10 GB*
- Video memory type
- GDDR6X*
- Memory bus
- 320 bit*
- Rated memory bandwidth
- 760.30 GB/s*
Power specifications
- Thermal design power
- 320 W*
System configuration
Detailed specifications
Specifications 1 · 13 fields
- Foundry
- Samsung*
- Vendor
- NVIDIA*
- FP16 GFLOPS
- 29,770* GFLOPS
- FP32 GFLOPS
- 29,770* GFLOPS
- FP64 GFLOPS
- 465.1* GFLOPS
- Product
- NVIDIA CMP 90HX*
- Type
- GPU*
- Release Date
- 2021-01-01*
- Process Size (nm)
- 8* nm
- TDP (W)
- 320* W
- Die Size (mm^2)
- 628* mm²
- Transistors (million)
- 28,300*
- Freq (MHz)
- 1,440* MHz
Specifications 2 · 31 fields
- proc transistors
- 28,300 million*
- proc die size
- 628* mm²
- cores
- 6,400*
- tensor cores
- 200*
- rt cores
- 50*
- sm count
- 50*
- base clock
- 1,500*
- boost clock
- 1,710*
- memory clock
- 1,188*
- ram
- 10*
- name
- NVIDIA CMP 90HX*
- ram type
- GDDR6X*
- bus width
- 320*
- bus interface
- PCIe 1.0 x4*
- ram bandwidth
- 760.3* GB/s
- fp16 performance
- 21.89*
- fp32 performance
- 21.89*
- fp64 performance
- 0.342*
- tdp
- 320* W
- suggested psu
- 700* W
- slot width
- Dual-slot*
- power connectors
- 2x 8-pin*
- api cuda
- 8.6*
- is retail board
- No*
- gpi value
- 12.57*
- vendor
- nvd*
- manufacturer
- NVIDIA*
- arch name
- Ampere*
- card release date
- 2021-07-28*
- proc foundry
- Samsung*
- proc process size
- 8* nm
Specifications 3 · 40 fields
- CUDA compute capability
- 8.6*
- DirectX / feature level
- 12 Ultimate (12_2)*
- OpenCL
- 3.0*
- OpenGL
- 4.6*
- Shader model
- 6.9*
- Vulkan
- 1.4*
- Reported theoretical BF16 peak
- 43.78* TFLOPS
Scope and conditions
- Arithmetic density
- Not stated*
- Graphics interface
- PCIe 1.0 x4*
- Memory bus
- 320* bit
- Reported shader processors
- 6,400*
- Reported theoretical FP16 peak
- 21.89* TFLOPS
Scope and conditions
- Arithmetic density
- Not stated*
- Reported theoretical FP32 peak
- 21.89* TFLOPS
Scope and conditions
- Arithmetic density
- Not stated*
- Reported theoretical FP64 peak
- 0.342* TFLOPS
Scope and conditions
- Arithmetic density
- Not stated*
- Declared L1 cache configuration
- 128 KB (per SM)*
- Declared GPU L2 cache
- 5 MB*
- Reported display outputs
- No outputs*
- Theoretical pixel fill rate
- 136.8 GPixel/s*
- External power connectors
- 2x 8-pin*
- Reported chip die area
- 628 mm²*
- Chip foundry
- Samsung*
- Process node declaration
- 8 nm*
- Transistors
- 28.3 billion*
- Graphics processor
- GA102*
- Reported video memory
- 10* GB
- Rated memory bandwidth
- 760.3 GB/s*
- Video memory type
- GDDR6X*
- Raster output units
- 80*
- Ray tracing cores
- 50*
- Reported shading units
- 6,400*
- Card or module format
- Dual-slot*
- Streaming multiprocessors
- 50*
- Suggested system PSU
- 700 W*
- Declared GPU TDP
- 320 W*
- Tensor cores
- 200*
- Theoretical texture fill rate
- 342.0 GTexel/s*
- Reported theoretical TF32 peak
- 21.89* TFLOPS
Scope and conditions
- Arithmetic density
- Not stated*
- Texture mapping units
- 200*
- GPU cores
- 6,400*
- Video memory
- 10* GB
- Video memory type
- GDDR6X*
* Provisional data