CMP 70HX

Key Metrics

Save this GPU, then open another GPU page to compare them side by side.

Identity
Entry Type	Reference GPU
Official Name	NVIDIA CMP 70HX
GPU Codename	GA104
Architecture	Ampere
Launch Date	2021-01-01
Market Segment Notes	Desktop in Mining GPUs
Core Specs
GPU Die	GA104
Chip Variant / Package	BGA-2713
Process Node	Samsung 8 nm
Transistor Count	17.4 billion
Die Size	392 mm²
CUDA Cores	3840
SM Count	30
TMUs	120
ROPs	64
AI Units / Tensor Cores	120
RT Units / RT Cores	30
Base Clock	1,365 MHz
Boost Clock	1,395 MHz
Pixel Fillrate	89.28 GP/s
Texture Fillrate	167.4 GT/s
FP32 Throughput	10.71 TFLOPS
FP16 Throughput	10.71 TFLOPS
FP64 Throughput	0.17 TFLOPS
Memory
Memory Size	8 GB
Memory Type	GDDR6X
Memory Bus Width	256-bit
Memory Clock	1,188 MHz
Memory Bandwidth	608.3 GB/s
L2 Cache	4 MB
L1 Cache	128 KB
Board and Display
Recommended PSU	200 W
Power Connectors	1x 12-pin
Interface	PCIe 1.0 x4
Slot Width	Dual-slot, 267 mm length
Card Length	267 mm
Board Partner Model	Reference or OEM specification
Reference Design	NVIDIA reference design
Software
CUDA Compute Capability	8.6
Shader Model	6.8
DirectX Support	DirectX 12.2
Vulkan Support	Vulkan 1.4
OpenGL Support	OpenGL 4.6
OpenCL Support	OpenCL 3.0

Sources

sysrqmts CMP 70HXExact-model clue checked on 2026-06-25. Confirms GA104-100-A1 identity, 8 GB GDDR6X, and 1x 12-pin board-power layout; not treated as a direct TDP source.Open source

TechPowerUp GPU SpecsModel-level specification reference for this GPU entryOpen source

RightNow GPU DatabaseOpen GPU database used to cross-check specifications A reference database: good for cross-checking specifications, but not proof of the original release details on its own.Open source

PC Tech Vault source reviewTDP left blank after checking existing online source leads; available sources were sufficient for identity or board clues but did not provide a direct model-level TDP value.