GeekRadar / The comparison desk

COMPARE.

GPU

NVIDIA

NVIDIA PG506-232

Graphics processor

Video memory: 24 GB^*
Memory bus: 3,072 bit^*
Rated memory bandwidth: 933.10 GB/s^*

No matched benchmark results are available for this entry yet.

Processing & platform

Architecture: Ampere^*
Process node: 7 nm^*
GPU cores: 3,584^*
Reported GPU frequency: 930 MHz^*
Graphics interface: PCIe 4.0 x16^*
Reported chip die area: 826 mm²^*
Chip/package TDP: 165 W^*
Transistors: 54,200 million^*
GPU L2 cache: 24 MB^*
Graphics processor: GA100^*
Texture mapping units: 224^*
Raster output units: 96^*
Streaming multiprocessors: 56^*
Tensor cores: 224^*
Chip foundry: TSMC^*
Card or module format: Dual-slot^*
Declared GPU TDP: 165 W^*
Suggested system PSU: 450 W^*
External power connectors: 8-pin EPS^*
OpenCL: 3.0^*
CUDA compute capability: 8.0^*

Memory & storage

Memory options: 24^*
Video memory: 24 GB^*
Video memory type: HBM2^*
Memory bus: 3,072 bit^*
Rated memory bandwidth: 933.10 GB/s^*

Power specifications

Thermal design power: 165 W^*

System configuration

Detailed specifications

Specifications 1 · 13 fields

Foundry: TSMC^*
Vendor: NVIDIA^*
FP16 GFLOPS: 10,320^* GFLOPS
FP32 GFLOPS: 10,320^* GFLOPS
FP64 GFLOPS: 5,161^* GFLOPS
Product: NVIDIA PG506-232^*
Type: GPU^*
Release Date: 2021-04-12^*
Process Size (nm): 7^* nm
TDP (W): 165^* W
Die Size (mm^2): 826^* mm²
Transistors (million): 54,200^*
Freq (MHz): 930^* MHz

Specifications 2 · 30 fields

proc transistors: 54,200 million^*
proc die size: 826^* mm²
cores: 3,584^*
tensor cores: 224^*
sm count: 56^*
base clock: 930^*
boost clock: 1,440^*
memory clock: 1,215^*
ram: 24^*
name: NVIDIA PG506-232^*
ram type: HBM2^*
bus width: 3,072^*
bus interface: PCIe 4.0 x16^*
ram bandwidth: 933.1^* GB/s
fp16 performance: 10.32^*
fp32 performance: 10.32^*
fp64 performance: 5.161^*
tdp: 165^* W
suggested psu: 450^* W
slot width: Dual-slot^*
power connectors: 8-pin EPS^*
api cuda: 8^*
is retail board: No^*
gpi value: 12.38^*
vendor: nvd^*
manufacturer: NVIDIA^*
arch name: Ampere^*
card release date: 2021-04-12^*
proc foundry: TSMC^*
proc process size: 7^* nm

Specifications 3 · 35 fields

CUDA compute capability

8.0^*

OpenCL

3.0^*

Architecture

Ampere^*

Graphics interface

PCIe 4.0 x16^*

Memory bus

3,072^* bit

Reported shader processors

3,584^*

Reported theoretical FP16 peak

10.32^* TFLOPS

Scope and conditions

Arithmetic density: Not stated^*

Reported theoretical FP32 peak

10.32^* TFLOPS

Scope and conditions

Arithmetic density: Not stated^*

Reported theoretical FP64 peak

5.161^* TFLOPS

Scope and conditions

Arithmetic density: Not stated^*

Declared L1 cache configuration

192 KB (per SM)^*

Declared GPU L2 cache

24 MB^*

Reported display outputs

No outputs^*

Theoretical pixel fill rate

138.2 GPixel/s^*

External power connectors

8-pin EPS^*

Reported chip die area

826 mm²^*

Chip foundry

TSMC^*

Process node declaration

7 nm^*

Transistors

54,200 million^*

Graphics processor

GA100^*

Reported video memory

24^* GB

Rated memory bandwidth

933.1 GB/s^*

Video memory type

HBM2^*

Raster output units

96^*

Reported shading units

3,584^*

Card or module format

Dual-slot^*

Streaming multiprocessors

56^*

Suggested system PSU

450 W^*

2:4 tensor sparsity

Yes^*

Declared GPU TDP

165 W^*

Tensor cores

224^*

Theoretical texture fill rate

322.6 GTexel/s^*

Texture mapping units

224^*

GPU cores

3,584^*

Video memory

24^* GB

Video memory type

HBM2^*

* Provisional data