NVIDIA Tesla P10 videocard review

Tesla P10 videocard released by NVIDIA; release date: 13 Sep 2016. The videocard is based on Pascal microarchitecture codenamed GP102.

Core clock speed - 1025 MHz. Boost clock speed - 1493 MHz. Texture fill rate - 358.3 GTexel/s. Pipelines - 3840. Manufacturing process technology - 16 nm. Transistors count - 11800 million. Power consumption (TDP) - 250 Watt.

Memory type: GDDR5X. Maximum RAM amount - 24 GB. Memory bus width - 384 bit. Memory clock speed - 1808 MHz, 14.5 Gbps effective. Memory bandwidth - 694.3 GB/s.

Specifications (specs)

Essentials
Architecture	Pascal
Code name	GP102
Launch date	13 Sep 2016
Place in performance rating	not rated
Technical info
Boost clock speed	1493 MHz
Core clock speed	1025 MHz
Manufacturing process technology	16 nm
Peak Double Precision (FP64) Performance	358.3 GFLOPS (1:32)
Peak Half Precision (FP16) Performance	179.2 GFLOPS (1:64)
Peak Single Precision (FP32) Performance	11.47 TFLOPS
Pipelines	3840
Pixel fill rate	143.3 GPixel/s
Texture fill rate	358.3 GTexel/s
Thermal Design Power (TDP)	250 Watt
Transistor count	11800 million

Video outputs and ports
Display Connectors	No outputs
Compatibility, dimensions and requirements
Form factor	Single-slot
Interface	PCIe 3.0 x16
Length	267 mm, 10.5 inches
Recommended system power (PSU)	600 Watt
Supplementary power connectors	1x 8-pin
Width	97 mm, 3.8 inches
API support
DirectX	12 (12_1)
OpenCL	3.0
OpenGL	4.6
Shader Model	6.7 (6.4)
Vulkan
Memory
Maximum RAM amount	24 GB
Memory bandwidth	694.3 GB/s
Memory bus width	384 bit
Memory clock speed	1808 MHz, 14.5 Gbps effective
Memory type	GDDR5X