NVIDIA OpenCL CUDA
AMD EPYC 7551 32-Core testing with a GIGABYTE MZ31-AR0-00 and NVIDIA GeForce GTX 1070 TiMicrosoft Basic Display on Microsoft Windows Server 2016 Essentials Build 14393 via the Phoronix Test Suite.


GeForce GTX 1070 Ti: 

	Processor: AMD EPYC 7551 32-Core @ 2.00GHz (32 Cores / 64 Threads), Motherboard: GIGABYTE MZ31-AR0-00, Memory: 8 x 4096 MB 2666MHz 9ASF51272PZ-2G6E1, Disk: 466GB Samsung SSD 960 EVO 500GB, Graphics: NVIDIA GeForce GTX 1070 TiMicrosoft Basic Display, Network: Realtek PCIe GBE Family + Microsoft ISATAP

	OS: Microsoft Windows Server 2016 Essentials Build 14393, Kernel: 10.0 (x86_64), Display Driver: 23.21.13.910110.0.14393.0, Vulkan: 1.0.65, Compiler: GCC 7.1.0, File-System: NTFS, Screen Resolution: 25602560x14401440


Blender 2.79a
Blend File: BMW27 - Compute: CUDA
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 1137.85 |================================================


Blender 2.79a
Blend File: BMW27 - Compute: OpenCL
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 1293.84 |================================================


Blender 2.79a
Blend File: Classroom - Compute: CUDA
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 740.71 |=================================================


Blender 2.79a
Blend File: Fishy Cat - Compute: CUDA
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 2953.08 |================================================


Blender 2.79a
Blend File: Barbershop - Compute: CUDA
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 2083.97 |================================================


Blender 2.79a
Blend File: Classroom - Compute: OpenCL
Seconds < Lower Is Better
GeForce GTX 1070 Ti . 751.99 |=================================================