Synthetic OpenCL performance of NVIDIA’s upcoming GeForce GTX 1080 graphics card was revealed by CompuBench database.
NVIDIA GeForce GTX 1080 has average OpenCL performance
It’s worth nothing that CompuBench score is based just on 9 separate results performed using different testing platforms. My point is that this value will change and become more stable over the next few weeks. We are specifically using ‘Radeon Fury X’ result page for comparison, because ‘Radeon Fury Series’ page also gathers results from Fury non-X and R9 Nano too.
We are observing 10-14% boost over GM200 and Fiji based solution, which in my opinion is not much, considering how well this card performs in pure gaming-related benchmarks.
Later in this post you will find graphs from CB website, which unfortunately are polluted with strange cards like ‘Kung Fury’. In any case GeForce GTX 1080 only tops those charts in two benchmarks: Bitcoin Mining and Face Detection.
| NVIDIA GeForce GTX 1080 OpenCL Performance | |||
|---|---|---|---|
| NVIDIA GeForce GTX 1080 | NVIDIA GeForce GTX 980 Ti | AMD Radeon R9 Fury X | |
| Face Detection | 100% | 85% | 63% |
| TV-L1 Optical Flow | 100% | 79% | 97% |
| Ocean Surface Simulation | 100% | 88% | 118% |
| Particle Simulation – 64K | 100% | 101% | 68% |
| T-Rex | 100% | 99% | 90% |
| Video Composition | 100% | 94% | 110% |
| Bitcoin Mining | 100% | 73% | 81% |
| AVERAGE | 100% | 88% | 90% |
Source: CompuBench via OCN