Original Post
Hi everybody
I'd be very grateful if you could help us gathering OpenGL performance data with 3 low level benchmarks we made to test various procedural and non-procedural rendering approaches across different hardware.
There are two procedural-type tests: grass and buildings, that primarily measure achievable triangle throughput vs the instance mesh size.
Here are some performance results from our procedural grass test:
Raw performance Mtris/s
Performance/price, Mtris/s per dollar (using http://www.videocardbenchmark.net/gpu_value.html for prices)
The third test (cubes) is focusing on finding the optimal settings for non-procedural model rendering. This one is measuring the achievable triangle throughput and number of draw calls in several rendering modes and with varying mesh sizes.
There are more than 150 test combinations total, each taking roughly 5 seconds, so it can take 13-15 minutes total.
If you would like to help us gathering performance data, the test suite can be downloaded here:
Outerra perf tests download
After unpacking run the three test batches there: grass_test.bat, buildings_test.bat and cubes_test.bat
When each tests completes, you will see your score, and you will be asked to send the results to us so we can incorporate results from different hardware into our results.
Please note that the test program requires OpenGL 4.x drivers and it may not run on some older cards that only support OpenGL 3.x. Also, the test program requires 64-bit Windows.
On some older AMD cards the cubes test will likely crash the drivers, also not all of the modes work (bindless textures will crash any AMD card/driver currently, and basevertex mode won't work on AMD either).
Thanks for your help! Complete results will be published on Outerra blog together with comments.
Sources for the tests are at https://github.com/hrabcak/draw_call_perf
I'd be very grateful if you could help us gathering OpenGL performance data with 3 low level benchmarks we made to test various procedural and non-procedural rendering approaches across different hardware.
There are two procedural-type tests: grass and buildings, that primarily measure achievable triangle throughput vs the instance mesh size.
Here are some performance results from our procedural grass test:
Raw performance Mtris/s
Performance/price, Mtris/s per dollar (using http://www.videocardbenchmark.net/gpu_value.html for prices)
The third test (cubes) is focusing on finding the optimal settings for non-procedural model rendering. This one is measuring the achievable triangle throughput and number of draw calls in several rendering modes and with varying mesh sizes.
There are more than 150 test combinations total, each taking roughly 5 seconds, so it can take 13-15 minutes total.
If you would like to help us gathering performance data, the test suite can be downloaded here:
Outerra perf tests download
After unpacking run the three test batches there: grass_test.bat, buildings_test.bat and cubes_test.bat
When each tests completes, you will see your score, and you will be asked to send the results to us so we can incorporate results from different hardware into our results.
Please note that the test program requires OpenGL 4.x drivers and it may not run on some older cards that only support OpenGL 3.x. Also, the test program requires 64-bit Windows.
On some older AMD cards the cubes test will likely crash the drivers, also not all of the modes work (bindless textures will crash any AMD card/driver currently, and basevertex mode won't work on AMD either).
Thanks for your help! Complete results will be published on Outerra blog together with comments.
Sources for the tests are at https://github.com/hrabcak/draw_call_perf
