Benchmark data sets back

(L) [2017/02/10] [tby knus] [Benchmark data sets] Wayback!

Hi,
Is there a commonly used dataset for benchmarking ray tracing? I know about the BART for dynamic tracing, but is there something widely used for static scenes? Preferably a model-only, not compileable code to run on 'any' machine. The reason I ask is that I am (for a hobby) developing a dedicated hardware ray tracer in an FPGA. It is definitely not real time, and there is no API to program it yet. It does BVH tree iteration in hardware, intersects triangles and shading (shadow, reflection, refraction etc). Currently no texturing but that is on the to-do list.
The current prototype runs at 33 MHz, and is able to trace the 'Fairyforest' scene at about 1 FPS in 256x256 resolution (diffuse lighting, 1 light, 1 shadow ray per pixel).
-knus-

(L) [2017/02/10] [tby papaboo] [Benchmark data sets] Wayback!

You can find a lot of the standard scenes here [LINK http://graphics.cs.williams.edu/data/meshes.xml].
I'm not aware of any generally accepted performance test. Performance depends on a lot of factors, BVH building times and quality, traversal speed, resolution, camera transformation, shading complexity, selection / coherence of secondary rays. If you just want to improve one of those, then you don't need a general reference, you can just use your own system and time things before and after.

(L) [2017/02/10] [tby knus] [Benchmark data sets] Wayback!

I'm using lots of different scenes to do internal benchmarking and enhancements, but I would like to compare my hardware (maybe scale up number of cores and core speed) to see how efficient it is compared to other state-of-the-art solutions.
-knus-

(L) [2017/02/10] [tby toxie] [Benchmark data sets] Wayback!

There was an attempt once to have a standardized environment where one could plug "any" ray tracing engine in and test exactly the same stuff:
[LINK http://bwfirt.sourceforge.net/]

(L) [2020/04/12] [ost by JasonSmith] [Benchmark data sets] Wayback!

I think the collection of meshes that used to be at Williams is now here:
http://casual-effects.com/data/index.html