I'm running the blender benchmarks from the open data project (https://opendata.blender.org/) and I'm confused about the results. It ran three tests and reported the results of each, and while the sum of the three scores is roughly equal to the score reported on the website for my hardware, each individual test is only about 1/3 of the score. I'm guessing they should not be summed since the score is samples/second and it does not make sense to sum a rate of things that happened in series (so I would expect to see the average reported as the overall score), so I am concerned I have a driver issue. However, I would like to confirm my interpretation of the benchmark before I start more drastic measures than reinstalling the driver.
I did look in the about tab of the opendata.blender website and it did not explain how the results should be interpreted other than implying that the ones for each scene/test are unrelated.
Note about relevance: I'm assuming this is on-topic as it is a question about official blender software and is not tied to any specific hardware, nor is it soliciting a hardware recommendation.
