- Joined
- Mar 3, 2018
- Messages
- 1,713
In response to the recent controversy surrounding the 9900k benchmarks by Principled Technologies, AMD passed out some slides outlining problems with those benchmarks, as well as general "best benchmarking practices". Aside from the numerous issues with the first round of testing, AMD also noted issues with Principled Technologies 2nd, updated round of testing. According to AMD, the Multicore Enhancement auto-overclocking setting was still unclear, the memory timings were still "suspect" and testing methodology specifics and other relevant system configuration settings still weren't addressed. In addition to the criticism, AMD outlined good, general benchmarking practices. AMD says that reviewers should make sure to sanitize the OS before running benchmarks, make sure the CPUs are running at their stock, warrantied specifications, and run repeatable, consistent benchmarks at least five times, among other things.
Check out the full slides by clicking the images here.
Never stop asking: "does this config, test, and result make sense to me as a consumer?"
Check out the full slides by clicking the images here.
Never stop asking: "does this config, test, and result make sense to me as a consumer?"