Only ONE tech review website deemed it worthy of a re-do based on inconsistencies of other site reviews, and yep, that's Anandtech.
Tom's Hardware admitted right in their review to having a problem with their testing methodology, namely that they made the mistake of not realizing that their X470 motherboard was already patched for Spectre, so they didn't apply patched firmware to their Intel motherboard, potentially providing Intel's CPUs with an unfair advantage in their test results. Whether that made any notable difference to how those processors performed is yet to be seen, but it's certainly possible. Unfortunately, they tucked that information away on the test setup page instead of posting it somewhere more prominent, so many readers probably missed it...
Unfortunately, we were only made aware that Variant 2 mitigations are present in our X470 board's firmware just before launch, precluding us from re-testing the Intel platforms with patches applied. We're working on this now, and plan to post updated results in future reviews.
The lack of Spectre Variant 2 patches in our Intel results likely give the Core CPUs a slight advantage over AMD's patched platforms. But the performance difference should be minimal with modern processors.
The gist of the matter is, Spectre and Meltdown mitigations decrease the performance of IO operations on Intel platforms. The HPET (High Precision Event Timer) is mapped in such a way, accessing it constitutes an IO operation. Forcing Windows to use HPET as it's primary timer has a major impact on Intel's CPU performance in cases where the software makes use of the timer. The difference hasn't been quantified between pre- and post-patched machines yet, but it's not an unfair assessment to say there won't be a performance difference, and the final result won't change anyway. HPET being used as the primary timer in Windows with a patched Intel system seems to come with a very high performance penalty.
It would be a nice addition to any benchmark setup pages going forward to include the state of the HPET timer settings, both in UEFI and the Windows setting used. Anandtech also points out that installing software can change the HPET setting in Windows, so the step at which the Windows HPET setting is looked at and potentially modified by the reviewer should obviously be after all pertinent software has been installed.
No conspiracies or intentional biases seem to have been necessary for the interesting results found by the Anandtech reviewers.
An honest mistake. Ian works hard, so this is surely a bit disappointing for him.
We test with the default OS behavior, which is HPET disabled. We also enforce this with a hard disable in our test scripts.
As an aside, there are several tools that you can use to check the resolution of the windows timer. It is important to make sure that this in-built timer is consistent across test platforms, as well. I check the resolution on every new test platform as SOP to ensure consistent and accurate benchmark results across platforms.
Interestingly, the primary reason that HPET still exists is to ensure backwards compatibility with older programs that rely on it, the same old kludge as always.