@Rigg
Those numbers are . . . interesting. Why'd they get such a poor minfps for the 7800X3D in the last slide, but not in any of their other slides? Am I missing something?
Yeah, the BG3 results in the 9600x review are a pretty big red flag. My best guess? A combination of bad testing methodology, limited time to get a review out, a lack of curiosity/due diligence, and possibly buggy AMD Adrenaline drivers. After spotting the obviously bizarre result in TYC's 9600X review, I made a good faith attempt to
reproduce his results on my 7800X3D/7900 XT system.
After posting my results (which I think clearly demonstrate that that TYC's 9600X review data and methodology is flawed) it dawned on me that I hadn't updated my GPU drivers in a while on my windows install. I've been running Bazzite on this system because Microsoft can redact off. I hadn't booted my windows 10 install in a few months up until that testing, so I went ahead upgraded my GPU drivers to 24.7.1. I wanted to see if the old driver I was using was affecting my results.
BG3 began to hard crash my system upon launching the game after the update. These are the only crashes of any kind I've had on this system since I built it at the end of last year. I haven't tuned the system (running 6000 MT/s EXPO and auto everything else) at all, but I still stress tested as if I had when I built it. Up until this driver update the system has been solid as a rock. After search results lead me to AMD's support site, and finding other users having the same/similar issues with the last two Adrenaline releases, I rolled back to 24.5.1. Everything seems fine now and I'm consistently getting around 80 FPS 0.1% lows in the BG3 exploding mushroom test with 24.5.1. Clearly the latest drivers are problematic in this game so it may have something to do with his results. I did at least a dozen test runs with no further crashing on 24.5.1.
I don't think you're missing anything. My conclusion is TYC's testing isn't trustworthy, and the BG3 data in his 9600X review is hot garbage. The results from his latest 6200 MT/s C34 7800X3D/7900 XTX testing correlates pretty closely to my testing. The fact he makes no mention of the difference between the data in the two videos doesn't bode well for his trustworthiness as a reviewer. I did a skim of the comments in the memory vid and was surprised nobody pointed out the discrepancy. Any TYC testing should be taken with a mountain of salt IMO.