Had a bit of fun running my two main systems through the benchmark this evening. Quite interesting results as well.
Test systems:
Desktop - Ryzen 5900X (CO -15), 16GB DDR4 3200C14, XFX RX6800 Speedster ~ 2200/16000
Laptop (Legion 7) - Ryzen 5800H, 32GB DDR4 3200C20 (2Rx8), RTX 3080M 16GB (150-165w) ~ 1750/14000
Both systems running latest drivers and connected to the same 3440x1440 144Hz monitor.
Settings in game are a mix of High, Ultra with the Volumetric fog / cloud options on Medium. Ray Tracing wise I had just enabled the lighting option at the medium pre-set.
Ray Tracing on:
Ray Tracing off:
Overall the 6800 seemingly struggles with any form of RT when compared to the 3080M. FSR balanced was the only option whereby it was what I would consider playable (without locking at 30FPS) and retain decent image quality. The 3080M did better with FSR than DLSS in the benchmark run but I personally found the DLSS image quality in motion to be a bit "cleaner". Neither were what I consider playable sans DLSS/FSR but the 6800 really struggled to come close to hitting even 30FPS.
With RT off the 6800 does much better, beating the 3080M - albeit not by a massive margin, at both native resolution and with FSR set to Ultra quality. DLSS on the 3080M was a little slower but again a little "cleaner" to my eye when in motion.
Need to play the game more on both systems to see how they fair in different scenarios actual in game. A quick run around the opening area from V's apartment returned similar results across both systems as the benchmark.
My main takeaway from the testing I have done so far is:
- 6800 does a fine job with RT switched off and enabling FSR Ultra Quality.
- 3080M is quite happy running with RT set to medium and DLSS on auto/balanced. I may try and squeeze it a little harder to see if I can enabled the other RT options and maintain 50+ FPS in gameplay. This may require lots of time spent messing with undervolting which I am totally ok with.
- I am guessing that CP2077 generally favours Nvidia hardware as I was expecting the delta between the GPU's to be a little greater than I experienced (both in the benchmark and in game).