Crysis Remastered Performance Transformed - Patch 1.3.0 Delivers Huge Performance Gains
CPU Optimisations - Higher framerates when RT is enabled
Published: 26th November 2020 | Source: OC3D Internal Testing | Price: |
CPU Optimisations - Higher framerates when RT is enabled
Patch 1.2.0 has already increased Crysis Remastered's performance in CPU-limited scenarios and increased framerate stability, but patch 1.3.0 delivers performance gains that are genuinely transformative for Crysis Remastered.
At 1080p High settings, the day-1 version of Crysis Remastered delivered 1st percentile framerates of 52 FPS on an RTX 2080 Ti and an average framerate of 83.7 FPS in our test area within Crysis' 3rd level (Relic). Larger performance dips can be seen in specific areas of Crysis, such as the town area in Level 2 (Recovery). Still, we have found that our Relic test area can deliver more repeatable benchmark results.
While patch 1.3.0 can still see our test system dip below 60 at 1080p High settings in the town area of Recovery, framerates in this area are now in the high 50s, rather than the low 40s on our Ryzen 9 3950X processor. In our Relic test scene, 1st percentile framerates increase from 52 to 6, while average framerates increase from 83.7 FPS to 96.4 FPS.
Regardless, Crysis Remastered remains a CPU-limited title in many areas at High and above settings. That said, 60+ FPS framerates are now a lot more achievable across the game.
Most Recent Comments
Ooo, ya sexy dancer.
Nice one Mark, thanks for putting the effort in mate !Quote
I realize this harder to do than to slap a card in a build and watch a FPS counter but perhaps this is so easy (albeit time consuming) that there is room to get into the other parts?
Once in a while, yes there is a comment about poorly optimized game but I think it should be taken a big step further. As it is, the focus is always on the hardware, but as we already know the performance levels of the cards then we get too little out of the game performance reviews. Perhaps title is slightly better for team red .. so what? The interesting part would be to evaluate the games optimizations like it is being done for new hardware. It should not be a comment, it should be analyzed and ranked in table form. Obviously very difficult, but who better to do it than the crew here?
Sure it may fail, but it would be a chance of doing something new. And to put the pressure on sloppy coders just like the pressure is on the hardware builders when they cut corners or make poor design choices.
And in the end, optimizations in software tends to be much more important than hardware. Just look at the 'demoscene' to truly understand what is possible using optimized code.
Yes I do understand this is a hardware review site :-) But measuring HW performance without even considering the 'input' is missing a important part of the 'ecosystem' of performance.Quote
|
Originally Posted by MiNo
I'm missing a evaluation of (any) games own performance - as in is it sloppy code or optimized code? This article puts much needed focus on it, but not enough.
I realize this harder to do than to slap a card in a build and watch a FPS counter but perhaps this is so easy (albeit time consuming) that there is room to get into the other parts? Once in a while, yes there is a comment about poorly optimized game but I think it should be taken a big step further. As it is, the focus is always on the hardware, but as we already know the performance levels of the cards then we get too little out of the game performance reviews. Perhaps title is slightly better for team red .. so what? The interesting part would be to evaluate the games optimizations like it is being done for new hardware. It should not be a comment, it should be analyzed and ranked in table form. Obviously very difficult, but who better to do it than the crew here? Sure it may fail, but it would be a chance of doing something new. And to put the pressure on sloppy coders just like the pressure is on the hardware builders when they cut corners or make poor design choices. And in the end, optimizations in software tends to be much more important than hardware. Just look at the 'demoscene' to truly understand what is possible using optimized code. Yes I do understand this is a hardware review site :-) But measuring HW performance without even considering the 'input' is missing a important part of the 'ecosystem' of performance. |
For the type of indepth detail you want would take way more time that it's worth for a review site mainly about the general performance of said hardware, There are places dedicated to this though, Digital Foundry being the main one who do get detailed information from game developers about various optimisations etc...Quote
CSS Valid