Krunkkracker Posted June 20, 2020 Share Posted June 20, 2020 Hello mouthbreathers and diverlungs! I have done testing on the scaling of performance from clockspeeds of the ram, cpu, and ring ratio. I posted this over on Reddit and figured people could make use of it here. First, System Specs: 9900k Aorus Pro z390 Gtx 1660 ti 3600 cl15 1.35v Gskill 4x8gb Mods: https://imgur.com/a/6tX7vNi For the 9900k, I tested with 4.8ghz core clock and 5.2ghz core clock. I also tested with 4.5ghz ring ratio and 4.9ghz ring ratio. For the ram, I tested with the push button XMP settings on the kit(3600 cl15-15-15-30 1.35v) and for the overclock I did 4000mhz cl15-16-16-35. For the actual testing, I had a several hundred cycle save file where I started a timer after pressing the unpause button. Once 5 cycles passed, I stopped the timer and the game. I decided to use how long it took 5 cycles to pass since it seemed much more logical to test. It's also much easier to test real time since FPS would involve doing also 0.1% and 1% lows as well as the average. If anyone wants the file, let me know and I'll try to find somewhere to upload it. Results(10x speed used): (All overclocked) https://imgur.com/a/PXrjujJ (Ring and ram underclocked) https://imgur.com/a/IMMZl0w (CPU underclocked) https://imgur.com/a/12Kcn98 (Ring underclocked) https://imgur.com/a/FKZMx9L (All underclocked, ram at 3.6) https://imgur.com/a/kCvsDX4 (8 core HT off) https://imgur.com/a/VbgOha6 (6 core HT off) https://imgur.com/a/IUhsWXF Extra: https://imgur.com/a/RFgZgak 5m55s 720p testing. ONI scales very well with core clock but not ring ratio. The ram speed also improves performance quite a bit, which means ram speeds and timings matter considerably when looking for the best ONI performance. What was most surprising was that ONI gets considerably more performance with HT disabled, something definitely to keep in mind if anyone is looking for "free" gains. Monitor resolution seems to not matter, at least on my system. Link to comment Share on other sites More sharing options...
Smife Posted June 21, 2020 Share Posted June 21, 2020 It will lag no matter what. Squeezing out a couple of seconds doesn't solve anything and barely noticeable. Link to comment Share on other sites More sharing options...
Gurgel Posted June 21, 2020 Share Posted June 21, 2020 5 hours ago, Smife said: It will lag no matter what. Squeezing out a couple of seconds doesn't solve anything and barely noticeable. That is so. I think the only potential for more speed is using more threads with loose coupling and thereby more cores. For example save/load should be able to do that, as input/target are frozen. Also, thermal calculation should basically scale unlimited. Electrical gas and liquid grids should be doable with one thread per segment, with cut-points at certain elements, maybe even at gas or liquid bridges. Planning is probably the thing where that only works on plan execution )one thread per dupe), not on plan creation. Link to comment Share on other sites More sharing options...
Krunkkracker Posted June 21, 2020 Author Share Posted June 21, 2020 14 hours ago, Smife said: It will lag no matter what. Squeezing out a couple of seconds doesn't solve anything and barely noticeable. That's okay! Small things add up over time, I'm sure every who plays Oxygen Not Included is aware of this. 10 seconds saved per cycle on a 400 cycle playthrough is over an hour saved. Overclocking, sweeping, and optimizing the base for performance could easily save hours of time per world playthrough. I did this originally because I was playing on my laptop, which I'm sure would take close to 10 minutes if not over after running this experiment. Hopefully someone gets something out of this, good to see that overclocking does make a decent difference. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.
Please be aware that the content of this thread may be outdated and no longer applicable.