It's definitely better with the 14 drivers than the previous ones. What I mean is if I set 8X AA, HDR, maximum number of visible cars, all settings turned to the highest they go in the video settings and don't turn on multiview, go to Malaysia, start non-private practice, it hovers around 30 fps - regardless of the type of car - . As I mentioned earlier crossfire (so a 2nd 290X) does nothing.
R9-290 or GTX-770 4GB ??
- Thread starter ViSo
- Start date
speed1
Banned
Maybe that's the CF configuration, even if you use only one ? Actually your sys should be much better than mine with one card ( 290 non X here ) but you have worse results.
I haven't tested malaysia but all the current content as silverstone with various vehicles and 21 opponents, and the result as I said ~ 60 fps average with a short run of 3 rounds.
The only thing I have not yet tested on max level is Indy.
I haven't tested malaysia but all the current content as silverstone with various vehicles and 21 opponents, and the result as I said ~ 60 fps average with a short run of 3 rounds.
The only thing I have not yet tested on max level is Indy.
Yes, that is with one card. A 2nd one doesn't have any effect, I complained to Amd a million times (a long time ago with a 4970 as well with flickering textures in Gothic 3 - this was never fixed), but so far nothing. I don't think Amd actually has the resources to address niche games such as rFactor2. There's a blog post from a Valve developer about the driver quality of different vendors. Even though it is about OpenGL, it reflects on the whole company as a whole I would say... http://richg42.blogspot.com/2014/05/the-truth-on-opengl-driver-quality.html . Vendor B is Amd, Vendor A is Nvidia. Basically he says Amd makes great hardware but crap software, Nvidia makes great hardware and semi-crap software 
.
I agree, some fps issues could be setup related, who knows what the differences are btw your setup and mine, except for the actual systems, such as software that is running, processes etc. I did notice that some tracks are better (Mills for example is better by 20 fps or so), but I figured I'll go for the lowest denominator. Either way, I remember having issues with two gtx680s in sli, those cards are in a different computer now, and I don't want to switch things around just to test them again - however I think I remember sli was fixed in an older build.
For single card setups I would say it's ok to get a new Amd card, but one that actually gets improvements from the new Amd drivers (e.g don't get a 6 series or even 7 series card such as 7970 or whatever, since they might drop support for them sooner). Who knows what happens in the future, Amd does many "fix one break 10 things" driver releases.
.I agree, some fps issues could be setup related, who knows what the differences are btw your setup and mine, except for the actual systems, such as software that is running, processes etc. I did notice that some tracks are better (Mills for example is better by 20 fps or so), but I figured I'll go for the lowest denominator. Either way, I remember having issues with two gtx680s in sli, those cards are in a different computer now, and I don't want to switch things around just to test them again - however I think I remember sli was fixed in an older build.
For single card setups I would say it's ok to get a new Amd card, but one that actually gets improvements from the new Amd drivers (e.g don't get a 6 series or even 7 series card such as 7970 or whatever, since they might drop support for them sooner). Who knows what happens in the future, Amd does many "fix one break 10 things" driver releases.
speed1
Banned
The upcoming 14.6 is interesting, perhaps it also helps your problem.
Performance update
Catalyst 14.6 will also provide performance boost to Watch Dogs, Thief and Murdered Soul Suspect (which is not yet released). You should expect 25-28% better performance in Watch Dogs with the new drivers. Catalyst 14.6 also has new CrossFire profiles specificly for WD. AMD claims it will provide up to 99% scaling to Watch Dogs.
http://videocardz.com/50583/amd-catalyst-14-6-brings-eyefinity-mixed-resolutions-support
http://www.anandtech.com/show/8058/...ailable-adds-new-eyefinity-functionality-more
Performance update
Catalyst 14.6 will also provide performance boost to Watch Dogs, Thief and Murdered Soul Suspect (which is not yet released). You should expect 25-28% better performance in Watch Dogs with the new drivers. Catalyst 14.6 also has new CrossFire profiles specificly for WD. AMD claims it will provide up to 99% scaling to Watch Dogs.
http://videocardz.com/50583/amd-catalyst-14-6-brings-eyefinity-mixed-resolutions-support
http://www.anandtech.com/show/8058/...ailable-adds-new-eyefinity-functionality-more
Instant crash with the new 14.6 beta drivers when I try to open the catalyst control panel. Reinstalled, uninstalled, reinstalled, switch from 3 screens to 1 screen, did Voodoo, nothing. Have to wait for the non-beta release.
Edit: I managed to get past the Catalyst Control Panel crash by not installing it, and just installing the new actual drivers. Remains to be seen if they are any better.
Edit 2: There is definitely a good improvement, I now get about 40 fps at the back of the grid at Malaysia with everything maxed out except FXAA and Multiview, full heavy rain, this settles into 40-50 while racing. I believe this is about 10-15 more than before. With the overclock I mentioned earlier I get 50 fps at the start and 50-60 when racing. This is with everything set to the highest possible (max cars on screen etc.), 8X AA but no FXAA, and triple screen (5760*1080). Should be ok to get more than 60 fps all the time with some settings lowered, but I don't have to lower as many as before.
Unfortunately Crossfire is still KO. Also, since these are beta, who knows what/where broke besides the Catalyst Control Panel. I'll keep them for now since I was able to get past the Catalyst Control Panel issue and they don't seem to be any worse than v14.4 .
If anybody tries them, you can uncheck the Catalyst Control Panel in the driver install if you pick the customized install option. The old one seems to work fine with the new drivers, some of the eyefinity settings look different but they work with the older Catalyst Control Panel.
Edit: I managed to get past the Catalyst Control Panel crash by not installing it, and just installing the new actual drivers. Remains to be seen if they are any better.
Edit 2: There is definitely a good improvement, I now get about 40 fps at the back of the grid at Malaysia with everything maxed out except FXAA and Multiview, full heavy rain, this settles into 40-50 while racing. I believe this is about 10-15 more than before. With the overclock I mentioned earlier I get 50 fps at the start and 50-60 when racing. This is with everything set to the highest possible (max cars on screen etc.), 8X AA but no FXAA, and triple screen (5760*1080). Should be ok to get more than 60 fps all the time with some settings lowered, but I don't have to lower as many as before.
Unfortunately Crossfire is still KO. Also, since these are beta, who knows what/where broke besides the Catalyst Control Panel. I'll keep them for now since I was able to get past the Catalyst Control Panel issue and they don't seem to be any worse than v14.4 .
If anybody tries them, you can uncheck the Catalyst Control Panel in the driver install if you pick the customized install option. The old one seems to work fine with the new drivers, some of the eyefinity settings look different but they work with the older Catalyst Control Panel.
Last edited by a moderator:
speed1
Banned
That is a pity, i'm sorry for you that CF doesn't work.
With how many vehicles have you tested ? Then I would suggest you to use the more optimized tracks such as example Silverstone, LochD or Mores because I don't know if Malaysia is suitable, and also the cars make a difference. The open wheeler had a better performance as I've tested the last, as for example the Civic, and we had used the civic for the last bench's we did to compare.
About Indy content I'm not sure to what extent it is finished.
edit: Driver 14.4 / Resolution 5760x1080 / Track Silverstone Int. / 20 AI / Level 8 AA / 16xAF / No MultiV. / No FXAA / Display-GFX Settings MAX / Weather Sunny / realistic HDR / AI Driver Start Position 21 / 3 Turns with trailing AI in front
No OC / CPU PhenomII@3200 Standard / GPU 947Mhz/1250Mhz Standard / 16GB DDR3 1600Mhz
Frames, Time (ms), Min, Max, Avg
16418, 299974, 42, 68, 54.731
With how many vehicles have you tested ? Then I would suggest you to use the more optimized tracks such as example Silverstone, LochD or Mores because I don't know if Malaysia is suitable, and also the cars make a difference. The open wheeler had a better performance as I've tested the last, as for example the Civic, and we had used the civic for the last bench's we did to compare.
About Indy content I'm not sure to what extent it is finished.
edit: Driver 14.4 / Resolution 5760x1080 / Track Silverstone Int. / 20 AI / Level 8 AA / 16xAF / No MultiV. / No FXAA / Display-GFX Settings MAX / Weather Sunny / realistic HDR / AI Driver Start Position 21 / 3 Turns with trailing AI in front
No OC / CPU PhenomII@3200 Standard / GPU 947Mhz/1250Mhz Standard / 16GB DDR3 1600Mhz
Frames, Time (ms), Min, Max, Avg
16418, 299974, 42, 68, 54.731
Last edited by a moderator:
The 14,6 is a bit faster than the 14,4 with my 7970 @5040x1050.
With 25 Ai on the Nurburg Ring 1.0 i reach 55 instead of 51fps @ the end of the grid and nearly the whole time constant 60fps after the start. With the 14.4 i got often drops down to 55 fps or below.
Not so much but unexpected.
Thanks again Amd.
With 25 Ai on the Nurburg Ring 1.0 i reach 55 instead of 51fps @ the end of the grid and nearly the whole time constant 60fps after the start. With the 14.4 i got often drops down to 55 fps or below.
Not so much but unexpected.
Thanks again Amd.
zonder
Registered
Driver 14.4 / Resolution 1920x1080 / Track Silverstone /Honda Civic 29 AI / Level 3 AA / 8xAF / No MultiV. / FXAA / Sync To - Video / Display-GFX Settings all max / Weather Variable
No OC / CPU FX8350 / GPU R9 290 reference / 8GB DDR3 1800Mhz
And I'm hitting 60 FPS, going in low 50's sometimes,, averaging at 59 after 20 min bench in Fraps.
Is limiting FPS advised for racing sims, I have 60Hz Display and belive there is no reason to go over 60 fps
No OC / CPU FX8350 / GPU R9 290 reference / 8GB DDR3 1800Mhz
And I'm hitting 60 FPS, going in low 50's sometimes,, averaging at 59 after 20 min bench in Fraps.
Is limiting FPS advised for racing sims, I have 60Hz Display and belive there is no reason to go over 60 fps
Last edited by a moderator:
DrR1pper, yes. Not sure why Zonder isn't hitting the numbers I am, my guess it's the Honda, as it uses what 4k textures someone wrote? I find the Hondas also bring me to 70-100 frames while racing.. Try running with vsync off and see how it feels for you. Thee tearing is not bad with this card.. or the stuttering
DrR1pper
Registered
Yeah, not a fair comparison running different cars and zonder's using Vsync.
Zonder, i would not be using the ingame vsync video though. If tearing is a real issue for you (as it is for me) there is an alternative vsync solution but only for nvidia users that i know of. It's called Adaptive vsync (perhaps there is now an amd equivalent, if not through the CCC then maybe through something like rivatuner). In short, it's vsync only when the fps is at or above your native refresh rate. This way you don't get those choppy moments that happen when your fps dips below 60 (to say 50fps) which causes vsync to churn out 30fps instead as well as the ugly screen tearing.
Having said that, i'm not 100% sure if the ingame vsync video option works in the same way (with regards to halving the fps to 30 if your fps dips below 60) because as i know from prior experience with it, it is drop dead smooth. However, it comes at a big cost to latency...around 200ms (if not more). When i was first exposed to this as were many of us (thanks to KeiKei) back in the earlier days of rf2 beta, i noticed that my lap times around LRP in the skippy (which i had been driving for a solid 1-2 months continuously at that time) suddenly improved around 0.35-0.5s (which may not sound a lot but considering i had plateaued for a couple of weeks prior) and my lap time consistency shot through the roof (compared to how i was driving with vsync video on). I started to become aware of how i was having to guess the correct moment to initiate turn ins into the corners and how they should be modulated throughout the corners by shear predetermined guesswork when i was using vsync video. It was also responsible for pretty much all the unexplainable un-savable yet only slight/relatively-small oversteer moments. It was only when i disabled it that i realised this and could start making decisions in the moment which made driving both a heck of a lot easier and more natural feeling (hence the consistency and lap time improvements).
Next time your in the sim, get into any car and oscillate your wheel left/right (best done quickly to easily highlight the issue) and see how much of a delay is happening onscreen with the virtual steering wheel. Then repeat the experiment with vsync video off (perhaps also making sure you don't have any vsync on in CCC as well).
Problem though is that even with Adaptive sync (for nvidia users) there is in fact some lag that still remains (thanks to someone on the forum for making me aware of this). It's worse for 60hz users and much less so for 120hz users (i.e. me) but still no where near as bad as vsync video's lag.
Here is KeiKei's thread on input lag (relating to vsync in this case), well worth the read/look: http://isiforums.net/f/showthread.p...ut-lag-(but-in-expense-of-details)?highlight=
It's a shame that this sort of issue exists. Nvidia has now made Gsync monitors that remove this issue completely, offering tear free gaming with minimal visual delay/input lag. But not exactly an option if you don't have an nvidia card either and not a cheap solution.
Those provided by KeiKei however should work well but as always, something needs to be compromised.
Zonder, i would not be using the ingame vsync video though. If tearing is a real issue for you (as it is for me) there is an alternative vsync solution but only for nvidia users that i know of. It's called Adaptive vsync (perhaps there is now an amd equivalent, if not through the CCC then maybe through something like rivatuner). In short, it's vsync only when the fps is at or above your native refresh rate. This way you don't get those choppy moments that happen when your fps dips below 60 (to say 50fps) which causes vsync to churn out 30fps instead as well as the ugly screen tearing.
Having said that, i'm not 100% sure if the ingame vsync video option works in the same way (with regards to halving the fps to 30 if your fps dips below 60) because as i know from prior experience with it, it is drop dead smooth. However, it comes at a big cost to latency...around 200ms (if not more). When i was first exposed to this as were many of us (thanks to KeiKei) back in the earlier days of rf2 beta, i noticed that my lap times around LRP in the skippy (which i had been driving for a solid 1-2 months continuously at that time) suddenly improved around 0.35-0.5s (which may not sound a lot but considering i had plateaued for a couple of weeks prior) and my lap time consistency shot through the roof (compared to how i was driving with vsync video on). I started to become aware of how i was having to guess the correct moment to initiate turn ins into the corners and how they should be modulated throughout the corners by shear predetermined guesswork when i was using vsync video. It was also responsible for pretty much all the unexplainable un-savable yet only slight/relatively-small oversteer moments. It was only when i disabled it that i realised this and could start making decisions in the moment which made driving both a heck of a lot easier and more natural feeling (hence the consistency and lap time improvements).
Next time your in the sim, get into any car and oscillate your wheel left/right (best done quickly to easily highlight the issue) and see how much of a delay is happening onscreen with the virtual steering wheel. Then repeat the experiment with vsync video off (perhaps also making sure you don't have any vsync on in CCC as well).
Problem though is that even with Adaptive sync (for nvidia users) there is in fact some lag that still remains (thanks to someone on the forum for making me aware of this). It's worse for 60hz users and much less so for 120hz users (i.e. me) but still no where near as bad as vsync video's lag.
Here is KeiKei's thread on input lag (relating to vsync in this case), well worth the read/look: http://isiforums.net/f/showthread.p...ut-lag-(but-in-expense-of-details)?highlight=
It's a shame that this sort of issue exists. Nvidia has now made Gsync monitors that remove this issue completely, offering tear free gaming with minimal visual delay/input lag. But not exactly an option if you don't have an nvidia card either and not a cheap solution.
Those provided by KeiKei however should work well but as always, something needs to be compromised.
Last edited by a moderator:
zonder
Registered
Ohh yeah thanks DrR1pper, just turned Sync to video off, and FPS is going 'SkyHigh' well it goes higher also jumps are bigger 2014-06-08 15:55:06 - rFactor2
Frames: 68673 - Time: 1000000ms - Avg: 68.673 - Min: 46 - Max: 109
Tearing is not really visible at least I couldn't spot it but FPS fluctuation is noticible, sometimes event feels like stutter.
Also just found out I accidentally 'lied' in previous post, my settings weren't all Max, must have missed that, opponent detail was Medium, just found out I can't load Silverstone with all settings far right and full grid of Honda cars, it gives me Memory allocation Error.
Anyway it is possible to Max rF2 with this card and still enjoy good racing
Frames: 68673 - Time: 1000000ms - Avg: 68.673 - Min: 46 - Max: 109
Tearing is not really visible at least I couldn't spot it but FPS fluctuation is noticible, sometimes event feels like stutter.
Also just found out I accidentally 'lied' in previous post, my settings weren't all Max, must have missed that, opponent detail was Medium, just found out I can't load Silverstone with all settings far right and full grid of Honda cars, it gives me Memory allocation Error.
Anyway it is possible to Max rF2 with this card and still enjoy good racing
DrR1pper
Registered
I have a gtx 770 and I can run on max but I don't. I always turn environment reflections off (because you still have them but updated reflection maps every 10 seconds instead which saves a good 40%-ish in fps) and road reflection on high (because you then still get the nice road reflections and road shimmer and wet road effects only sacrificing 10-20% fps).
I found these to be good compromises that give a near max setting appearance without as significant a performance drop. Very handy when you don't use vsync and need to keep the fps high to make the game stutter less from sudden fps drops. Perhaps try regular vsync and see if the lag is significantly less (i.e. acceptable) than vsync video and if it also removes the stutter moments.
Here are my settings for my gtx 770:
I think for AMD users the major culprits are in te environment and/or road reflections settings. I recall someone saying that disabling one or both of them made their AMD card perform as it should vs its nvidia equivalent competitor in other games. Worth testing I think. I know how annoying stutter is and if even regular vsync causes it (due to the dips below 60) then perhaps setting the graphics settings so it never dips below 60 is what you need to do to fix it (if your willing to compromise...I did for a time before I bought a better card).
I found these to be good compromises that give a near max setting appearance without as significant a performance drop. Very handy when you don't use vsync and need to keep the fps high to make the game stutter less from sudden fps drops. Perhaps try regular vsync and see if the lag is significantly less (i.e. acceptable) than vsync video and if it also removes the stutter moments.
Here are my settings for my gtx 770:
I think for AMD users the major culprits are in te environment and/or road reflections settings. I recall someone saying that disabling one or both of them made their AMD card perform as it should vs its nvidia equivalent competitor in other games. Worth testing I think. I know how annoying stutter is and if even regular vsync causes it (due to the dips below 60) then perhaps setting the graphics settings so it never dips below 60 is what you need to do to fix it (if your willing to compromise...I did for a time before I bought a better card).
Last edited by a moderator: