As I moved to GTX 780 Ti, here are my scores (sort of a confirmation, as mine fit between other 780 TIs from the list). CPU the same: i5 2500k @ 4,5GHz drivers: 331.65 Mine is Gigabyte Windforce 3x OC, so it's factory overclocked - boost goes up to 1150MHz (although on paper, it's core 1020 and boost 1085MHz). 1920x1080 Avg: 98.582 - Min: 79 - Max: 133 5760x1080 (MultiView off) Avg: 63.045 - Min: 47 - Max: 78 5760x1080 (MultiView on) Avg: 47.861 - Min: 37 - Max: 59
I'm not sure the test is 100% on equal terms anyway, since the benchmark uses AA on level 3. On my GTX 760 that is level 3 of 8 but on my old HD 6870 there were only 3 levels anyway (so lvl 3 should be a much higher setting there). When I got my gtx 760 I did a comparison with the same combo as with the older card (Marussia @ Silverstone with mostly med/high-settings) and my fps with the same settings went up by about the expected amount (~80%). But I always used AA lvl1 on the HD 6870 since everything above that massively crushed the fps with this card.
^^ correct. level 3 in RF2, is NV 8xCSAA. AMD drivers do not support CSAA. Not even in software mode. My info is from here: http://isiforums.net/f/showthread.p...ow-They-Relate-to-Video-Card-Driver-AA-Levels
Yes, and not every level looks and performs good. As example the level 1 looks better than level 2 which blurrs the image to much. Level 3 is even more worse and level 4 works again, as the next step working is level 8 for me, but i don't like higher levels as they blurr the image to much with very little efficiency on the jaggies. Nothing else than level 1 or Fxxa is worth to use here, because the result of different levels isn't worth wasting more performance atm. As it seems some things changed with this build and my config doesn't work anymore. I've struggled yesterday to find the balance again. However there are things i like more and things i don't but in general this build is an improvement for me. The visuals are much better with the same power requirement, while it still isn't possible to see the full extend of te visuals with my machine, the gfx is becoming better. Finally i end up with the following settings , since everything else cause any issues, or don't performs well. I can go with High Env Reflections and still have FPS above 60 but i don't need it ( wasting power atm). Here the Result of the bench replay with RadeonPro/rf2 optimized settings and the resolution of 4900x900 ( multiview off ) : Frames, Time (ms), Min, Max, Avg 16551, 201553, 43, 115, 82.117 View attachment 11195 View attachment 11196 I rebalanced clock speeds of the cpu after benching it with max clock settings and check the avarage level of FPS, which decrease about 5-6 fps with max clock while the min FPS increases about 4-5 fps. The Result of a more constant and higher avarage fps level was a clock speed of 3600Mhz instead of 4000Mhz before. As well intresting is the increased fps rate with setting the Flip Queue to 0 in RadeonPro. this decreases lag as well. Another intresting fact is, the 290 runs better without cap fps and any sync option. The card throttle's to much with cap and sync options in use. Single car performance with this settings are brutal, starting with ~150 fps in the box of LochD and reach levels above 170 fps on track. With setting the Transparency AA to 0 in the plr gains another 15-20 fps on track for me but i keept it in use for now. From here each optimization should only deliver a better result. I have hopes
This thread is a bit of a bummer to read. I haven't been racing in any form for a little while but my racing machine has a 6970 and I was thinking of upgrading my main machines GPU and giving the racing machine the 7970 as a hand me down. I thought it should have been a reasonable upgrade but it seems not really.
So which is better a 770 with 4GB ram http://www.memoryexpress.com/Products/MX47143 Or a 780 with 3 GB of ram? http://www.memoryexpress.com/Products/MX46083
Ok, thanks, got the part about XP, but why is the 780 the better choice with respect to rfactor2? I thought rfactor2 liked more video ram?
The benchmark should have a standard setting of in-game AA level of 0 as the in-game level AA is a different amount of AA depending on the GPU/driver. It should use a standard, equivalent for all GPUs, control panel set AA for all tests.
Yes the benchmark is not equal. Another reason is texture quality differs between the ATi and Nvidia, you have to run the Nvidia in quality or high quality for the textures to look the same as the ATi so this is another area where the ATi figures look worse than they actually are, but even when you level these things out the Nvidias are still crushing the ATi in this title, even when you bring price into the equation. I have both the 7970 and the gtx770 (4gb) and there is no way that ATi is going back in my machine, running everything on low with the lowest res possible across 3 screens worked out about the same as the nvidia with everything set on high and max resolution, so no comparison with equivalent cards. All other games ran superb on the 7970, as the do on the Nvidia. I think to see the effects of video ram 3gb v 4gb or 780 against GTX770 4gb we need a different benchmark with a higher resolution . I would still love to see a SLi V no SLi comparison and the FCAT test performed. How does it scale? Is it worth it? Does it stutter? Also be interesting to see what you can get with a pair of 290x in xfire now they need no bridge. Since the new build I think things have changed as well probably making the original benchmark a bit more invalid for instance I now find the in game AA settings to work just as good if not a little better than the NVidia ones, I am just going by feel I haven't benchmarked it. Possibly the true type fonts thing (in the 382 release notes)?
Need to create a new topic for tests. I noticed an increase of minimum and average fps, +8. But may be this is due to the updated track?
It is more acceptable for now but requires further optimization imho. The options are a real mess anyway and the odd some low or medium settings perform more bad than higher levels with my pc and 'm forced to use high levels of something i may don't like or need, as example the special effects on low or off cause stutter and get reduced just by increasing to greater levels as low reflections perform bad here also and need to be set off or high while the low level looks terrible and cause stutter as well. The opposite it is with the soft particles which i need set to low because here the high level cause stutter and reduces performance extremly. The Track Details are another performance killer on AMD's as it looks like. 3 Turns on LochDL, ridden with tha Corvette against 20 AsideIntelligence driver going by a resolution of 4900x900 and AF 8x/ AA Level Off / FXXA On / Soft Particle Low / Shadow Blur Optimal / Track Details at Low : Frames, Time (ms), Min, Max, Avg 12914, 256341, 30, 71, 50.378 View attachment 11224 View attachment 11225 It is well driveable but not enough for my like and thereby I am unfortunately forced to put some back here and there, and what i don't need for driving is something like high QA reflections but slowmo, stuttering and awful low resolution ones as well not, so i set them off. About road reflections i'm not sure, if i ever see any high glossy reflection in real on wet roads, so i'm not sure if i should waste performance with that and set it also off. Just setting the reflections both off catapults the fps rate up to a average of ~80.
Hi, here are my results. My System: HD 7970 3GB , i5-3570k@4,3GHz . Catalyst 12.11Beta with 12.11CAP2 installed 1. 1920x1080 Single Screen Min= 40 Max= 65 Avg= 48 2. 5908x1080 Triple Screen / Multiview on Min= 21 Max= 30 Avg= 24 Which Settings do you have in Catalyst?.
Actually standard settings here, without the use of RadeonPro with my last test results to not confuse people. Don't know about other's but i have had used the HQ texture filter setting with my previously bench's but used CCC standard settings with the last one. I'm not sure about that but in the past, year's ago i heard something about the efficiency of that option, where it was topic NV would cheat about the texture filter option, while ex ATI's standard was NV's HQ setting but i'm not sure if it is actually still a topic. However, the last test's are with standard CCC settings.
Because of poor Performance with the new drivers. The best driver for me ist the 12.11 beta + the Cap2 Profiles. With this driver i have a increase of 40-50%. Some others have the same result: http://isiforums.net/f/showthread.php/16951-R9-290-or-GTX-770-4GB?p=223501&viewfull=1#post223501 and here http://isiforums.net/f/showthread.php/16951-R9-290-or-GTX-770-4GB?p=223544&viewfull=1#post223544 . I`m thinkink about to buy me a Nvidia Card, AMD-Cards sucks in Racing-Games.
Actully it don't look's that bad anymore, if we consider on what kind of base the 290 gpu is accelerated in my case. I'm using still a 890 FX chip with a Phenom 955BE on 3600MHZ and a standard of PCI-E 2.0. It is not really the standard for the current gen GPU's but unfortunately rf2 is still the most bad performing current gen game on it, while it can't really competed ( similar conditions ) against anything out there, it still leaves a bit of bad feelings on me but i'm also aware of the release date of a new gpu gen, which for sure isn't on it's max optimization level. I'm sure it would be somewhat better already still with a more current gen Base.
