AMD FX-8350 @ 4Ghz (8 cores), ATI Saphire R9 280x Tri-X 3Gb OC, 8Gb DDR3 RAM 1600Mhz CL9, Asus M5A99FX Pro R2.0 (Latest Bios Version) --Windows 7 64bit-- Mirrors OFF --> Avg: 41.282 - Min: 21 - Max: 80 Mirrors ON --> Avg: 31.194 - Min: 15 - Max: 53 Next test windows 8.1 64bit
Yes, i have run the benchmark again and the 12.11+CAP2 is still better for my 7970. The 12.11 has higher values in the minFPS (more important) and the 14.1 has some higher values in the maxFPS. Anyway is the proper direction from AMD with the 14.1, the 13.x were all junk. These are the results with the 12.11 beta + CAP2 and in the clamp (red) the values for the 14.1 driver My System: HD 7970 3GB , i5-3570k@4,3GHz 1. 1920x1080 Single Screen Min= 40 (29) Max= 65 (88 ) Avg= 48 (52) 2. 5908x1080 Triple Screen / Multiview on Min= 21 (15) Max= 30 (40) Avg= 24 (26)
You should already removed in rFactor 2 for the AMD card: FXAA - OFF Antialiasing - OF and configure the catalyzed Post in a long say about it: http://isiforums.net/f/showthread.php/18087-Experimenting-with-AMD-graphics-card-settings/page2#24 Your tests will be different
I have been a long time and user, mostly because of the eyefinity. I have complained here about awful performance in rf2. Not just RF2, nvidia seems to perform better with a number of sims. Yesterday I took the plunge and removed my AMD card and put in my first nvidia card for a number of years. I didn't get the visual difference I thought I might get but I certainly got a large framerate improvement. I did go for a much more powerful card than I had but the gains I have got surpasses the gains I should have received based on only the card difference.
I did the Renault Clio and Loch Drummond Benchmark from B342, but I did it using B660 and all up to date content. Some things to know: 1) I used a brand new, untouched controller.ini and player plr file 2) I mapped all the buttons to my wheels and controls to the way I normally play, but I left FFB smoothing to the default 4 and same with head movement and exaggerate yaw; I left them at 100% and 25%. 3) The original settings for the Benchmark say to enable "Sun Occlusion" however that option has been replaced by "Soft Particles" which I set to max, however even with "soft particles" set to max I noticed in the PLR that the line "Sun Occlusion" is set to "0", so it still seems to be disabled. I left "sun occlusion" in the PLR file at "0" in order to keep the PLR file 100% un-modified. 4) Under replay options I changed "replay fidelity" from "medium" to "maximum" 5) I use EVGA Precision-X in order to enable K-Boost which limits most of the NVIDIA dynamic GPU voltage and Core MHz adjustments. This keeps things very consistent compared to stock, but it's still not 100% perfect. My System: CPU: Intel i7-4930K @ 4.4 GHz, Hyperthreading disabled GPU: NVIDIA GTX 780 Ti @ 1190-1205 MHz core / 7000 MHz memory. GPU Driver Version: 337.81 Beta RAM: G.SKILL RipjawsZ 16 GB (4 x 4 GB quad channel) @ 1866 MHz 8-9-9-24 2T Operating System: Windows 8.1 64-bit My Results: Avg: 112 - Min: 93 - Max: 136 If anyone would like me to do the same benchmark with NVIDIA 3D Vision enabled, in triple screen mode w/ Multi-View enabled, or both, let me know. I also just did the new benchmark (Civic @ Silverstone, w/slightly different settings). Results are here --> http://isiforums.net/f/showthread.php/19682-NEW-Benchmark-comparison-between-AMD-Nvidia?p=275613#post275613
Sun Occulusion has not been "replaced" by Soft Particles as those are two completly different effects. As you noticed later, sun occulusion is still there, in PLR file but might not work as designed (reason why ISI took it off from UI).
So can we conclude that amd and nvidia has now the same performance? I bought a gtx 770 2gb just because amd had poor performance in rfactor 2 and I want to return it and buy a r9 290.
for me, I was going to switch back to Nvidia, after dealing with my old card long enough with this sim. Like a month before I bought the nvidia card, AMD released their 14.4 drivers, and I was pretty stoked about the performance.. So I bought the 290x, mostly because of the heat issues with the base 290/ I bought a windforce, and my temps stay at 40c while playing anything. It idles at 25c... Makes no noise to speak of.. And I bought a full tower case to fit the monster in as there's no way the card will fit in a mid-tower setup.. lol.. Awesome card, really is. Very happy with it, only had it two weeks now though
If it helps, I can overclock the 290X to 1215 MHz Core Clock (from stock 1000 Mhz) and 1550 Mhz memory (from 1250 stock). It gives some more fps, some 5-15 extra if your cpu can take it. I use MSI Afterburner, set the Core Voltage to +100 mV, Power Limit to +50, and I personally don't use the Fan Speed since it's watercooled. If anybody does this, make sure you don't cook the card, set the fan to 100%, but I'll sound like a train. I should say, I run a I7 3770K @4.5 Ghz, and triple screen (5760*1080) and can't turn everything on. If I do, I get between 5 and 30 fps. With "fancy" things off, no reflections, no wind movement, no fancy this and that and 4 AA, no FXAA, with HDR it can stay around 55-65 fps. No multiview, with multiview on it kills it and drops to around 30 fps at most.
I use the 290 in a relative old base and yet the performance is surprisingly good. I'm sure AMD has not yet achieved the optimum but they are definitely recommended now, especially the performance in other applications, such as BF4 with Mantle or some other software. I for one am very pleased and I have seen many graphics cards but this is one of the best. What do you mean with you can't max out, what conditions ? Apart from FXAA and Multiview in use i can set everything else on max on triples and still have an avarage of ~60 fps (short run ~3 turns) with 21 AI on Silverstone as example, and this with a Phenom II 955 sys which even doesn't support PCI-E 3.0.
