Do you mean the nVidia drivers are more stable now on Metro: Last Light? Last time I heard, and according to the reviews, Metro: Last Light runs better on AMD's cards, as you can see by yourself on the link below. http://www.techpowerup.com/reviews/AMD/R9_290/17.html
That i call customer service. Ordered yesterday some min before i post here and now i'm holding my new 290 in my hands.
M8 just an idea. Create a race with 3 laps and 20 cars. It can be done only with AI's. Then save the replay of the race to share with us. If possible use some mod and track provided by ISI to be easier for us to test it too. Then for example, start the replay in cockpit view of the last car on grid at the start and configure fraps to save the min/average/max FPS during that 3 laps and always in cockpit view. that way all of us can compare the results with other car. Here i can test one 5870 1gb and one GTX660 ti PE 2GB. That way we could create some results to compare and help people like me too chosing the next graphic card. Just one last note, you must share the entire settings of rfactor and if possible test resolutions like 1680x1050, 1920x1080 and 3 monitors if you have three. Thanks and waiting for results Note: When i get this 660 ti i test it against the old ATI card with one lap alone in Portugal with DTM car in cockpit view. All maxed out with 3xAA and 16xAF at 1680x1050 and i get 42fps average on 5870 and 82fps with 660 TI In overall the 660 is about 50/55% faster than 5870 but in rfactor 2 this number becomes almost 100% faster... At 5040x1050 Disabling reflections and putting shadows on low i can have 20 cars in Portugal track with a average of 50fps on 660 TI
I will try to follow your advise as i'm not really the king when it comes to vids but at least some replay capturing should be possible. I'm a bit nervous since i couldn't install it yet but in a hour or so. When i have something usefull to share, i'll let you guy's now
you don't need to record the vid . Just share the rfactor replay file for us to be able to open the same exact replay and the numbers that FRAPS save from minimum,average and maximum framerate in that 3 laps of race. only with the same exact replay file we can have a fair comparison .
But don`t use paid mod content like DTM, please! Not everyone has it. And use "ALL cars& tracks", not custom vmod, because replays will not open without vmod.
And he has to do the fraps benchmark of the replay in case running there is a performance difference with running live AI vs just running a replay. Also make sure to start at the back of the grid with the AI's so you have the highest amount of graphics load. Speed1, you can also just let the AI drive your car by pressing "I" at the start of the race and then after 1-3 laps (up to you how many, imo 1 lap is enough) quit the race. Then play the replay and run with the fraps benchmark tool (please remember only starting and stopping the benchmark at the start/finish line).
No, I mean Metro Last Light is a nV product, and the 290 runs better on it. So the R9 290 should be fine on rF2. Since it's optimized for nV as well. And no, I meant AMD had new drivers out. 13.11 beta 9.2 http://support.amd.com/en-us/kb-articles/Pages/latest-catalyst-windows-beta.aspx
It's kinda weird that it runs better on AMD since it's a Nvidia game, as you mentioned. However, you shouldn't guide yourself that way since (I think) rFactor 2 isn't optimized for any of the two, it just runs better on Nvidia. Normally the sponsored games have publicity for either AMD or Nvidia and this isn't the case. It would be interesting to see some feedback regarding this matter from any of the ISI's programmers.
I recorded a replay with 20 AI cars with Renault Clio, 2 laps @ Loch Drummond. Here you can find the replay and detailed information how I did the test if anyone wants to test their computer-> https://mega.co.nz/#!8csV3B4I!GkBzylN-aHpGuKN9F9yV73pmh2cgLJiNSAXedhrilE4 My results: CPU:i5-2500K @ 4.0GHz GPU:Gigabyte GTX 670 with 331.65 Frames: 11319 - Time: 200047ms - Avg: 56.582 - Min: 43 - Max: 76 Did 3 tests and the average results were almost indentical
I certainly wont be basing my buying decision on assumptions like this, but if your willing to gamble and take the risk for us all then we will be only to happy to read about your findings.
Just did the benchmark with the replay you provided and the results were: Resolution: 1680x1050 Drivers: 331.65 CPU: Intel Q6600 @ 3.6GHz GPU: MSI GTX660 Ti PE (Defaults) RAM: 8GB @ 800MHz
CPU:i5-2500K overlocked to 4.0GHz GPU: Gigabyte GTX 580 Driver version: 320.49 (later will try 331.65) 1920x1080 GPU clocks @ 772/2000 (stock) Avg: 50.438 - Min: 37 - Max: 61 -------------------- 1920x1080 GPU clocks @ 926/2150 Avg: 58.920 - Min: 43 - Max: 72 -------------------- 1680x1050 GPU clocks @ 926/2150 Avg: 61.491 - Min: 47 - Max: 75
Unfortunately, i have bad news, not only that MickeySoft's Win7 means to annoy me to the beginning of the installation with tonns of updates, the performance is poor and sobering in this combination but i'm not sure if the card is running as it should be because it is very silent. I'm a bit confused because of all the reviews of a very loud fan and such but i would call it silent. The configs of the last 4 year's i use where CF and the last hd6950's where louder in stock without watercooler than the 290 actually (max 47% fanspeed). It seems a bit of overrated to me. However, after installing the current beta driver and restarting the sys i get an error message, anything with the Mantle API win7 couldn't load up or find. Very annoying because the CCC won't load up, and i need to reinstall the driver by cleaning with sweeper and start the installation procedure again. Finally everything went fine and the first and very basic i did was to start a standard bench with 3dmark11. The result of it 14xxx gfx score / 4xxx physics score and general result 95xx Seems a bit low to me. The same machine with cf hd6950's general result where 8xxx It looks like the cpu (955BE4000Mhz) is crap, there is no need to bench something because rf2 runns poorly on it, even with single screen the perfomance is worse than with the two hd6950. edit: fast test with single screen and 1920x1080 : rf2 on max except fxaa - silverst./camaro/no AI/just standing on grid pos 1 the fps move up and down between 40-55 fps AC on max except fxaa is constant on 66 fps with BMWGT2 Both are to low to enjoy imho. Intresting is with AC the fan of the 290 really starts to annoy and with rf2 it is more silent.
speed1, thats sad to hear. May be there is something wrong with the card or installation/drivers... My test: 2013-11-10 16:30:18 - rFactor2 Frames: 4071 - Time: 199822ms - Avg: 20.373 - Min: 11 - Max: 40 1680x1050 HD6950 @ stock i5-750 @ 3.8Ghz So basically ATI is 2x times slower than nvidia with rF2? Thats quite a unique situation. I`m planning to upgrade to R9 280x in upcoming weeks, I`ll redo benchmark then. No way I`m going to nvidia just to run rF2, all other games and my own OpenGL apps run flawlessly with my 6950, zero problems with drivers or card. UPD. Actually that might be natural gap from 6950 to 670/580. We need more benchmarks with comparable cards. Tim, may be this benchmark posts can be transfered to new thread, where everybody can post their results? Would be nice.
Speed, definitly check the clockspeeds during gaming or benching and watch for throttling of the speeds. It could be that when you have a small case in a hot environment that she gets hotter then she should and never reaches proper clock speeds. I know the automatic fan speed profile now limits to 47% but is there absolutely no way to speed it up some more? Also check wether it actually runs at that speed all the time.
2013-11-10 15:34:17 - rFactor2 Frames: 13329 - Time: 200087ms - Avg: 66.616 - Min: 51 - Max: 90 3570K @ 4.4GHz and Gigabyte 670 GTX @ 1280MHz boost. 1920x1080.
@Timpie The fan i can adjust manually with gigabyte's OC tool what i did as well and you can't believe how much noise it is able to produce on 100%. For a moment i tought the pc would lounch to the moon.