I am thinking of going SLI I have a i7 2600k P8P67 mb, therefore pci2 with a single 970. Should I go for one 980ti and run three 1080p monitors from that or go the SLI route with two 970's. Although the 970 runs three monitor well, with reasonable setting's. it is noticeable that under quite a few conditions it struggles to maintain 60 FPS even if I try lesser settings. When it drops to 30 odd FPS it make driving difficult. Also from reading the forum I know there are some issues with scaling SLI, what ever that means.. Could some one give me some advice on my next move before I splash the cash. Or has a bench mark been performed for these circumstances . I have looked
Personally I would go with one 980: - less power consumption - less heat generation - no SLI problems like micro stutter or wrong/missing SLI profiles
970 SLI has much better peformance and more fps then single 980 or 980 ti in all games also rFactor2 Nvidia support or not. Just remeber that SLI is not allways just klick and play and sometimes you must be ble to setup to you spesific hardware.
I've a 980ti -All graphical settings to full -Downsampling from 4k and level 3 AA (super clean image) -50 vehicles, but only rendering 35 -Framelimited to 70fps -No vsync It rarely drops under 70fps and only hovers in the 60s when on the grid.
I'd pick single 980ti to reduce frustrations. Same power requirements as single 970, good enough for triples with multiview, no fiddling with SLI. Hah, I see Ozzy already said the same thing!
If anyone here has triple 1080p screens and a 980 Ti, we could help Kek700 and others out by doing a benchmark because I have 2x 970s in SLI and triple 1080p screens...
Now that would be interesting especially for me I much prefer the simplicity of one card as opposed to SLI. But also it would be good to know I can future proof by SLI'ing with an added 980 ti when the price point drops in the future.
That would indeed be very interesting to know. It could be that a 970 sli is better for rf2 over a single 980 Ti. However, from my brief search of some performance comparisons online, it looks to me like the 980 ti is consistently the better option in all other games, especially considering the price. If you play other games as well, it may well be better to go the 980 Ti route. If however the majority of your gamings interests lie with rf2 and the performance results come back in favour of the 970 sli option, then ofc probs best to go the 970 sli route. But based on the performance in all other games...doesn't look likely it'll be in favour of 970 sli.
GTX970 SLI at 8x 8x ( Z170) would be best bang for buck. GTX980Ti only needs a H170 motherboard, with stock Intel i5 does not bottleneck a GTX980Ti. --- Does 8x 8x vs 16x 16x have as much effect as much as Pcie2.0 vs Pcie3.0 for RFactor2 would be my only question. lol
Ah, s***. How did i forget about that giant pitfall. That alone is sufficient excuse not to go the sli route unless you have the mobo+cpu combo to negate its heavy affects in rf2. I believe so. Look to my most recent bench results: http://isiforums.net/f/showthread.p...her-end-cards!?p=382841&viewfull=1#post382841 In rfactor 2 with my 970 with a single 1080p display, PCI-e 3.0x16 vs 2.0x16/3.0x8 = 22% average fps increase. The performance difference should also be the same if going SLI. So if he intends on running each 970 on pci-e 3.0x8/2.0x16 mode only capable hardware, for sure it will perform worse than a single 980 Ti running on full 3.0 x16. Ok, so i just realised he's using a i7 2600k P8P67 mb meaning he's limited to a maximum of pci-e 2.0 x8/x8 in sli mode. That'll equate to only 56% the performance of 970's running rf2 on the latest hardware allowing for pci-e 3.0 x16/x16 mode with a single 1080p display. 2x 970's in pci-e 2.0 x8/x8 mode would probably get no more than 20-30% higher average fps (if even that) than a single 970 in pci-e 2.0 x16 mode. Therefore it makes zero sense to go sli with his current hardware for rf2. Having said that, performance drop seems to decrease when you increase the pixel count. For example, at 2560x1440 the performance drop from pci-e 3.0 x16 to pci-e 2.0 x8 goes down to only 63% compared with the larger 56% at 1920x1080. (Just did some extrapolating and triple 1080p at pci-e 2.0 x8 should be around 68% the performance otherwise achievable in pci-e 3.0 x16 mode. A little less of a drop in performance potential but still a signficant loss none the less.) Lastly, a GTX 980 Ti in your current PC that is limited to pci-e 2.0 x16 for a single graphics card is going to hinder the cards performance potential in rf2 a fair bit as well. If i had to guess, you'd probably be looking at around 75-80% of what it would otherwise be capable of if ran the 980Ti on a pci-e 3.0 x16 capable mb+cpu at triple 1080p. If you are going to keep your current mb+cpu combo, fore sure 980Ti is the better option but knowing it's till going to be limited to only around 70-75% of what it would otherwise be capable of with newer mobo+cpu hardware, the choice to run a 980Ti instead of 2x 970's on your hardware feels more like a choice of picking the lesser of two evils. If i were you, i would consider also upgrading my cpu and motherboard. You would't need to splurge out for the very best either to make the 980Ti run at 100% unlike with 2x 970's that require pci-e 3.0 x16/x16 capable cpu+mobo combination (which costs a LOT) in order to get 100% of their performance out of them in rf2. In 99.9% of games, pci-e 2.0x16 vs 3.0 x16 doesn't affect frame rate performance in any way. rFactor 2 is unfortunately for whatever reason(s) a big exception to the rule. (sorry, finished with my edits now).
I thought as much if you can afford 40 lanes it is very nice to have. Very jealous right now P.S. which sort of belies doing test like this https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_970_SLI/20.html ( thats average of 20 games ) 16 lane Processor used or is that only making it fair for single cards or don't new gen games get a good boost like rF2 ? I mean I know they don't with pcie3.0... so is relatively the same for 8X vs 16x ? Sorry not 40% that is 80% faster / scaling ( 61% - 100% ) ? thats okay.
By the way R1pper I left out that line out of post as I did not want to take chance to upset Spin someway thinking I think 16x 16x is not good or a waste, which if you can afford it is obviously NOT, especially if you are you trying to get Hi-Res/ Multi screen synced to 120/144MHz nice. Us mere mortals on 1920 @60Hz forget this sometimes. hehehe
fyi, i made some edits along the way but finished now. To summarise my point in my previous post, I think the GTX 980 Ti is the better option with or without an upgrade to your current cpu+mobo hardware. Having said that, if you can also upgrade your mobo+cpu to something that'll allow a single graphics card to run in pci-e 3.0 x16 mode, you'll get an extra 33-43% extra average fps in rf2. That's a lot. Fyi, i upgraded my i5-2500k to an i5-4690k a while back for this very reason. So my cards could run in pci-e 3.0 x16 mode for rf2. "Bitch...please. I'm rolling 1440p @144hz" lol. sorry. couldn't resist.
well yeah lol OT really happy with GTX970 @1920 Boxer and I raced 30 ai and Belgium ( 120% AI 4x tyres/Fuel ) 70 fps at start smooth as silk. gr fun too
Yeah, it's a good value for money card. Having said that though, i'm suprised how good the bang for buck for the 980Ti appears to be. I've traditionally always gone for the GTX -70 cards since they usually have the best bang for buck. However the 980Ti looks to have an equally good bang for buck which is odd. Usually expect to see the top Ti cards having almost twice the price to performance ratio of the GTX -70 cards of the same gen. Unless i'm missing something?
I think I've made a mistake actually regarding my statement that the 980Ti performs twice as well as a 970. Need to look further later.
