The relative performance between equal Nvidia and AMD cards in rF2

Does anyone know the current relative performance between equal Nvidia and AMD cards in rFactor 2?

So an example would be the GTX 460 vs the HD 6850 achieve the same average results (in fact almost identical) across all games. But how does that correlation stack up against rf2 in its current state and for other cards?

I know it's unlikely that some here will own two different cards of equal performance but I'm interested as I am weighing up whether to switch back to AMD.

 

All I know is the new nvidia cars has a lot of tweeks and their drivers are much more optimized. even though amd cards gives you more fps, nvidia cars gives less stuttering. I will switch to nvidia as long as possible.

 

You can use hwcompare.com to find their theoretical performance based on their specification.

 

I am curious about this too. I have traditionally been an Nvidia guy (not a fanboy though) but will jump on what ever card is the best bang for the buck. The only games I play are rF1 and rF2. rF1 is maxed out on any current card (hell even Intel Integrated) but rF2 is a whole new ballgame. I currently have an HD5770 and I finally managed to get decent performance out of rF2 with it but would like to upgrade. It seams like traditionally ISI has been bias towards Nvidia but. If I remember correctly they even had an Nvidia intro logo on rF1 at one point. Maybe that was SCGT.

I would love to know which route to go next. AMD or Nvidia.... hmmm.

 

I too am not a fanboi, and will pick any good price/performance card so long as it isn't an AMD/ATi due to their bad Linux driver support, or an S3 because they smell.

Currently, I'm running an NV660-TI which was pretty sweet for the price when I got it.

 

I purchased a HD 7970 last night, only to experience buyers remorse this morning after i discovered that i had not checked the GTX 670 which is only 6-7% lower performance than a GTX 680 and costs a heck of a lot less at £238 vs £360 respectively (51% cheaper!!!). The HD 7970 is as good as the GTX680 and better in games like BF3 and i found one for £283 but i feel more comfortable staying with Nvidia having had so much success with them for so long.

I'm really torn between price to performance ratio and brand success.

Thanks to everyone who posted.

 

Where are you getting these prices from DrRipper, I recently purchased an Asus gtx670 recently for ~£330. I see they're available for £315 now on Amazon. Made the switch from Ati to Nvidia due to purported more stable performance and slightly better power draw, although there is a price premium to pay for nvidia.

 

As I am no GPU guru I am trying to wait out the purchase of a new card until the game is further along and I can read feedback from other users about which card they feel does the best job handling rf2. It hasn't been until recently that I've started to encounter issues with some stuttering, some screen tearing and lower fps at some tracks. I imagine it could be any number of things but it's beating up my 6950 2gb pretty bad. (I run triples).
Being the novice that I am it makes my head spin when I read an article about issues with using 2 cards and another taunting their magnificence. It seems like a lot of research is necessary to make the right choice and even then you start to worry you might be missing something important that will effect your results.
Anyone else ever feel it's a bit overwhelming to keep up with the hardware side when all you really want to do is race? lol I suppose gamers must be tweakers.

 

I have used a HD 7970, HD 7950 and a GTX 680 in rFactor 2 and their is no difference in performance. All 3 are more than powerful enough to run rFactor 2 at 50+ FPS at 2560x1600 for me. The problem is that rFactor 2 does not seem to like AMD when it comes to FSAA. No matter what settings I use rFactor 2 has a lot of shimmer with distant textures on an AMD card. So if you want the best from rFactor 2 stick with Nvidia. Unfortunately that means paying a higher premium for less performance.

HD 7970 is around 10% faster than a GTX 680 and is cheaper.
Even a decent overclock on a 7950 will beat a decent overclocked GTX 680 for a lot less money.

Here is a link to another forum where I show graphs and benchmark results from an overclocked 7950 compared to an overclocked GTX 680.

http://hardforum.com/showpost.php?p=1039505601&postcount=66

Right now a GTX 680 is overpriced and slower when compared to the competition. Though as I said this doesn't help when AMD doesn't work well with rFactor2. I wouldn't blame AMD on that though, clearly the devs need to work on their graphics code. It is a beta after all.

If you do decide to go with a Nvidia GPU don't go any lower than a GTX 670. Anything lower is horribly constrained with low bandwidth IMHO.

PS. The GTX 670 is NOT only 6%-7% slower than a GTX 680 if we look at more recent GPU intensive games. Stock for stock a GTX 670 is around 10%-20% slower overall. Things get even worse if you compare against a similar priced 7970. It then ends up around 15%-30% slower for the same price. So DrR1pper you should be happy that you bought a GPU that is undoubtedly faster than a GTX 670.

 

51% Cheaper ?



My store same brand

Cheapest decent vanilla 670 - 389
Cheapest decent vanilla 680 - 519

That is 25% cheaper


Gigabyte 670 OC 4GB - 499
Gigabyte 680 OC 4GB - 649

That is 23% cheaper

 

http://www.waeplus.co.uk/product/1981538/PNY-GeForce-GTX-670---2-GB-GDDR5---PCI-E



I've scowered the internet in search of updated benchmark comparisons as there have been improvements from drivers from both sides of the fence.

Here is one i used for example: http://kotaku.com/5964022/the-best-graphics-cards-nvidia-vs-amd-current+gen-comparison

And example benchmark comparison from that review:



Yes the GTX 670 actually has 10% lower fps than a GTX 680 in this example but it's 51% cheaper to the cheapest GTX 680 i could find (@ £360).

 

IT IS NOT 51%

Do your calculations again.

It is 34%

All you do is compare cheap 670 to a dearer 680 but it is no way fair.



Anyways it is around 25% less is what you should be paying for a 670 ( equal model cards)



BTW
If you compare the cheapest 670 to the dearest 680 then it is 100% DEARER
I can get a 670 for under 400 the dearest EVGA 680 4GB classified is 799 .......

 

Sorry, the GTX 680 is 51% more expensive than the GTX 670.

So if i were to purchase a GTX 680 it would cost me 51% more for only a 10% increase in performance.

That's not true, the £360 is the cheapest price i can find for a GTX 680.

 

You can easily overclock the 7970 to 1200MHz(mine is watercooled but even with stock fan you should get there.... With a bit higher temps... Depends on your system temp.)
Its a great card.... Rf2 is getting more and more amd friendly with every release.... Dont worry, for me it runs fine with tripple screen and nearly all maxed out.... To be honest its a bit like the pepsi coke question... At the you get what you pay, if its nvidia or amd.

 

The cheapest 670 I can see at OC is £300, with its like-branded 680 counterpart costing £480 (the one with twice the memory is £60 cheaper) - whatever the % cheaperness(*), the This Week Only £30 off the 670 makes it look pretty sweet right now.

But yeah, £360 for the more budget end of the 680s.

(*) That's at least two goats.

 

The 79xx are generally underclocked/have more room for overclocking vs the nvidia 6xx series, and although it's game dependent a 7950 with a good clock is up there with a gtx 680 sometimes, a 7970 neck and neck and game dependent with a 680. Once overclocked, I believe the 7970 has slightly better potential than the 680, again though it can be very game dependent at times.

You're right about them being ripoffs though. I would never pay the extra money for a 7970/680, when you can get so close, for so much cheaper with a 7950/670, then overclock it and get another 10-20% frames.

 

My store same brand

Cheapest decent vanilla 670 - 389
Cheapest decent vanilla 680 - 519

That is 25% cheaper


Gigabyte 670 OC 4GB - 499
Gigabyte 680 OC 4GB - 649

That is 23% cheaper


Check yourself

http://www.pccasegear.com/index.php?main_page=index&cPath=193_1377&vk_sort=1

http://www.pccasegear.com/index.php?main_page=index&cPath=193_1387&vk_sort=1

My figures above are CORRECT.

So I can get from 25% to 100% difference depending on models

All I was saying is 25-33 % for equal specs ( ram, cooling , brand ) is what you should pay.

51% is way too much ( Sorry I fudged the figures it is 51% )

You can see my store comparative cards is more like 25%.

 

If you got the right motherboard and power supply spend 250 bucks and get 2 more used 6950s and trifire. Trifire generally scales very well, it also gets rid of almost all stuttering that may be brought on when using only 2 cards. Or, sell the 6950 for 140 to 180, and get a used 7950 for 240-270. The 7950 will be about the performance of 2 6950s, can't remember if this is with an OC or not, but OC it anyways.

Don't ever use that site, it's all theoretical bull**** and means nothing. Look at proper reviews and benchmarks from gaming sites, forums, YouTube, etc

With supersampling?? Hmmmmm, I highly doubt that. My 2 6970s in crossfire can struggle at times, let alone any current card or especially an integrated gpu.

 

.

 

Above 1920x1080, the AMD cards start to really outshine the Nvidia cards in the varies game benchmarks. Comparing an overclocked 7970 vs a stock 680 isn't really helpful since you should then be comparing it to an overclocked 680 and the benchmarks say the same jump in performance so they are still tied. That is not to say the 7970 is a ridiculously cheap alternative to a GTX 680 (at least for where i live) as the GTX 680 costs 27% more than the HD 7970 for the same performance at 1920x1080 and the 7970 gets better performance when you start to go above 1080p and with more Anti-Aliasing level.

That alone is a great reason to purchase a HD 7970 but i only wanted to find something that could replace my GTX 460's in SLI with a single card solution and slightly better fps and at a great price/performance ratio. The jump from GTX 670 to GTX 680 (or HD 7970 @ 1920x1080) is a 10% performance gain but at a steep 51% extra cost (or at a 19% extra cost for the HD 7970).

The simple fact is that i am very lucky to have found this single bargin for the GTX 670 that gave it a decently better price/performance ratio than the HD 7970. If this single bargin didn't exist, then the HD 7970 would not only perform 10% better than the GTX 670 but also have a better price/performance ratio.

Although the benchmarks did not include GTX 460's in the comparisons, it did however include a GTX 580. I know from other reviews that a GTX 580 performs nearly identically to a pair of GTX 460's which allows me to guage the expected performance increase. Anywhere between 18-32% across all the games on that benchmark site with the average hovering around 24%. This is not exactly a giant leap but is welcome non the less and i look forward to seeing if swapping to a single card solution will alleviate the stuttering i experience when graphical settings in rf2 produce lower fps. And by lower i mean below 140fps (stutter starts) and at 60fps it becomes unplayable without using Vsync: Video which suddenly makes the game become unbelievable silky smooth (which is still really puzzling me). I'm hoping that a single card might provide me me with the silky smooth performance without any vsync at around 60fps so i can re-crank all the graphical qualities back up to how they were when i was using Vsync: Video with my two cards.