RX 480 Crossfire or GTX 1070?

[LyricalRain018]

Hey guys i just wanted to clear a doubt. Should i get rx 480 crossfire over gtx 1070. Cause i'm getting a monitor with the respective adaptive sync. But g sync is 170 dollars more expensive than freesync. So i really want to get the 480s. But will they run fine? Any reply will be appreciated.

[FallenDark7]

You shoud go everytime for the best GPU you can buy, crossfire and sli, are not worth the pain.

[Mr.Humble]

Hi,

 

what resolution do you plan to get? Chances are you won't even need to Crossfire the RX480s. Personally I don't have any experience with a dual card setup, so afraid I can'T help you much with that.

 

Another thing to consider is, what is your timeline? Because Vega is coming in Q2 2017, so if you have the option it might be worth your while to wait until those cards hit the market.

[Morgan MLGman]

Crossfire 480's are faster, but it's still a dual-GPU setup so if you're looking to get something to just throw in your PC and expect it to work, get the 1070.

 

If you like to tinker a little, you'd get quite a bit more performance out of a dual-480 setup.

[AlTech]

Wait for VEGA.

[Maxxtraxx]

13 minutes ago, AluminiumTech said:

Wait for VEGA.

Agreed, even if you still get a 1070 the chances of a price drop due to Vega/1080Ti are still worth the wait(if your willing to wait).

[Princess Luna]

13 minutes ago, Maxxtraxx said:

Agreed, even if you still get a 1070 the chances of a price drop due to Vega/1080Ti are still worth the wait(if your willing to wait).

1080ti won't hit the markets anytime in 2017 though, VEGA is still a couple of months away.
The GTX 1070 already is a very capable card, if he wants it now it is the best alternative, unless he has a freesync monitor or something... though even I use nVidia on a freesync monitor using v-sync instead and feel no stutter at all so who knows.

[Gonio]

Since dual GPU setup can cause a less smooth experience despite more fps I would go for the single card solution always. And allthough Gsync/Freesync help on their complete refresh range, they have diminishing returns the higher the Hz screen it is. I asume you talk about a 100+ refresh rate. So it is more a question if you need it vs want it imo, allthough having it is still great.

 

Vega is currently sceduled for 2nd quarter, but expect june. Wish it was sooner, because it could cause prices of the better cards to drop. Now there is no reason to go lower.

[grimreeper132]

4 minutes ago, Princess Cadence said:

1080ti won't hit the markets anytime in 2017 though, VEGA is still a couple of months away.
The GTX 1070 already is a very capable card, if he wants it now it is the best alternative, unless he has a freesync monitor or something... though even I use nVidia on a freesync monitor using v-sync and feel no stutter at all so who knows.

I recon the 1080ti will come out this year, as the 980 ti only came out about 9 months after the 980, so it would seam reasonable

[Princess Luna]

1 minute ago, grimreeper132 said:

I recon the 1080ti will come out this year, as the 980 ti only came out about 9 months after the 980, so it would seam reasonable

Nah nVidia still holds the strongest high end cards, until AMD releases VEGA to compete with the 1070 and 1080, nVidia can just wait sitting, if I had to guess, provided that VEGA is worth it, the 1080ti will come out only in 2018 Q1

[i_build_nanosuits]

51 minutes ago, LyricalRain018 said:

Hey guys i just wanted to clear a doubt. Should i get rx 480 crossfire over gtx 1070. Cause i'm getting a monitor with the respective adaptive sync. But g sync is 170 dollars more expensive than freesync. So i really want to get the 480s. But will they run fine? Any reply will be appreciated.

GTX 1070 and Gsync monitor is 100% worth it.

Crossfire/SLI is not supported by all games, it's very finicky...it's prone to micro-stuttering, driver issues, artifacting etc. and even when supported the performance scalling is often very poor...in most games, overall, a single GTX 1070 will perform a lot better, will be smoother, consume a lot less energy, will be quieter, cooler etc. and Gsync is superior to freesync to begin with etc. i could go on all day on why the 1070 is a much better choice but i'll stop here.

 

8 minutes ago, Princess Cadence said:

Nah nVidia still holds the strongest high end cards, until AMD releases VEGA to compete with the 1070 and 1080, nVidia can just wait sitting, if I had to guess, provided that VEGA is worth it, the 1080ti will come out only in 2018 Q1

i think the 1080ti will come out a couple weeks before or a couple weeks after AMD has a product that can give a GTX 1080 a run for it's money...if vega can't do that and only slot in between the 1070 and 1080 there won't be a 1080ti imho...but why this discussion about 1080ti?! op is looking at a GTX 1070 which is a much better bang for the buck anyways

[grimreeper132]

5 minutes ago, Princess Cadence said:

Nah nVidia still holds the strongest high end cards, until AMD releases VEGA to compete with the 1070 and 1080, nVidia can just wait sitting, if I had to guess, provided that VEGA is worth it, the 1080ti will come out only in 2018 Q1

I yea that being said, if VEGA is a serious challenge to NVidia, and starts to pull the market away from them then they might bring the launch date forward. Also as AMD has recently shown, they are refining their drivers so they get the best out of the cards meaning that VEGA, will probably have good working drivers, as they have been using polarsis to make sure they work properly, and if the drivers aren't the best on launch they will be refined to the point where they are good.

[i_build_nanosuits]

51 minutes ago, Morgan MLGman said:

Crossfire 480's are faster, but it's still a dual-GPU setup so if you're looking to get something to just throw in your PC and expect it to work, get the 1070.

 

If you like to tinker a little, you'd get quite a bit more performance out of a dual-480 setup.

i kind of disagree mate...the GTX 1070 is roughly 50% more powerful than an RX 480...and there are many titles where crossfire will provide less than a 50% boost...many games ranging from 0% (many titles at 0%) all the way up to 60 or 65% but the chunk of the games i believe will scale anywhere from 30% to 40%...and that's when it works properly.

http://gpu.userbenchmark.com/Compare/Nvidia-GTX-1070-vs-AMD-RX-480/3609vs3634

[Morgan MLGman]

48 minutes ago, i_build_nanosuits said:

i kind of disagree mate...the GTX 1070 is roughly 50% more powerful than an RX 480...and there are many titles where crossfire will provide less than a 50% boost...many games ranging from 0% (many titles at 0%) all the way up to 60 or 65% but the chunk of the games i believe will scale anywhere from 30% to 40%...and that's when it works properly.

http://gpu.userbenchmark.com/Compare/Nvidia-GTX-1070-vs-AMD-RX-480/3609vs3634

Yeah, I know, I wonder though as it's known that two R9 290X's are faster than a 980Ti, an RX 480 performs very similar or better than the 290X while being more efficient thus I assumed that it's gonna outperform the 1070 as it's around the same in terms of performance as a 980Ti... 295x2 beats a Titan XM in most AAA titles if you take averages, but I don't fully understand this graph:

 

Why is a 1070 here 18% faster than a 980Ti? It's not right...

[ivan134]

5 minutes ago, Morgan MLGman said:

Yeah, I know, I wonder though as it's known that two R9 290X's are faster than a 980Ti, an RX 480 performs very similar or better than the 290X while being more efficient thus I assumed that it's gonna outperform the 1070 as it's around the same in terms of performance as a 980Ti... 295x2 beats a Titan XM in most AAA titles if you take averages, but I don't fully understand this graph:

 

Why is a 1070 here 18% faster than a 980Ti? It's not right...

Maybe stock 980 ti vs boosting 1070?

[i_build_nanosuits]

2 minutes ago, Morgan MLGman said:

Why is a 1070 here 18% faster than a 980Ti? It's not right...

i HATE techpowerup, simple as that...over the years they have posted litteraly DOZENS of charts that make absolutely no sense what so ever...i don't go there, EVER.

In many games RX 480 CFX will indeed pretty much as fast or even maybe a bit faster than a GTX 1070...but what about all these other games that did not receive multi-GPU optimization by the developers...and you know as well as i do that these games are many, especially these days where multiple GPU is not a common thing...i believe less than 5% of users in the steam surveys use more than 1 graphics processing unit...so why would game developers waster money, time and ressources to improve the gaming experience of their titles for like...5% of the people out there...sure, in some games it works fine and trowing in CFX/SLI support is simple, and those games will have support in drivers soon after the game launch...but some other game engines are completely not friendly with multi-GPU support.

[Morgan MLGman]

1 minute ago, ivan134 said:

Maybe stock 980 to vs boosting 1070?

That's most likely, though it flaws their entire testing methodology and it's not some shitty little site but techpowerup... I mean, I expect reliability from sites like that...

[Gonio]

An average number is a good way for an indication in general, but it is still a indication. Also % say not much since the actual fps gain can differ from game to game and settings used. GOing from 60 to 66 fps is a 110 % in this example, but not a day and night difference in practise.

[ivan134]

51 minutes ago, Morgan MLGman said:

That's most likely, though it flaws their entire testing methodology and it's not some shitty little site but techpowerup... I mean, I expect reliability from sites like that...

Doesn't boost 3.0 do that on its own?

[i_build_nanosuits]

15 minutes ago, Gonio said:

An average number is a good way for an indication in general, but it is still a indication. Also % say not much since the actual fps gain can differ from game to game and settings used. GOing from 60 to 66 fps is a 110 % in this example, but not a day and night difference in practise.

and even then...even that number...even it WAS consitent from game to game(it's not) but it's still an AVERAGE...but gaming experience can't be measured in AVERAGE FPS...you get to learn that over time...frametiming and framepacing, minimum FPS... 0.1% low...perceived FPS...these are much more important and much more accurate than an ''average FPS over time''...let alone an ''agregate of 15 games frames per second averages over time average'' this become so so far from the actual user gaming experience it's beyond me. it shoud NOT EVEN be considered...AT ALL.

Look, i can tell you this: with my own machine, depending on the game and how it's working and pushing out frames etc. i've seen 85FPS gaming being stuttery and slightly choppy and felt ''off timing''...and i've seen perfectly smooth 45FPS animation...it depends on so much more than ''average FPS''...and when it comes to consistency, lack of micro-stuttering and complete smoothness of animation and enjoyeable gaming experience, a single more powerful GPU will ALWAYS be king...always shoot for the strongest GPU you can afford, unless you're buying a top-end card, and you think you want/need more and you swim in money, then buy two of these and have a hell of an e-peen...but at the end of the day i might very well have a better experience than this guy anyways because i know how to tweak my settings to have beautiful clean and sharp images, good framepacing, no input lag etc. Gsync/Freesync IMHO nowadays is pretty much a must have...some games  play fine with Vsync, some will have horrible imput lag...if you want no imput lag and no screen tearing you need variable refresh monitors.

[Morgan MLGman]

1 minute ago, ivan134 said:

Doesn't boost 3.0 do that on its own?

Unless they can disable the Boost 3.0, they should run all of the cards with such setting turned on cause it just makes the cards incomparable cause one would be theoretically "overclocked" and the other wouldn't...

[ivan134]

1 minute ago, Morgan MLGman said:

Unless they can disable the Boost 3.0, they should run all of the cards with such setting turned on cause it just makes the cards incomparable cause one would be theoretically "overclocked" and the other wouldn't...

Every publication leaves it as is though. It technically shows the performance you get out of the box.

[Morgan MLGman]

1 minute ago, ivan134 said:

Every publication leaves it as is though. It technically shows the performance you get out of the box.

Well yeah, but doesn't the 980Ti boost itself out of the box as well? That's my problem with their methodology...

[ivan134]

1 minute ago, Morgan MLGman said:

Well yeah, but doesn't the 980Ti boost itself out of the box as well? That's my problem with their methodology...

Boost 2.0 is more like what you have on Polaris. Boost 3.0 is similar to what ryzen will be doing. It looks at all your limits and automatically overclocks to the max those limits can handle. Boost 2.0 is a pre-determined clock.

[Morgan MLGman]

2 minutes ago, ivan134 said:

Boost 2.0 is more like what you have on Polaris. Boost 3.0 is similar to what ryzen will be doing. It looks at all your limits and automatically overclocks to the max those limits can handle. Boost 2.0 is a pre-determined clock.

Ah, okay I understand now. Though it's still not a good way of testing imo: for reliable benchmarking I'd just set the clock manually for both cards at their stock frequencies