• Hey Guest. Check out your NeoGAF Wrapped 2025 results here!

R9 390 or GTX970? Whic one?

It can bring tangible improvements in some games (Crysis 3, Watch Dogs, Battlefield 4 MP and some others I'm forgetting), with DX12 it's also possible more logical cores = more performance in some cases.
It would still be nearly the same for most cases. I remember a few games where you got a 50% increase though
 
Nvidia cards dominate the top of those charts as far as single gpu solutions go.

The charts are showing how the 970 and 390 pull punches against each other in context of the OT. The 390 edges out the 970 in some modern games and even beats it out when some games becomes more demanding. More of a case of just analyzing upcoming trends.
 
I would like to remind people that Nvidia's drivers have been particularly bad in the last year. Of the games I followed, both GTA V and Witcher 3's "game ready" drivers were crap with crashes and stutter which forced a lot of people on this board to roll back to older drivers. There was also the Windows 10 fiasco where Windows Update was installing the actual newest Nvidia drivers but then Geforce Experience would apparently try to install the previous driver because that's what it thought was newest which would cause crashing and BSODing. In my limited personal experience, the Windows 10 driver issue actually damaged a friend's laptop card (a 760M), possibly by causing it to overheat (ie. artifacting followed by BSODing when this didn't happen under Windows 8.1).

So, basically what I'm saying is that the whole mantra of Nvidia's drivers being worlds ahead of AMD is no longer true. In terms of DX11 performance in general, yes, Nvidia's drivers are much more efficient but, in terms of stability and drivers on game launch, recently AMD has been performing a lot better. I've had no crashing and stability problems with my 280X while I watched a lot of people with Nvidia cards on this forum end up having to do rollbacks this year which was surprising.
 
I wouldnt buy the 970 because of the gimped vram. 970 may be good now until a year later more vram is needed so a 390 is a better long term investment.
 
But Star Wars Battlefront already performs better on AMD without DX12. Given that, it'd hardly say anything one way or the other.

As I said: It is too early to say something about dx12 performance on either side. Not enough games are out right now to make assumptions. We have to wait for Fable, SW:Battlefront and most interesting will be the next Deus Ex. Then we can start pointing fingers and start to claim which gpu is going to be more 'future proof'. Till then it is all speculation and wishful thinking on both sides.

And now for the big secret: Both cards won't be future proof.

It is also impossible to make assumptions how gpus will perform with dx12 based on their relative dx11 performance to each other. The performance difference isn't big enough.

battlefrontbenchmark2oow1.png
 
The charts are showing how the 970 and 390 pull punches against each other in context of the OT. The 390 edges out the 970 in some modern games and even beats it out when some games becomes more demanding. More of a case of just analyzing upcoming trends.

so youre just completely ignoring gpu tiers all together?

no? its just expected that the r9-390 and gtx970 trade blows if not the r9-390 should be better. the cards are similar in price range and the r9-390 is 9 months newer than the gtx 970. its falling into the gap that it should do. It does not mark a rise of AMD.

Personally I think the card is disappointing for this reason the r9-290 whipped the card it was designed to go up against and even nearly meets the gtx 970. for 18 months newer card to merely trade blows was just personally disappointing.
 
huh, i thought i had my mind set on a 970. thing is, im looking to jump up to the 144hz world pretty soon. i was going to make a post in the pc building thread and 144hz monitor thread but i figure ill do it here. my past two upgrades were amd cards because of they offered the best performance in my budget at that time so now i thought about going to nvidia just because. having said that, now i need to make the decision of nvidia + gsync or amd + freesync. would like to stay at 1080p because ultimately getting close to or staying at 144hz is my goal. at 1080p, my choices are pretty limited. suggestions?
 
huh, i thought i had my mind set on a 970. thing is, im looking to jump up to the 144hz world pretty soon. i was going to make a post in the pc building thread and 144hz monitor thread but i figure ill do it here. my past two upgrades were amd cards because of they offered the best performance in my budget at that time so now i thought about going to nvidia just because. having said that, now i need to make the decision of nvidia + gsync or amd + freesync. would like to stay at 1080p because ultimately getting close to or staying at 144hz is my goal. at 1080p, my choices are pretty limited. suggestions?

Neither will do. You probably want a 980, 980 Ti, Fury, or Fury X.
 
huh, i thought i had my mind set on a 970. thing is, im looking to jump up to the 144hz world pretty soon. i was going to make a post in the pc building thread and 144hz monitor thread but i figure ill do it here. my past two upgrades were amd cards because of they offered the best performance in my budget at that time so now i thought about going to nvidia just because. having said that, now i need to make the decision of nvidia + gsync or amd + freesync. would like to stay at 1080p because ultimately getting close to or staying at 144hz is my goal. at 1080p, my choices are pretty limited. suggestions?

I got a 970 paired up with the Asus VG248QE and I'd hit 144hz depending on some games. Sometimes hitting 144 if I lower some graphical settings enough. There's a lot more to monitors than just selecting one with G-sync or Freesync like display type and build quality. If you wanted to hit 144hz in games today at 1080p, the 980 TI and Fury-X will easily do that.
 
huh, i thought i had my mind set on a 970. thing is, im looking to jump up to the 144hz world pretty soon. i was going to make a post in the pc building thread and 144hz monitor thread but i figure ill do it here. my past two upgrades were amd cards because of they offered the best performance in my budget at that time so now i thought about going to nvidia just because. having said that, now i need to make the decision of nvidia + gsync or amd + freesync. would like to stay at 1080p because ultimately getting close to or staying at 144hz is my goal. at 1080p, my choices are pretty limited. suggestions?

If you're going above 60, then you'll take all the frames you can get and the 390 is the better choice. Freesync is also generally less expensive than Gsync.

I game on a 1080p144 monitor with crossfire 290s. When I had just one card, I would compromise with lower AA solutions and a mix of high and ultra settings to to play demanding modern games in the 90s range, everything else pretty much 120-144. The 390 should be even better. With AMD being so price competitive, you can also wait to get a second card at a good deal and rebate combo for crossfire. Multi GPU always has problems but overall I've been happy. And, again, that's another area DX12 looks to help out with.
 
no? its just expected that the r9-390 and gtx970 trade blows if not the r9-390 should be better. the cards are similar in price range and the r9-390 is 9 months newer than the gtx 970. its falling into the gap that it should do. It does not mark a rise of AMD.

Personally I think the card is disappointing for this reason the r9-290 whipped the card it was designed to go up against and even nearly meets the gtx 970. for 18 months newer card to merely trade blows was just personally disappointing.

its not a new card, its a rebrand of a card released 2 years prior. its also an architecture dating back to 2011. its going up against nvidias latest architecture from 2014. and it beats the 970 in every single one of those benchmarks
 
You also have a chance to unlock the 390 to a full 390X using a simple bios flash!

I honestly wouldn't try that. The chance of it happening is even less with the 300 series. There's a reason why some of the chips are cut down and have less cores. Binning process etc etc. Could try but you'd end up with massive artifacting. There is that one program that would check if it's doable if that's any fun.

EDIT:
http://www.overclock.net/t/1567179/...iji-unlockability-tester-ver-1-6-and-atomtool
 
Neither will do. You probably want a 980, 980 Ti, Fury, or Fury X.

I got a 970 paired up with the Asus VG248QE and I'd hit 144hz depending on some games. Sometimes hitting 144 if I lower some graphical settings enough. There's a lot more to monitors than just selecting one with G-sync or Freesync like display type and build quality. If you wanted to hit 144hz in games today at 1080p, the 980 TI and Fury-X will easily do that.

980ti and fury x arent out of the question. but at that point, should i make the jump to 1440p? i know itll be harder on the gpu and of course harder to hit my desired refresh rate. but i feel i need to say that my main purpose of going to 144hz is for mp fps. and at the moment, those games would be bf4/5 and battlefront. so im not exactly looking to max out current gen games at those refresh rates, although that would be nice.
 
980ti and fury x arent out of the question. but at that point, should i make the jump to 1440p? i know itll be harder on the gpu and of course harder to hit my desired refresh rate. but i feel i need to say that my main purpose of going to 144hz is for mp fps. and at the moment, those games would be bf4/5 and battlefront. so im not exactly looking to max out current gen games at those refresh rates, although that would be nice.

there is no single gpu thats overkill for 1080p. dont know where this constant stigma comes from
 
980ti and fury x arent out of the question. but at that point, should i make the jump to 1440p? i know itll be harder on the gpu and of course harder to hit my desired refresh rate. but i feel i need to say that my main purpose of going to 144hz is for mp fps. and at the moment, those games would be bf4/5 and battlefront. so im not exactly looking to max out current gen games at those refresh rates, although that would be nice.

Both should still be capable of pulling 120-144fps at 1440p. The cards were essentially built to tackle 1440p at serious settings. At 1080p, the cards would be seen as overkill by some, but you would have some of the most consistent frame times.
 
The only way for overkill at 1080p 60hz is if you do not use VSR/DSR (AMD/Nvidia). VSR can't change the fact that your monitor has 1920x1080 pixels and only refreshes 60hz a second, but rendering the game in 2560x1440 or higher really does wonders for aliasing. For games with limited AA or even no AA options it is a godsend, even if a native 1440 monitor would be better.
 
Not sure but ever since Windows 10 (which includes OS native capturing), I havent touched shadow play... I really hate all the proprietary software

It does. It's within the Raptr client and is called "Game DVR." It didn't work very well the last time I tried it, but I have my fair share of issues with ShadowPlay too, so YMMV.

Ok, now I'm choosing between a Sapphire R9 390 or a MSI one.

You really can't go wrong with either. Depends on what exactly you are looking for between the two.
 
May I ask what posts lead you to choosing the 390? Because I'm in the same position.

It's just the better card. Reviews, Neogaf, benchmarks all put them in the lead. It may consume a lot more, but that's another story.
Also, the possibility of getting a 390x by flashing it gives it points.
 
It's just the better card. Reviews, Neogaf, benchmarks all put them in the lead. It may consume a lot more, but that's another story.
Also, the possibility of getting a 390x by flashing it gives it points.

good thing you also read a bunch of reviews on your own. if you read a large enough sample size, its impossible not to come to the conclusion that the 390 is faster. its also architecturally stronger in the areas that matter more IMO.
 
good thing you also read a bunch of reviews on your own. if you read a large enough sample size, its impossible not to come to the conclusion that the 390 is faster.

I always do. I love computers and I always try to be up to date when it comes to them.
It's just hard to choose XD.
 
It does. It's within the Raptr client and is called "Game DVR." It didn't work very well the last time I tried it, but I have my fair share of issues with ShadowPlay too, so YMMV.

Tbh, I never could get Raptr's recording to work. Shadowplay works around most problems with window recording. Not exactly the best solution, but it's sufficient I suppose.

But what about heat? Heat would be a big problem for me in the summer

Heat will always be a problem. But the 300 series are tuned with better voltage regulation so they produce less heat than the 290 and 290X. The new coolers are also a big plus over the 290 and 290X.
 
How far behind is the 970 from the 390 in general? Because the 970 overclocks very, very well, better than the 390 AFAIK. So maybe that makes up some of the gap.

Either way, they're close enough that I would be more driven by what ecosystem (I know some of y'all don't like that term) you want to go with, e.g. g-sync, shadowplay, etc. vs AMD's offerings.
 
It's not FUD, the benchmarks show it.

Call it better driver support, whatever you want, but we all know it's architectural.

Benchmarks? As in multiple of them? Which benchmarks are those? These ones?

yxlylSx.png


fable-fps.gif


fable4k-fps.gif


What's so special in these benchmarks compared to the average DX11 performance landscape?

rh17I9y.png


Let's check:
- 980Ti is a bit faster than Fury X in both DX12 benchmarks and in DX11 on average
- 390 is a bit faster than 970 in both DX12 benchmarks and in DX11 on average

Unless I'm missing something it doesn't look like DX12 is changing anything in performance landscape of DX11. So this means that even these two DX12 benchmarks we have at the moment do not show any "better DX12 support" from AMD cards at all.

change that line to "better dx12 performance" and its not fud.

It's FUD in its purest form simply because you can't base such judgements on two benchmarks available at the moment. 970 and 390 are both good cards for 1080p, they are both likely to choke and die in 4K and will be hit and miss in 2560x... The choice is simple - if you feel like you need more RAM go with 390, if you prefer NV for some reason (PhysX I dunno) - go for 970. It's very unlikely that going with 390 over 970 you will get any DX12 benefits which won't be available to 970 and vice versa.
 
Benchmarks? As in multiple of them? Which benchmarks are those? These ones?

yxlylSx.png


fable-fps.gif


fable4k-fps.gif


What's so special in these benchmarks compared to the average DX11 performance landscape?

rh17I9y.png


Let's check:
- 980Ti is a bit faster than Fury X in both DX12 benchmarks and in DX11 on average
- 390 is a bit faster than 970 in both DX12 benchmarks and in DX11 on average

Unless I'm missing something it doesn't look like DX12 is changing anything in performance landscape of DX11. So this means that even these two DX12 benchmarks we have at the moment do not show any "better DX12 support" from AMD cards at all.



It's FUD in its purest form simply because you can't base such judgements on two benchmarks available at the moment. 970 and 390 are both good cards for 1080p, they are both likely to choke and die in 4K and will be hit and miss in 2560x... The choice is simple - if you feel like you need more RAM go with 390, if you prefer NV for some reason (PhysX I dunno) - go for 970. It's very unlikely that going with 390 over 970 you will get any DX12 benefits which won't be available to 970 and vice versa.

In the 3 or 4 available dx12 benches, the 390 beats the 970 across the board. What part of that doesnt equate to "the 390 performs better in dx12?"

Were no longer suppose to base judgements on benchmarks?
 
Benchmarks? As in multiple of them? Which benchmarks are those? These ones?

yxlylSx.png


fable-fps.gif


fable4k-fps.gif


What's so special in these benchmarks compared to the average DX11 performance landscape?

rh17I9y.png


Let's check:
- 980Ti is a bit faster than Fury X in both DX12 benchmarks and in DX11 on average
- 390 is a bit faster than 970 in both DX12 benchmarks and in DX11 on average

Unless I'm missing something it doesn't look like DX12 is changing anything in performance landscape of DX11. So this means that even these two DX12 benchmarks we have at the moment do not show any "better DX12 support" from AMD cards at all.



It's FUD in its purest form simply because you can't base such judgements on two benchmarks available at the moment. 970 and 390 are both good cards for 1080p, they are both likely to choke and die in 4K and will be hit and miss in 2560x... The choice is simple - if you feel like you need more RAM go with 390, if you prefer NV for some reason (PhysX I dunno) - go for 970. It's very unlikely that going with 390 over 970 you will get any DX12 benefits which won't be available to 970 and vice versa.

Yup, I'm definitely getting a 390 now.
 
In the 3 or 4 available dx12 benches, the 390 beats the 970 across the board. What part of that doesnt equate to "the 390 performs better in dx12?"

Were no longer suppose to base judgements on benchmarks?

Name these 3 or 4 available DX12 benchmarks.
Look up DX11 benchmarks and check if 390 in all of them are slower then 970.
 
Top Bottom