• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

5970 vs 280 SLI benchmarks

It sounds like there's something up with JCDentons machine.
Posted via [H] Mobile Device
 
actually if u think logically 5970 is clocked to 5850 speeds and 5850 isn't that different from gtx 285 in crysis or other nvidia titles and unless game uses shader power there won't be much difference. i suggest you to oc to 5870 speeds and then test games that ati is superior like shift street fighter or something like that
 
Let's add this up:

(1) No fan slider in the CCC. Had not read anywhere that this is a common problem.
(2) Cannot launch Dragon Age or NFS Shift (which run perfectly on my single 5850)

Sure seems that something is amiss here.

I know the ATI drivers are still a little flaky. I'm hoping to see some improvement by the time the 9.12 finals get here sometime next month.

I agree with the people here that say if you are not happy with the card, you should return it. I think your conclusions about the performance of the card are a little premature until we get some good stable drivers for it. There may be an issue with the card itself. I had to RMA the first 5850 I bought, as the fan would barely spin in auto mode.
 
Hi there thought I would jump in here two.I have 3x285trisli also have 5970 with another on the way.
To be honest 5970 does beat the trisli in a few games never mind sli.
Crysis runs great for me on both setups but I much prefer it on the 285 cards.No particular reason why.

There are slider controls in CCC if you update your Bios.
Think dopey already posted it here somewhere
http://www.overclock.net/ati/614095-5970-bios-update-fan-control-fix.html

Dragonsage and shift run fine.

Vantage with single 5970 note max temp reached was 70c on the core.did not check vram thru gpu-z
core 1000
mem 1150
 
Hi there thought I would jump in here two.I have 3x285trisli also have 5970 with another on the way.
To be honest 5970 does beat the trisli in a few games never mind sli.
Crysis runs great for me on both setups but I much prefer it on the 285 cards.No particular reason why.

There are slider controls in CCC if you update your Bios.
Think dopey already posted it here somewhere
http://www.overclock.net/ati/614095-5970-bios-update-fan-control-fix.html

Dragonsage and shift run fine.

Vantage with single 5970 note max temp reached was 70c on the core.did not check vram thru gpu-z
core 1000
mem 1150
your card is champ mate :D probably one of the best oclers out there
 
best thread ever. The poor guy is simply stating that he was expecting a little more of the results and everyone jumps on him telling him that he is wrong.
It's stil a great card and better than any other single card offerings.
However the OP was comming from a dual card setup comparable to a GTX295 and expected a bigger jump in the the most graphic intensive game Crysis and didn't get it. Reviews back his up...end of story.
The HD5970 IS top dog, so everyone can relax...he just expected more of a leap in performance for the huge outlay that's all.

no, that score in Crysis is slower than my 5870.

there is definitely something wrong there, I had tried to CF my 5870 and the performance is at least half more than his number. that is what everyone else is talking about.
 
Last edited:
Er no...

The 5970 even at stock aces Crysis in 2560x1600 16xAF max settings, to be getting those frame rates in about HALF the resolution is just wrong.

Comparing the 5970 stock agianst anything isn't really a good comparison anyway, the card is designed around overclocking, come back with at least 5870 clocks speed benchmarks and you'll get a more accurate idea of what this card is going to be used for.

your card is champ mate probably one of the best oclers out there

Actually I think most of the 5970 cards will happily hit them speeds, the reason he's got better OC's is because he's volt modded past the preset values that come with AMDs overvolt tool (and most other board partners since most are limiting it as well) Read any review of that card and no one exceeds 1.1625v on the core and 1.15v on the memory, because that is the preset overvolt values.

If you're interested, the tool used for overvolting past the preset limit is MSIs afterburner 1.4 which works on all 5970 cards, thanks to Gerardfraser for passing on the benches for this card and the info on voltmodding it past the presets.
 
Last edited:
LOL, this calculation is full of fail. You should know better than that. Why not just rely on the actual numbers instead of your made up theories? Oh, that's because the actual numbers show things like:

http://www.tomshardware.com/reviews/radeon-hd-5970,2474-8.html
Crysis 1920x1200
GTX285 SLI: 46.9 fps , 5970: 49.4 fps

LOL @ a <10% gain for "next-gen" dual gpu.

No, you were talking about Warhead, not Crysis 1. Different games, different numbers.

So I think I'll take a page out of your book and dispute your Crysis numbers by providing NFS: Shift numbers: http://hardocp.com/image.html?image=MTI1ODQ5OTg1NkJiV3FOdkVhR1JfNF80X2wuZ2lm

100%+ FPS increase, booya! :rolleyes:
 
I discovered the source of the poor performance in Crysis. Sadly the card just can't handle what I was throwing at it. Digging around in the configs, I was using one of the ones floating around for beyond max settings that has enhanced HDR, sunshafts, and mid/distant textures. Well these extra enhancements caused the 5970 to choke and bog down a lot whereas the sli setup had no trouble with them. At a stock max config (which doesn't look nearly as good), the 5970 ended up beating the 280s by 10% and didn't have any of the stuttering from before. So we have a $600 graphics card that can't handle a tweaked config. Hopefully it was just because drivers were rushed. The 5970 struggles pretty badly against last gen cards here, even losing to GTX285 sli sometimes:

http://www.bjorn3d.com/read.php?cID=1741&pageID=8057
 
I discovered the source of the poor performance in Crysis. Sadly the card just can't handle what I was throwing at it. Digging around in the configs, I was using one of the ones floating around for beyond max settings that has enhanced HDR, sunshafts, and mid/distant textures. Well these extra enhancements caused the 5970 to choke and bog down a lot whereas the sli setup had no trouble with them. At a stock max config (which doesn't look nearly as good), the 5970 ended up beating the 280s by 10% and didn't have any of the stuttering from before. So we have a $600 graphics card that can't handle a tweaked config. Hopefully it was just because drivers were rushed. The 5970 struggles pretty badly against last gen cards here, even losing to GTX285 sli sometimes:

http://www.bjorn3d.com/read.php?cID=1741&pageID=8057

their benchmark is way off, no idea where he got 74fps on GTX 295 with very high..

from that point, this review is not even legit... their numbers are entirely off the chart...
 
I discovered the source of the poor performance in Crysis. Sadly the card just can't handle what I was throwing at it. *Snip
http://www.bjorn3d.com/read.php?cID=1741&pageID=8057

One of the reasons that the hardware review site like hardocp need to exist is that some users "test" their cards by thinking they know what they are doing, actually they don't and spread the misinformation around.

did you check this link? does it look like any struggle to you?

http://hardocp.com/image.html?image=MTI1ODQ5OTg1NkJiV3FOdkVhR1JfNF80X2wuZ2lm
 
Lol, when eric posted a review showing the 5970 smoking everything nobody says anything, but when the OP shows a link from the same site showing that the 5970 isn't that much better and loses in some tests now the site is magically wrong? Lol.
While i agree that maybe that site isn't the best, you cant not say something to 1 person about it and not the other.

From looking at all the different reviews out there, it looks more like it all depends on the game being reviewed depends on how well said cards perform.
Once drivers mature for the 5970 then yea it will hopefully perform much better in every test. Not that it don't seem to do well now :D.
 
Lol, when eric posted a review showing the 5970 smoking everything nobody says anything, but when the OP shows a link from the same site showing that the 5970 isn't that much better and loses in some tests now the site is magically wrong? Lol.
While i agree that maybe that site isn't the best, you cant not say something to 1 person about it and not the other.

From looking at all the different reviews out there, it looks more like it all depends on the game being reviewed depends on how well said cards perform.
Once drivers mature for the 5970 then yea it will hopefully perform much better in every test. Not that it don't seem to do well now :D.

since when does Crysis with GTX 295 can get to 60+fps on max setting with 1680*1050 resolution?

that review is entirely inaccurate from that point. ;)
 
One of the reasons that the hardware review site like hardocp need to exist is that some users "test" their cards by thinking they know what they are doing, actually they don't and spread the misinformation around.

did you check this link? does it look like any struggle to you?

http://hardocp.com/image.html?image=MTI1ODQ5OTg1NkJiV3FOdkVhR1JfNF80X2wuZ2lm

There was no misinformation, I said up front that Crysis was using a modified config. How was I to know that a brand new $600 "next generation" graphics card would choke on a few extra textures and sunshafts that didn't even phase 1.5 year old cards? Both cards should suffer from the extra effects equally, and it says bad things about the 5970 when it suffers a lot more
 
Set graphics to Very High w/ 4xAA.

Play the fucking game.

Come back in an hour and tell us if its still about as fast as GTX 280 SLI set up.
 
There was no misinformation, I said up front that Crysis was using a modified config. How was I to know that a brand new $600 "next generation" graphics card would choke on a few extra textures and sunshafts that didn't even phase 1.5 year old cards? Both cards should suffer from the extra effects equally, and it says bad things about the 5970 when it suffers a lot more

What custom config are you using?

Just curious as 1920 x 1080 very high 4xaa with doomsword's cfg feels very playable (smooth) on my system with 1 5870.
 
My 260 core 216 Tri-Sli still rapes ATI's latest offering, of course not in dx11. :p
But honestly i don't think DX11 is a big deal, i actually see a nice performance margin with DX10 over DX9. I find it amusing that people would argue over such trivial benchies. I hope their drivers mature and it does beat out my setup. And at the same time Nvidia is still improving the tri-sli performance. The 1366 boards really opened up a performance gain for SLI users, and even more so for tri-sli. I think we need to give ATI some appropriate time before getting some accurate benches on the new cards. IMO catalyst has always been way behind the release of the cards.
 
My 260 core 216 Tri-Sli still rapes ATI's latest offering, of course not in dx11 :p
But honestly i don't think DX11 is a big deal, i actually see a nice performance margin with DX10 over DX9. I find it amusing that people would argue over such trivial benchies. I hope their drivers mature and it does beat out my setup. And at the same time Nvidia is still improving the tri-sli performance. The 1366 boards really opened up a performance gain for SLI users, and even more so for tri-sli. I think we need to give ATI some appropriate time before getting some accurate benches on the new cards. IMO catalyst has always been way behind the release of the cards.

Man I gotta buy those 260 cards.I noticed alot people say there great.
I have tried out in the last year
3x285
3x275
3x9800+
3x4890
Never the 260 cards.There even some guys on the EVGA forum that swear the 3x260 beats 3x285,I just gotta try them out.
 
My 260 core 216 Tri-Sli still rapes ATI's latest offering, of course not in dx11. :p
But honestly i don't think DX11 is a big deal, i actually see a nice performance margin with DX10 over DX9. I find it amusing that people would argue over such trivial benchies. I hope their drivers mature and it does beat out my setup. And at the same time Nvidia is still improving the tri-sli performance. The 1366 boards really opened up a performance gain for SLI users, and even more so for tri-sli. I think we need to give ATI some appropriate time before getting some accurate benches on the new cards. IMO catalyst has always been way behind the release of the cards.


What ?? one 5970 (5870 clocks) will smoke GTX260 tri sli on high resolution with AA and (9.12 drivers)
 
LOL, this calculation is full of fail. You should know better than that. Why not just rely on the actual numbers instead of your made up theories? Oh, that's because the actual numbers show things like:

http://www.tomshardware.com/reviews/radeon-hd-5970,2474-8.html
Crysis 1920x1200
GTX285 SLI: 46.9 fps , 5970: 49.4 fps

LOL @ a <10% gain for "next-gen" dual gpu.

Not that I have either card, or combination thereof, but do you seriously buy the most expensive cards and play with no AA/AF on whatsoever? (or really care about a benchmark with both off?).

Just look at the same page you linked to prove your point, and then look at the chart with AA on. The Nvidia cards pretty much die there. Even the 5850x2 (3rd for ATI) manages to beat Nvidia's best.

I personally hope Nvidia comes out with something good soon, as I can't afford the above MSRP prices the latest ATI cards are being sold for now. But to pick one irrelevant (no AA for high end card) benchmark to draw conclusions is of questionable motive imo.
 
since when does Crysis with GTX 295 can get to 60+fps on max setting with 1680*1050 resolution?

that review is entirely inaccurate from that point. ;)
Of course it must be...I mean, it does not show the numbers you WANT it to show so it must be wrong, right? ;)
 
Man I gotta buy those 260 cards.I noticed alot people say there great.
I have tried out in the last year
3x285
3x275
3x9800+
3x4890
Never the 260 cards.There even some guys on the EVGA forum that swear the 3x260 beats 3x285,I just gotta try them out.

They are awesome, but sincle you've tried 3 of the GTX275's there would be little point since the 275 clocks better and has more stream processors.
Price wise (for the US anyway) the 260 is the way to go though.

Back on topic eveything I'v seen about the HD5970 looks good, I think it's only drivers that are getting in the way at this stage in some instances,
 
They are awesome, but sincle you've tried 3 of the GTX275's there would be little point since the 275 clocks better and has more stream processors.
Price wise (for the US anyway) the 260 is the way to go though.

Back on topic eveything I'v seen about the HD5970 looks good, I think it's only drivers that are getting in the way at this stage in some instances,

There is no way 3x260s beat 3x285s unless the 260s are OCd and the 285s are left at stock speeds.

I also expected a bit more performance from the 5970. It barely beats out the 5870 on most benches I've run so far.
Perhaps I'm resolution limited. And would see more of a difference with 3 screens, who knows...
 
Of course it must be...I mean, it does not show the numbers you WANT it to show so it must be wrong, right? ;)

Show me when did a single 295 can get 60fps in max setting with crysis...

I highly suggest you to stop trolling and personal insult first before you make the next post..
 
There was no misinformation, I said up front that Crysis was using a modified config. How was I to know that a brand new $600 "next generation" graphics card would choke on a few extra textures and sunshafts that didn't even phase 1.5 year old cards? Both cards should suffer from the extra effects equally, and it says bad things about the 5970 when it suffers a lot more

Reason why ATI struggles with crysis is not because of their hardware. Its because they dont pay crytek like Nvidia does so they can get their product optimize.
 
Reason why ATI struggles with crysis is not because of their hardware. Its because they dont pay crytek like Nvidia does so they can get their product optimize.

actually, Crysis on Max setting is ATI flavor. This is not just from my experience, but also [H] review shows that..
 
actually, Crysis on Max setting is ATI flavor. This is not just from my experience, but also [H] review shows that..

Which is a testament to AMD's cards. Crytek worked closely with Nvidia, and even pimped the 8800 series during the lead up to Crysis. It was shameless.
 
Thanks for this. Looks like I will be sticking with my 280s for a bit longer!
 
What custom config are you using?

Just curious as 1920 x 1080 very high 4xaa with doomsword's cfg feels very playable (smooth) on my system with 1 5870.

It was the 1.2 version of this:

http://www.crymod.com/thread.php?threadid=31242

I think I also increased motion blur and swapped in the HDR settings from a different one, called Crysis Ultra Config or something like that. I think I'll try the latest 1.34 version and see how it runs.
 
hey JCDenton, what version of windows are you using? if its vista or windows 7, you should be able to switch between monitors with out having to use the CCC in ati's drivers. I always just use the built in functions of windows for pulling up multiple monitors or switching between screens, its as easy as right clicking on the desktop and going to change resolution, but i guess if your doing something with eyefinity then yes will will need to use CCC
 
Last edited:
Not that I have either card, or combination thereof, but do you seriously buy the most expensive cards and play with no AA/AF on whatsoever? (or really care about a benchmark with both off?).

I personally hope Nvidia comes out with something good soon, as I can't afford the above MSRP prices the latest ATI cards are being sold for now. But to pick one irrelevant (no AA for high end card) benchmark to draw conclusions is of questionable motive imo.

I'll tell you what my "questionable" motives were: I am a hardcore gamer who just bought a $600 card and wanted to see the performance in the two games I have that consistently dipped under a 60fps minimum. I'm not that excited to test Left 4 Dead to see if it will jump from 500fps to 550fps. Everyone already knew that ATI cards will give ridiculous framerates in undemanding games. So people who already have high end setups are going to mainly be wondering if the 5970 provides a boost where it matters. Right now it looks like there's no rush for those people to upgrade and it's probably smarter to wait on Nvidia/the next ATI cards and see if games will come out to use them or if the console-centric development continues.

As for the tests I ran, maybe you should re-read. 16xaf was used for all the tests. 4xAA was used for the unigine tests, and some AA was turned on for ARMA although it can only be done through the in-game menu by choosing "low", "med", ect so I don't know the actual level applied. Crysis was not tested with AA because it looks better without. As last I checked, you can either turn on normal AA through the config file or game menus (it can't be forced in drivers), or you can turn on edge aa through the config but both can't be used simultaneously. Crysis looks much better with edge AA on because it cleans up the vegetation, and there is far more vegetation than there are hard edges that would be helped by normal AA. Crysis does seem to be overclocking friendly, with clocks increasing FPS almost 1:1. So in the end I'm able to get a solid experience out of it since my second GTX 280 wouldn't overclock much at all, forcing me to leave them both stock.


hey JCDenton, what version of windows are you using? if its vista or windows 7, you should be able to switch between monitors with out having to use the CCC in ati's drivers. I always just use the built in functions of windows for pulling up multiple monitors or switching between screens, its as easy as right clicking on the desktop and going to change resolution, but i guess if your doing something with eyefinity then yes will will need to use CCC

Thanks, will check that out later. I found a somewhat easier way to switch in CCC, but it's still messed up when I go from the plasma to the PC monitor. The resolution is set at 1920x1080 and the refresh is 25hz. I have to select "force" from CCC to get it back to 1920x1200 60hz. The nvidia cards just automatically set it properly. If this works it'll save a lot of trouble.
 
it seems like you're on a crusade to bash the 5970, whatever floats your boat man.

from my experience, my 5850 beat out the GTX 285 (oc'd) in batman, by a fair amount with Physx on High too. ~40fps to ~60fps (2560x1600)

anyway enjoy whatever card you like, no one here is stopping you, but if you decide to go on some ill-informed fanboy crusade like it seems you are doing, you're just going to get flamed off the boards :p
 
Why is that surprising ? he return the 5970 and replace with Nvidia card/card so now he must make hes replacement worth :rolleyes:
 
I've been doing some objective research on the 5970, and let's put it this way: No matter what you throw at this dual-gpu card, there's no budge in performance. And that's how it's going to be until 2011, or for as long as videogame companies continue to work on 360/PS3 development and continue placing PC ports/DX11 development on the back burner...

But I have two relevant questions for this thread:

1. Is the 5970 overkill at 1900x1200? Or does the increased AA ability make it worth the purchase?
( Cuz I can't even find a decently priced 30" monitor with a 2560x1600 resolution... and I may be stuck at 19x12 for a while )

Also, I have a 2nd GTX280 on the way to my house right now. No doubt these cards scale well in SLi, so I can't wait to gain some performance! ...BUT!

2. Aside from its DX10 Limitations and heat issues aside: Is Tri-SLi GTX280's comparable to 5970 performance? And is Tri-SLi worth it at 1900x1200??

Because I can save myself some trouble if I just buy a third GTX280. And I just want to wait for prices to come down on DX11 cards so I can watercool them. :)
 
Last edited:
One question that's kind of important:

Is the 5970 overkill at 1900x1200? Or does the increased AA ability make it worth the purchase?

I can't even find a decently priced 30" monitor with a 2560x1600 resolution...

At that resolution, the 5870 is perfect. I run just about everything at the highest in game settings with 8xAdaptive AA and 16xAF at 1920*1200. The 5970 will only really show it's true potential with an Eyefinity setup. If we could run a 5970 vs GTX295 at 5760*1200, we would see just how good AMD's new GPU is.
 
Depends from games.

Stalker CS, Crysis, Arma 2 will make one 5870 to a serious test with AA/AF.
 
Depends from games.

Stalker CS, Crysis, Arma 2 will make one 5870 to a serious test with AA/AF.

Which is why I said "just about everything" in my post. I don't play Stalker or Crysis, so i can't comment on playable settings for them, but with ARMA 2, I was able to put everything on highest or a notch down with 2xAA and had a very enjoyable experience. Surprisingly, a game that i really like that puts a stress on my card at highest settings is Red Faction:Guerrilla. Even with my overclock, I only average around 45-50FPS until buildings start collapsing, which it then drops down to 27FPS due to the high CPU dependency of the physics engine.
 
Back
Top