• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

GTX 570 - so disappointing

gbdtil1

gbdtil1

gbdtil1

Associate
Joined
24 Mar 2006
Posts
219
Just installed my new Zotac 570, having decided to move back to Nvidia following a brief spell with ATI via their 5870. I thought, rightly or wrongly, that the GTX 570 would give me another 10-12% performance hike over the 5870 but it appears to have gone the other way instead, ie 12%+ decrease.

I've installed the latest driver and would like to highlight some of my reference points for any comments you knowledgable folk may have:

3D Mark '06 - 5870 = 17563 (average of 3 runs)
- GTX 570 = 17,150 (again 3 run average)

Crysis - Level 19 (Recovery), 3 passes = min 23, max 43, ave 31 (5870)
= min 16, max 36, ave 26

Crysis run at 1920 x 1080; Very High detail and 4x AA

Key system specs in brief:

Q6600 @ 3.2ghz
4 Gig DDR 2 @ 800htz
ASUS Maximus x38 mobo
GTX 570
Win 7 64 bit

Really confused by these results as I thought the 570 would murder the 5870 in Crysis with AA enabled.

Grateful for any thoughts.........
 
Forget 06 it's of no use when comparing GPU performance, use vantage or mark 11, as for crysis, the 5870 has always been up there with the 480/570 in performance in that particular game, clock the 570 to 800+ core and you will see some very nice gains.
 
Last edited:
AMD's latest GPUs seem to be better with Crysis, than nVidia's, I seem to remember that the 5870 beat the GTX480, which means it'll also beat a 570, additionally, why are you using 3D Mark 06? It's a 5 year old benchmark now, and its results are extremely irrelevant.
 
RavenXXX2 said:
Forget 06 it's of no use when comparing GPU performance, use vantage or mark 11, as for crysis, the 5870 has always been up there with the 480/570 in performance in that particular game, clock the 570 to 800+ core and you will see some very nice gains.
Click to expand...

is my q6600 @ 3.2Ghz likely to be causing a bottleneck? Not up to date with Nvidia anymore, used to use Ntune to overclock - what's the current tool and where do I get it?

''AMD's latest GPUs seem to be better with Crysis, than nVidia's, I seem to remember that the 5870 beat the GTX480, which means it'll also beat a 570, additionally, why are you using 3D Mark 06? It's a 5 year old benchmark now, and its results are extremely irrelevant. ''

I guess I'm just familiar with '06 and have a better feel for what the scores actually represent. I've just run the basic 3d mark 11 and produced a score of P4536 with a graphics score of 4800 and Physics of 4019. These seem very low to me but I guess it's the latest standard so not likely to be at 3d mark '06 levels?!
 
3DMark 06 levels will always be higher than the latest versions. However, 3D Mark 06 stopped being reliable about 2 years ago now, my old 3870X2 gave me around 20,000 points, a card which is much much slower than a GTX570 or 5870 (about 1/3 the performance). It'd be much better for you to test more than just Crysis as well because of what I said about it too.
 
Tried it with the DiRT2 benchmark? Or Heaven? Or Tropics? Any of them I'd value over Crysis or 3DMark06. (And I'm a Crysis lover. Barely play anything else).
 
3dmark06 is cpu bound, original crysis 1 performed well on 58xx but not good on fermi.

Try some other stuff and you should see the 570 pull away but the real world difference won't be that noticeable.
 
Down load msi afterburner unlock voltages in settings.

Set volts to 1.1 and core clock to 900 memory to 2000 make sure u set up your own fan profile using 50% fan speed as your min and run heaven 2.1

If heaven crashes or shows artifacts just up voltage a little rerun benchmark

The 570 loves to be clocked running mine both at 950 core 2200 mem @ 1.160v
 
Sindri said:
This is BS, some games are very cpu intensive and your q6600 could prove to be a bottle neck at just 3.2Ghz. Just check a few cpu reviews to see the difference a higher overclock can make.
Click to expand...

It's not BS, just not qualified. For most games the q6600 won't be a bottleneck, for a small % of generally more recent games it will bottleneck it to some degree.
 
Sindri said:
This is BS, some games are very cpu intensive and your q6600 could prove to be a bottle neck at just 3.2Ghz. Just check a few cpu reviews to see the difference a higher overclock can make.
Click to expand...

Surely thats a CPU bottle neck that you would get regardless of GPU though?
I think he is asking if the GPU is being bottle necked by the CPU, not the game being bottle necked by the CPU.
 
straxusii said:
It's not BS, just not qualified. For most games the q6600 won't be a bottleneck, for a small % of generally more recent games it will bottleneck it to some degree.
Click to expand...

But as we we don't know which games the OP uses that guy cannot say that the cpu is in no way a limitating factor.
 
gbdtil1 said:
is my q6600 @ 3.2Ghz likely to be causing a bottleneck?
Click to expand...

It is a bottleneck no matter how many people on the same platform try to defend their ageing hardware! But both of the GPU's were on the same bottleneck, maybe one performs better with the bottleneck? Don't really understand why you'd go from a 5870 to a GTX570.
 
Vegeta said:
It is a bottleneck no matter how many people on the same platform try to defend their ageing hardware! But both of the GPU's were on the same bottleneck, maybe one performs better with the bottleneck? Don't really understand why you'd go from a 5870 to a GTX570.
Click to expand...

http://www.anandtech.com/bench/Product/306?vs=294

Seems to be some not bad gains looking at that. What don't you understand ?
 
Vegeta said:
My bad. A extra 3fps in Crysis Warhead at the settings he uses. MASSIVE gains.
Click to expand...

Don't be the pedant, it is quite obvious i was talking overall performance. Or are you saying he upgraded purely for Crysis Warhead as i got the impression he was using that as an example of the performance gain(or not).

Maybe the original poster could clarify.
 
Stanners said:
Don't be the pedant, it is quite obvious i was talking overall performance. Or are you saying he upgraded purely for Crysis Warhead as i got the impression he was using that as an example of the performance gain(or not).

Maybe the original poster could clarify.
Click to expand...

Really appreciate all the feedback on this guys even allowing for indifference in opinion ;)

Crysis was merely used as a benchmark as I believe it still sets the bar in terms of pushing GPU's. I also have Warhead but didn't think to use that as I believe it's not as demanding as Crysis??

Bad Company 2 is the current flavour for me and I generally play the shooters and driving (Dirt, Grid etc). The comparisons at AnandTech provide some comfort when viewed across a broad spectrum of games but I had no idea the 5870 would outperform the 570 in Crysis - which I always enjoy going back to and will now need to 'de-tune' to a certain extent to make it playable again!

Would it make much / any difference if I open up my CPU from 3.2Ghz to 3.4Ghz? I've managed to get it to 3.6Ghz in the past just for benching, but wouldn't want to run it at that for extended periods.
 
Only real significant gains I can see is when games are runned at 2560X1600 which the OP does not and when the game is at high frames per second (i.e average 70fps over 100 which won't matter in the real world assuming he has a 60hz monitor which i'm confident he will). And considering the only game he mentions in his OP is Crysis I think it's fair to say he'll judge the performance and his happiness with his upgrade based on the performance of that game.
 
You must log in or register to reply here.
Back
Top Bottom