5870 (vs GTX 280) Results: Confusing

Adding more RAM is not going to solve his problem. If RAM was an issue the games he'd playing would go from smooth to a stuttering mess. I reckon the only option at this point is to reformt and try Windows 7, but, I personally don't think it is the issue either. Doesn't make sense that the software would impact performance *that* much.
 
the time its taken you to type these giant posts, you could have done the clean install. and even, on the off chance its not the problem, at least you would have a fresh install instead of a debate.
 
I don't know if it's been mentioned, but why not download the drivers from AMD/ATI's website? From what I've read those drivers are based on cat 9.10 with performance improvements versus most driver disks which have a modified version of Cat 9.9

As far as software impacting hardware, well hardware is virtually worthless without software and it's totally dependant on it. I think it's entirely possible to cause that much of an impact.
 
Hey just one question i'm to lazy to read the whole tread, are you using the drivers from the cd?

LOL Ramon was 20 sec faster then me :)
 
Why don't you do a test not using your "gutted" install? Do a fresh install, use the latest drivers from ATI and for everything else in your system and see if your performance is still poorer than you think it should be.
 
I would just start with a driver uninstall/reinstall with updated ones. I have NEVER needed to reinstall an OS for a video card upgrade, and i've done well over 100 of them. I wouldn't completely discount it, but I would certainly save it as a last resort.
 
Relax man i'm 90% sure its driver issue had the same problem with my GTX295 after some driver sweeper action, did fresh install of newest drivers everything worked great ;)
 
Update: ...who would have thought...

Customer Review on Newegg

Cons: "You can't use this for gaming on XP, at least not right now. ATI/AMD is not showing support for XP, not 32 or 64 bit. The drivers will install, but the card is sluggish and unresponsive, as well as the fps are no better than with my old 4850 in XP. CCC in xp also does not read the core or ram speeds in XP correctly, so I don't recommend this for folks hanging on to XP."

There ya go... those who thought it could be the OS, no matter how unlikely that seems, I suppose were right.

I already had a fresh install, reinstalled every game and my results last night were the same. I was testing all night (hence why no reply until now).

As I said in one of my earlier posts: if the OS was the problem, ****ing AMD should have been quite aware and made it clearly known to the public. Not a "coincidence" that someone "just happened" to write the above while we were all here trying to figure this out.

I will keep the GPU and wait until I get a hold of Win7 64. I will then try the 5870 again, with 4GB RAM.

Again, the E8600 stock or not is not going to bottleneck this GPU. High-end dual-cores are still the way to go for gaming, and though I'm not doubting the great performance of the i7's, the 920 is not even clocked as high as a stock E8600 (3.33), and games do not take advantage of quad-cores.

So, I'm just going to drop this here, because I've wasted too much time with this damn thing already.

When I can get a copy of Win7 64, I will try it again. If the results at that point still do not equal/come far closer to this thing running akin to a 295 and crush the 280, I'll return it for another.

If it still does not run up to par, then and only then can I tell anyone whether or not any of the hardware I run is indeed causing some weird serious bottleneck.

I will do so at that point, to let people know... just don't know when that's going to be.

Thanks to everyone who tried to figure out this debacle.
 
Not really surprising. As of October 22nd XP will be 2 generations old. It's still a great OS for the typical business/office environment since things don't change much, but this is another animal.
 
yes, we were saying "no, it couldnt be the os, amd would definitely put a warning on the box, or it would have got out in the community". silly us. but at least you know it was that overclocking or ram crap.

as for win 7, i am pretty sure the ms program has run out, no more dl, no more keys (i could be mistaken). but it is very easy otherwise to get a copy that you can use until sometime in the spring.
 
Yeah, Microsoft claims the XP architecture would have to partially (major) rewritten to even support dx10, so new cards not running faster, is no shock on an old OS.
 
RamonGTP said:
Not really surprising. As of October 22nd XP will be 2 generations old. It's still a great OS for the typical business/office environment since things don't change much, but this is another animal.
Click to expand...

Ya it is surprising on how long XP has been around. It's about 8 years old now. Just think about 8 years or so before XP release (2001). People were probably using 3.1, according to wikipedia it came out in 1992. I know you can't make a straight comparison on progress considering all the factors, just a weird thing to think about comparing the 2 time lines.
 
Those good 'ole Win98 days. I was running a 400 Mhz AMD K6 and some hardcore integrated graphics at the time lols.
 
Slachtbeest said:
Update: ...who would have thought...

Customer Review on Newegg

Cons: "You can't use this for gaming on XP, at least not right now. ATI/AMD is not showing support for XP, not 32 or 64 bit. The drivers will install, but the card is sluggish and unresponsive, as well as the fps are no better than with my old 4850 in XP. CCC in xp also does not read the core or ram speeds in XP correctly, so I don't recommend this for folks hanging on to XP."

There ya go... those who thought it could be the OS, no matter how unlikely that seems, I suppose were right.

I already had a fresh install, reinstalled every game and my results last night were the same. I was testing all night (hence why no reply until now).

As I said in one of my earlier posts: if the OS was the problem, ****ing AMD should have been quite aware and made it clearly known to the public. Not a "coincidence" that someone "just happened" to write the above while we were all here trying to figure this out.

I will keep the GPU and wait until I get a hold of Win7 64. I will then try the 5870 again, with 4GB RAM.

Again, the E8600 stock or not is not going to bottleneck this GPU. High-end dual-cores are still the way to go for gaming, and though I'm not doubting the great performance of the i7's, the 920 is not even clocked as high as a stock E8600 (3.33), and games do not take advantage of quad-cores.

So, I'm just going to drop this here, because I've wasted too much time with this damn thing already.

When I can get a copy of Win7 64, I will try it again. If the results at that point still do not equal/come far closer to this thing running akin to a 295 and crush the 280, I'll return it for another.

If it still does not run up to par, then and only then can I tell anyone whether or not any of the hardware I run is indeed causing some weird serious bottleneck.

I will do so at that point, to let people know... just don't know when that's going to be.

Thanks to everyone who tried to figure out this debacle.
Click to expand...

Just would like to point out that games DO use quad cores. There's plenty of quad aware titles out there now and there will be more in the future (eg Left 4 Dead, Far Cry 2, Arma 2, Lost Planet, Resident Evil 5, GTA IV). If you are a gamer it makes no sense to go dual at this point. Sure they will suffice for now but not when games start to get more computational heavy, which probably won't be for another couple years at least. Still I wouldn't recommend a dual to anyone buying a PC today.
 
punkrockpolak said:
Those good 'ole Win98 days. I was running a 400 Mhz AMD K6 and some hardcore integrated graphics at the time lols.
Click to expand...

lol, yeah those days when some people swore up and down that they would switch to Linux when XP was introduced
 
RamonGTP said:
People were using Windows 98 prior to XP
Click to expand...

Was just using 3.1 as a starting point for around 92-93 on the changes between those years to the 2001 release of xp. Had 3.1, 95, 98, SE, 2000, Me for the unfortunate. I remember complaining about some things during the early life of xp. I think my mom actually tried to reformat a fairly new computer to 98 and got confused in the process since NTFS was foreign to her at the time. Not sure what all was up, I was busy playing nfl2k1 online via 56k with my dreamcast ^_^
 
Last edited:
If XP has no gains with this card I would be interested to see the difference between Vista and Win7 given the same hardware.
 
It's probably a driver issue. ATI doesn't have the latest support out for the 5870 and XP yet. don't know when they will either. I believe that MSI released a 9.10 beta driver and there's an XP version. Try that and see if it helps. Otherwise there's probably a bottleneck with the OS/RAM/or CPU speed.

Here's the link for the driver. They do have one for XP.
http://www.rage3d.com/board/showthread.php?t=33952980
 
First of all, 2GB DDR2 is a bottleneck. Technically, your CPU is a bottleneck too.

That does not explain the performance deficiency of the 5870 here, however. I suspect that is attributable too (1) some control panel setting difference (are the settings REALLY COMPARABLE?) (2) driver issues, (3) lack of optimization.

The partnerships that NVIDIA has with developers are in part why I stick with them.
 
Don't worry, I'm running Windows 7 Ultimate x64 with 8 gigs of memory and a quad clocked at 3.6ghz...seems like the perfect system to bench this on. Now just go ahead and send me that 5870 and we'll get this matter all straightened out :D
 
So, after reading this thread, we need to have 7 in order to benefit from 5870?

Are there any others who run 5870 and xp and have the same problem?
 
i don't get why everyone is claiming his cpu / ram bottlenecking is causing the results. if it was a bottleneck in the cpu/ram, shouldn't the results be the same for both cards?

my vote is for drivers.
 
stainremover said:
i don't get why everyone is claiming his cpu / ram bottlenecking is causing the results. if it was a bottleneck in the cpu/ram, shouldn't the results be the same for both cards?

my vote is for drivers.
Click to expand...

Might be a driver issue but both cards would perform so much better at 4Ghz. It's an E8600 for christ sake. Take that bitch to 4.5Ghz :)
 
I'm just going to go out on a limb and say that the early 5870/eyefinity reviews were probably cherry picked by giving the earliest cards/monitors to review writers who have a historical pattern of representing ATI products in a favorable way.
 
gravityparade said:
I'm just going to go out on a limb and say that the early 5870/eyefinity reviews were probably cherry picked by giving the earliest cards/monitors to review writers who have a historical pattern of representing ATI products in a favorable way.
Click to expand...

So this means his 5870 is suppose to perform worse than his GTX-280. Problem solved everyone, time to go home.
 
gravityparade said:
I'm just going to go out on a limb and say that the early 5870/eyefinity reviews were probably cherry picked by giving the earliest cards/monitors to review writers who have a historical pattern of representing ATI products in a favorable way.
Click to expand...

That's a mighty unstable limb. My results mimick the reviews pretty accuratly. If you use a powerfull card like the 5870 in a system significanly inferior in every way to the test system used in reviews, you're going to get worse results.
 
Tbird87 said:
So this means his 5870 is suppose to perform worse than his GTX-280. Problem solved everyone, time to go home.
Click to expand...

I agree with your sarcastic comment. The 5870 should perform better than a GTX280 considering the praise it's been given by the small niche of the enthusiast community that's had the opportunity to try it -- but lets also consider that, as per usual, ATI hardware is being released before the drivers are complete for all platforms, which is evident by the failure of the 5870 on the XP i86 and XP x64 platforms.

Let's also consider that the real life benches of fully optimized hardware configurations are probably not going to achieve the romance of some of the earlier reviews with the 5870 and the eyefinity, for the reason I stated. I think people are being set up to be dissappointed when they find out that the performance of the 5870 will actually be pretty close, if not marginally better than the GTX280.
 
RamonGTP said:
That's a mighty unstable limb. My results mimick the reviews pretty accuratly. If you use a powerfull card like the 5870 in a system significanly inferior in every way to the test system used in reviews, you're going to get worse results.
Click to expand...

Consider the implications of what you're saying according to the OP's dilemma. If you use a system with inferior hardware, you're destined to get worse performance than if you used an older generation graphics card. To me, that excuse is not going to cut the mustard.

The hardware for the 5870 may be top of the line, but if CCC is not tuned for your hardware and the game profile optimizations for the particular game you're playing aren't ready in CCC for that hardware configuration, then the product is a lesser value any way you slice it. CCC has a historical pattern of sucking eggs. Is AMD's strategy to release this hardware before the software is ready, so they can get the jump on potential sales? If that's the case then I'm a little frightened by their marketing strategy trumping QA. Maybe I buy my hardware from the wrong company.
 
This is why I'm skeptical of buying a 5870. I don't know who is benchmarking what on other sites, but real world, it doesn't seem like such a big step up. But, op, even if your 280 benchies look low.
 
Well the 4870 x2 is faster than the 280 ,and the 5870 is faster than the 4870x2 so , maybe its just a driver glitch at the moment
 
I guess no one takes into account Nvidia and their driver optimizations?

If any of the Nvidia control panel optimizations are in play, you will DEFINITELY see differences...even in advantage of the Nvidia card easily.
When I did a test run of Age of Conan on a GTX 260 versus a 4890 for several articles I did..I had a major discrepancy and got BETTER marks in certain places on the Nvidia card...yet, other zones the 260 died on compared to the 4890.

Anyways, when I did the test runs of the 5870 vs the 4890 in a fair comparison...we are talking some places got between 80-100FPS better performance on the 5870.

Anyways, I would like to know what all the settings are for the Nvidia Control Panel and what the ATI CCC settings are before going further on this issue.
 
gravityparade said:
I agree with your sarcastic comment. The 5870 should perform better than a GTX280 considering the praise it's been given by the small niche of the enthusiast community that's had the opportunity to try it -- but lets also consider that, as per usual, ATI hardware is being released before the drivers are complete for all platforms, which is evident by the failure of the 5870 on the XP i86 and XP x64 platforms.

Let's also consider that the real life benches of fully optimized hardware configurations are probably not going to achieve the romance of some of the earlier reviews with the 5870 and the eyefinity, for the reason I stated. I think people are being set up to be dissappointed when they find out that the performance of the 5870 will actually be pretty close, if not marginally better than the GTX280.
Click to expand...

Ah, I just thought that the driver optimizations for XP was already a given on his poor performance. So real world XP performance would be a better term than just saying real world in general along with them just sending it to sites that favor ATI. I say if you are buying a 5870 at launch, you should be looking at a OS upgrade in the near future IMO. I have an extremely small budget to play around with my pc toys, but planning on buying Windows 7 at launch. RC1 has been great to me. As far as real world feel difference running on Windows 7 between a 280 and 5870, I can't really say myself. My current system is the fastest I've used but ALOT of people on here would commit suicide if they had to use it for gaming. Luckily I only play source games.
 
Last edited:
You must log in or register to reply here.
Back
Top