4200 X2 bottlenecks a 8800GTX

[Greek]

hey guys,

just wondering i the above card would run fine on my processor, ir it currently just overclocked to about 2.475, so not a hughe difference to its stock levels.

i was thinkikng of either getting the 8800 only if it will do well with nforce 4 chipset and my cpu, otherwise a 360

what you guys think.

let me know

 

[Exceededgoku]

well my bro's 8800GTS gets bottlenecked by a 3800 x2 at 2.7Ghz... which is the level of which the bottleneck starts to get relieved, you really need about 2.8-3.0ghz to have no bottleneck...

 

[Wile E]

Get the 8800, then overclock your cpu a little more.

 

[demonbrawn]

Hahah top of the line Conroes have shown to bottleneck the 8800. But, who cares man it's an 8800! If I could afford one, I would get one.

 

[sno.lcn]

It's going to bottleneck maybe a good bit more than you're thinking. My 8800gts (OCed to 660/960) was giving me a 3dm05 score of 12103 with my 930 @ 4056mhz. I swapped it out for an e4400 (@ 3500mhz) and got my score up to 18367.

 

[Wile E]

It's going to bottleneck maybe a good bit more than you're thinking. My 8800gts (OCed to 660/960) was giving me a 3dm05 score of 12103 with my 930 @ 4056mhz. I swapped it out for an e4400 (@ 3500mhz) and got my score up to 18367.

A score of 12103 is still nothing to scoff at. In actual gameplay, in all but the highest resolutions, you're not gonna notice much of a difference.

 

[Exceededgoku]

A score of 12103 is still nothing to scoff at. In actual gameplay, in all but the highest resolutions, you're not gonna notice much of a difference.

no it's actually the lower resolutions that will suffer... the more settings you turn on the higher your FPS. Since it takes more work off the CPU.

 

[theonetruewill]

With processor, why not just the the 8800GTS 320MB, seems like a better deal and the bottleneck will be less of an issue. The 8800GTX will be pointless with its performance at this level as it will never be able to reach its true performance. Thus simple answer get the cheaper card thats more suited to your system.

 

[Greek]

mm well when i gt the 7900 gtx i had money for the 8800 gts but i waited to see what was coming out, so at the minute the gts is out of the picture because i only want to go for top of the line at the minute. i can overclock my cpu to about 2.7 safelly, u thinnk that will be ok?

 

[i_am_mustang_man]

it will only bottleneck it by a considerable amount when the resolution is low, such as at 1280x1024, but at 1920x1200 for example, the difference will be a few fps.

all this means is that at low res you will get 100 fps instead of 120
and at high res you will get 40 instead of 43.

get the card

 

[Greek]

rite i see, well i will be runnig at low res anyway, havent gt the money for a monitor upgrade.

 

[Exceededgoku]

it will only bottleneck it by a considerable amount when the resolution is low, such as at 1280x1024, but at 1920x1200 for example, the difference will be a few fps.

all this means is that at low res you will get 100 fps instead of 120
and at high res you will get 40 instead of 43.

get the card

there is a bigger difference...

 

[Wile E]

there is a bigger difference...

It's still not like the games will be unplayable. But, don't take offense to this, I'd like to see proof that the falloff will be bigger than mentioned above. You have me doubting myself. lol

 

[Chewy]

Hahah top of the line Conroes have shown to bottleneck the 8800. But, who cares man it's an 8800! If I could afford one, I would get one.

I dont really see where top of the line c2d would be bottlenecked.. a e6300 @3.2 is a little bottlenecked by an 8800gts from what someone who tested told me. TKPenalty I think it was.

I think an 6700 around stock would prob be enough for a gtx.

 

[Exceededgoku]

I can provide you with test scores me and my bro ran with his XFX 8800GTS 320MB:

My system has an E6600@3.2Ghz with 2GB RAM in 3dmark06 I score:
3DMark Score 10116 3DMarks
SM2.0 Score 4544
HDR/SM3.0 Score 4341
CPU Score 2688

On my bro's comp which has the exact same graphics card but an A64 X2 3800@ [edit]2.6Ghz[/edit] and 2GB ram he scores:
3DMark Score 8398 3DMarks
SM2.0 Score 3830
HDR/SM3.0 Score 4078
CPU Score 1813

when his CPU is at stock (2Ghz) he scores:
3DMark Score 6865 3DMarks
SM2.0 Score 3182
HDR/SM3.0 Score 3334
CPU Score 1452

All are running on 165.01 beta drivers on Vista x64. The 8800GTS 320MB is overclocked to 590/1400/950. 3dmark06 1.1.0 is used.
Hope this helps....

 

[Wile E]

What about in game performance? Has your brother's proved unplayable at any point?

 

[WarEagleAU]

I would say to not get the 8800GTX unless you just really want it. The 8800 GTS 320mb is the sweet spot right now.

 

[Exceededgoku]

What about in game performance? Has your brother's proved unplayable at any point?

well since he only really plays CSS then I can say yes... he plays at a low resolution sa well which doesn't help (1280x1024). Irrespective of settings he gets the same FPS even when 16x CSAA is applied.... from memory he hovers around the 50-60fps (on Dust2) mark and this fps decreases at lower resolutions, at the same setting that I run on my X1900XTX which is all highest and 6xMSAA he gets around 55-60fps whilst I get around 150+ on teh same map... Note, however, that CSS is very CPU limited. In games like Test Drive Unlimited the difference between graphics power is obvious, his PC is getting at least twice as many frames as me.
One thing you should know is that for some reason (and it was reproduceable on my system and his) is that there were hiccups every now and again, sort of pauses... Googling "8800GTS 320 pauses" finds quite a few posts. So if anything don't get the 320MB version..

 

[i_am_mustang_man]

let's not do canned benchmarks, mmk? 3D06 relies heavily on cpu for all the tests, not just the cpu tests.

look at these results showing 1280x1024 for 320 gts, 640gts, and gtx

http://www.xbitlabs.com/articles/video/display/msi8800gts-640_9.html

this bout sums this up

I can provide you with test scores me and my bro ran with his XFX 8800GTS 320MB:

My system has an E6600@3.2Ghz with 2GB RAM in 3dmark06 I score:
3DMark Score 10116 3DMarks
SM2.0 Score 4544
HDR/SM3.0 Score 4341
CPU Score 2688

On my bro's comp which has the exact same graphics card but an A64 X2 3800@ 2.7Ghz and 2GB ram he scores:
3DMark Score 8398 3DMarks
SM2.0 Score 3830
HDR/SM3.0 Score 4078
CPU Score 1813

when his CPU is at stock (2Ghz) he scores:
3DMark Score 6865 3DMarks
SM2.0 Score 3182
HDR/SM3.0 Score 3334
CPU Score 1452

All are running on 165.01 beta drivers on Vista x64. The 8800GTS 320MB is overclocked to 590/1400/950. 3dmark06 1.1.0 is used.
Hope this helps....

 

[Exceededgoku]

let's not do canned benchmarks, mmk? 3D06 relies heavily on cpu for all the tests, not just the cpu tests.

look at these results showing 1280x1024 for 320 gts, 640gts, and gtx

http://www.xbitlabs.com/articles/video/display/msi8800gts-640_9.html

this bout sums this up


http://www.xbitlabs.com/articles/video/display/msi8800gts-640_10.html

it keeps on the same throughout all the real games, so you can peruse that for yourself.

also, i'm more likely to trust a review site with a reputation than one person on a forum.

okay if you say so... I'm just saying what I've experienced. And I don't disagree that 3dmark06 isn't CPU dependent, but that is something that must be considered.

 

[AshenSugar]

well since he only really plays CSS then I can say yes... he plays at a low resolution sa well which doesn't help (1280x1024). Irrespective of settings he gets the same FPS even when 16x CSAA is applied.... from memory he hovers around the 50-60fps (on Dust2) mark and this fps decreases at lower resolutions, at the same setting that I run on my X1900XTX which is all highest and 6xMSAA he gets around 55-60fps whilst I get around 150+ on teh same map... Note, however, that CSS is very CPU limited. In games like Test Drive Unlimited the difference between graphics power is obvious, his PC is getting at least twice as many frames as me.
One thing you should know is that for some reason (and it was reproduceable on my system and his) is that there were hiccups every now and again, sort of pauses... Googling "8800GTS 320 pauses" finds quite a few posts. So if anything don't get the 320MB version..

thats an OLD nvidia problem, been around since the 6 seirse and they havent fully fixed it yet, some people get it others dont, just isnt as notable now as it use to be back in the day.

as to your bench results, alot of things can cause those diffrances in 3dmark scores, running prosesses, windows settings, program settings, driver settings.

run them off windows installs that are 1:1 installs, just windows+drivers+bench apps, no uniq software or settings, i have benched my buddys 4 systems and gotten drasticly diffrent scores in 3dmark, pcmark, and such b4 and after wipes/reinstalls and guess what, the top perf hits come from Norton/mcafee/avg or the like, one alot of people dont realise is that large hosts file can cause drastic perf hits in benching, no joke.

personaly i wouldnt buy any videocard right out, at least not till there are some dx10 games to test and we can see the compotition.

im happy with the card i have tho, x1900xtx@700/1600 is plenty of power

 

[Zubasa]

Norton/mcafee/avg apps like these can really kill a computer on 512MB ram running XP.

 

[AshenSugar]

zubasa, any ammount of ram, they are cpu,ram,hdd,full system resorce intencive, they can cause as much as 50% system wide perf loss in some cases.

compare them in use to nod32, where the most i have seen is a 1-3% perf hit in extencive testing on everything from a 233mmx 208mb ram system to a quad-dual core opteron system with 2gb ram per chip(used for rendering) and norton/mcafee and their like really really do hit perf baddly.

check out Alec's post about nod32 vs norton he compared them and posted results, and he was a very large norton fan!!!!(who has sence moved over to use nod32)

 

[Wile E]

zubasa, any ammount of ram, they are cpu,ram,hdd,full system resorce intencive, they can cause as much as 50% system wide perf loss in some cases.

compare them in use to nod32, where the most i have seen is a 1-3% perf hit in extencive testing on everything from a 233mmx 208mb ram system to a quad-dual core opteron system with 2gb ram per chip(used for rendering) and norton/mcafee and their like really really do hit perf baddly.

check out Alec's post about nod32 vs norton he compared them and posted results, and he was a very large norton fan!!!!(who has sence moved over to use nod32)

I was gonna post the same basic thing, Ash. lol. I also wanted to add that Kaspersky Internet Security is a nice middle ground with more features than NOD, but slightly higher resource usage.

 

[Crisao23]

Here is my Opteron 165 at 3006Mhz/ RAM at 273 Mhz and a 8800GTX on 3DMark 2006/Vista Ultimate x64: