bottleneck question

[Lopyswine]

so i just built a new amd 64 box.

i got a 2800 with a pretty good mobo. 512mb of kingston ram. a new 160gb HD etc...

however i have a geforce 3 ti200 video card...

and some of the new high end games are running choppy...

i am pretty much sure that my video card is what is bottlenecking my system...

do you guys agree?

if you think it is the video card than what card do you recommend for between 150-200$

 

[Acanthus]

if you have an Nforce 3 250GB based motherboard, NVIDIA GeForce FX5900SE/XT/Non Ultra for $190ish

if you have any other chipset, ATi Radeon 9800 128MB for $200ish

 

[Insomniak]

UH...yeah....definitely that GF3 holding the system back. Try looking into a 5900NU/9800NP, as already mentioned.

 

[modedepe]

3 year old video card paired with a brand new cpu? I'll say without question that that gf3 is a bottleneck. I'd go for a 9800 pro.

 

[Lopyswine]

so what is this NVIDIA GeForce FX5900SE/XT/Non Ultra for $190ish?

cause my board is a nvidia based board.

but yeah otherwise i am thinking 9800 pro

 

[Cawchy87]

nvidia chipsets don't favor one type of vid card over another.

 

[Acanthus]

Originally posted by: Cawchy87
nvidia chipsets don't favor one type of vid card over another.

Theres a proprietary performance increase for NVIDIA cards on NForce 3 chipsets.

 

[Acanthus]

Gigabyte

BFG

PNY

Albatron

Gainward 128MB

Gainward 256MB

Leadtek (usually highest IQ and best overclocker for NV)

XFX

 

[VIAN]

I'd still go with a 9800 Pro, the 5900 XT is only compareable to the 9800, not the Pro, and the 9800 is the winner in that comparison.

 

[Acanthus]

Originally posted by: VIAN
I'd still go with a 9800 Pro, the 5900 XT is only compareable to the 9800, not the Pro, and the 9800 is the winner in that comparison.

Weeee hi vian.

Got a link to 5900XT/SE vs 9800p on Nforce 3 250? Of course you dont, youve never even seen a chart im sure, but you're ready to state your OPINION.

Link

 

[Pete]

You see a gaming performance gain worth noting in that review?

 

[Digic]

if you get a 9800 beware of cheap 128 bit versions

 

[Acanthus]

Originally posted by: Pete
You see a gaming performance gain worth noting in that review?

Gaming performance - Halo - ATi (difference less than 2fps)

9800XT - 60.7
5950 - 59.0
-----------------------------------------

Splinter Cell - ATi (difference less than 2fps)

9800XT - 50.48
5950 - 49.17
-----------------------------------------

X2 - NVIDIA (Clear win 13.7fps)

9800XT - 116.0
5950 - 129.7
-----------------------------------------

Aquamark 3 - NVIDIA (Clear win ~ 4800 points)

9800XT - 47209
5950 - 52094
-----------------------------------------

Commanche 4 - Tie

9800XT - 72.45 (on SiS 755, Nforce 250GB was slightly slower at 70.72)
5950U - 72.45
-----------------------------------------

Quake 3 Arena - Tie (20fps difference, but fps is so high ill count it as a tie)

9800XT - 446.4
5950U - 466.0
-----------------------------------------

UT2003 Flyby - Tie (again 13fps difference, but fps is so high ill count it as a tie)

9800XT - 290.2
5950U - 303.4
-----------------------------------------

UT2003 Botmatch - Tie (within margain of error)

9800XT - 114.0
5950U - 113.4
-----------------------------------------

SPECViewperf 7.1 - 3DSMAX - NVIDIA

9800XT - 13.09
5950U - 16.81
-----------------------------------------

SPECViewperf 7.1 - DRV - Clear NVIDIA Win (over 100% faster)

9800XT - 40.45 (SiS 755, Nforce 3 250 was 37.55)
5950U - 88.08
-----------------------------------------

SPECViewperf 7.1 - DX - NVIDIA

9800XT - 77.53
5950U - 87.69
-----------------------------------------

SPECViewperf 7.1 - LIGHT- NVIDIA

9800XT - 14.56
5950U - 17.57
-----------------------------------------

SPECViewperf 7.1 - ProE - Clear NVIDIA Win

9800XT - 13.64
5950U - 22.73
-----------------------------------------

SPECViewperf 7.1 - UGS - Clear ATi Win (~300% faster)

9800XT - 28.22
5950U - 9.65
-----------------------------------------

NVIDIA wins by a lot - 4
NVIDIA wins by a small margain - 3
Ties - 4
ATi wins by a small magain - 2
ATi wins by a lot - 1
-----------------------------------------

Are the charts that hard to read? Its clear that the 5950 trades blows with the 9800XT with the optimizations, i really wish the review showed 5950U on 755. Couple that with better driver stability, and i think its clear that its down to personal preference for him on his platform.

I am not saying the 5900 is a better card, nor am i saying the R9800 is completely unstable in any way, im saying that ON THAT CHIPSET the 5900 is just as good a choice as the R9800, especially since its $20 cheaper.

Edit: this is also a decent review because the FX5900XT/SEs and 9800P often OC easily to the speeds of the cards in this review.

 

[Cerb]

Leadtek's probably won't be the best overclocker. The IQ of mine is fantastic, but I don't seem to be in a minority with the GPU maxing out around 450MHz w/ normal cooling.
Gainward's and Creatives have shown excellent overclockability, as XFX, and I'd trust Gainward to have IQ very close to Leadtek.

A 9800 Pro won't serve you badly, but for 2d, my Leadtek looks better than any BBA I've seen. AA does look better on ATI cards, though.

 

[Acanthus]

Originally posted by: Cerb
Leadtek's probably won't be the best overclocker. The IQ of mine is fantastic, but I don't seem to be in a minority with the GPU maxing out around 450MHz w/ normal cooling.
Gainward's and Creatives have shown excellent overclockability, as XFX, and I'd trust Gainward to have IQ very close to Leadtek.

A 9800 Pro won't serve you badly, but for 2d, my Leadtek looks better than any BBA I've seen. AA does look better on ATI cards, though.

It is normal for FX5900 Vanilla/XT/SE to top out around FX5950 speeds, the memory wont OC as high as a 5950, but the latency is lower on the 5900, so the performance is very very similar even though the memory speeds are not the same.

The 5950s are OCing monsters though .Check out the Leadtek A380

Leadtek definately has a good cooling design, they use the same cooling across their entire line of high end cards, even the 6800 uses the same style of HSF.

 

[BFG10K]

Theres a proprietary performance increase for NVIDIA cards on NForce 3 chipsets.

Doubtful. Even if true a driver update from ATi should be able to resolve the negligable performance delta.

Link

What an inept set of benchmarks. How about some shader games like Far Cry or Call of Duty? Or do you really think he'll be running SPEC all day?

Its clear that the 5950 trades blows with the 9800XT with the optimizations,

I think it's clear that you're avoiding the big picture, either intentionally or because you don't know any better.

The R3xx absolutely kills the NV3x in shader intensive titles and to claim the cards are equal after two years' worth of evidence is quite simply ludicrous.

 

[Cerb]

Originally posted by: Acanthus

Originally posted by: Cerb
Leadtek's probably won't be the best overclocker. The IQ of mine is fantastic, but I don't seem to be in a minority with the GPU maxing out around 450MHz w/ normal cooling.
Gainward's and Creatives have shown excellent overclockability, as XFX, and I'd trust Gainward to have IQ very close to Leadtek.

A 9800 Pro won't serve you badly, but for 2d, my Leadtek looks better than any BBA I've seen. AA does look better on ATI cards, though.

It is normal for FX5900 Vanilla/XT/SE to top out around FX5950 speeds, the memory wont OC as high as a 5950, but the latency is lower on the 5900, so the performance is very very similar even though the memory speeds are not the same.

The 5950s are OCing monsters though .Check out the Leadtek A380

Leadtek definately has a good cooling design, they use the same cooling across their entire line of high end cards, even the 6800 uses the same style of HSF.

It's normal for them to top out at 5950U speeds with a modified BIOS. RAM will get there if you use a card with faster RAM. Most of the Leadteks come with 2.2ns. And yes, the stock cooler isn't too shabby.

 

[Acanthus]

Originally posted by: BFG10K

Theres a proprietary performance increase for NVIDIA cards on NForce 3 chipsets.

Doubtful. Even if true a driver update from ATi should be able to resolve the negligable performance delta.

Its a proprietary performance increase NVIDIA claims is due to optimizations between the way the chipset drivers and graphics drivers interact. They claim the increase is only for NVIDIA based graphics cards. This is not new information... its been around since Nforce 3 150, there were even multiple threads about "lawsuits" for anti-competitive behavior.. Until benchmarks later proved that the FX series got a boost, while the ATi cards did NOT lose performance.