I had posted another thread on the problem I was having with the video cards I had selected but got no responses. I have fixed the issue and I wanted to share what the problem was.

I have never been so clear as to what brand to select than now.

I have the following components on my media center that I was upgrading:

Gigabyte GA 945GCM-S2C motherboard
PD 3.2 Presler cpu
2G Patriot 667mhz memory
Ultra 500 watt psu
Media Center 2005 OS
Thermaltake Lanbox Lite Case

I had an older motherboard that was an AGP setup and wanted to change over to PCI express for a faster card. I started out with a PNY 9600GT and was disappointed from the 3dmark06 scores I was getting then I took it back and picked up a Visiontek 3870 and couldnt get 3dmark to even start and I was getting horizontal lines when I played movies. Then I took back that card and picked up the one I started to get in the first place, the EVGA 9600GT.

Here are the 3dmark06 scores with the same components other than the video card:

PNY9600GT - 4750
Visiontek 3870 - would not load 3dmark06
EVGA 9600GT - 6715

Brand selection is obviously very important. I would have to speculate that other brands would get different results as well. I was limited to what brands to select from at Microcenter due to all of their ATI cards are Visiontek. DONT BUY THESE. At least not the 3870. I may have gotten a bad apple out of the bunch but I didnt want to drive back and return video cards for the next week to find out.

This is the final kicker for me. The PNY and EVGA had the EXACT same specs. I had the sales rep to look up both of the cards and compare when I first purchased the PNY. The PNY was 20.00 cheaper but obviously not worth it compared to the EVGA.

My old AGP setup with an ATI brand X850XT scored 2000 with the same cpu and memory.

My CPU scores were differential between the two cards:

PNY - 987
EVGA - 1621

This proves that not only does the cpu let the video card shine as I think most of us believed in but the video card also lets the cpu shine. The proof is the scoring using the same cpu.

I purchased an E2180 and will later test the EVGA 9600GT and see if that improves the score to any considerable degree.

I may need some assistance getting it to OC over 2.4ghz on this motherboard. So if any of you overclocking gurus want to get involved the give me a pm or reply here and we will set a time.

I thought this was a very useful bit of information and wanted to share what I learned with this...

i can assure you that you are so wrong i have no idea why you would post this thread.

Many review sites have done comparisons before and found little to no difference.

I would think it is your comp and your testing methods that are the problem, no wait, i do not think i KNOW.

I am sorry to burst your bubble, and try to put this as graciously as possible.

It is quite apparent that hardware upgrades are not your area of expertise.

trying to be polite as well, personally i think he should be flamed to hell for thinking he alone can refute years of independent testing by hardware sites and there respective forum members.

I am refraining due to me being far too harsh of late.

Yeah, you are just dying to get that Deadliest Catch video game, right?

damn right i am, my enthusiasm shines through whenever i think about trawling through the sea instead of trolling hardware sites

Well brand dose matter to a certain extent, but mainly in quality and not scores*, after all PNY and EVGA dose use the same GPU.

*It's also worth noting that if the EVGA was factory OCed it could result in the higher score. Different RAM speeds/size could have also effected the score, along with different drivers and/or settings.

While it is possible the card had some sort of doa issue, odds are it wasn't simply a brand issue, perhaps poor quality control.

I am not sure if you know this, but the card are universal, nvidia has a great plan in place and that is, sell the chips and design and third parties build them market them and handle any issues.

I know I had a similiar issue with xfx and the 7900gt when those chips were having issues. I bought an evga card and had no more issues. But that is most like only due to a different overclock amount, quality of ram or something else, not the manufactuere or chipset itself.

Just my two cents.

Did you use most recent drivers from the website? Did you make sure everything was identical?

Ok guys. Take it easy.

I simply took out one card and installed another and I did get the latest drivers for all of the cards from ati and nvidia.

Idential hardware environment. Nothing changed but the video cards and the drivers. Both video cards were stock, identical speeds with same memory size. Identical spec cards.

What else can I say?

These are the results that I got plainly as listed.

As I said about the ATI card if you read it carefully, I explained that maybe I got a bad apple. I have a hard time thinking all Visiontek or all ATI 3870's are bad.

This is not a biased posting and you cant say this is from a fanboy of either brand. I didnt lean one way or another and I currently own both brands of cards and buy what suits my needs for the price, other than my main rig that was strickly for performance regardless of price at the time.

You cant really argue with the facts here other than maybe I got bad cards from PNY and Visiontek.

The PNY and EVGA comparrison doesnt get any plainer or closer in comparrison than this. Go buy these two cards and try it yourself. Sure there could have been something wrong with the PNY card. It would seem as if there was something wrong to benchmark so differently, thats why I took it back.

I guess maybe I got 2 bad cards out of 3. That would be the only thing that I would be able to say about that. I am no beginner at building pcs. I have been doing this for myself and others for 7 years now.

Also, how much expertise does it take to install a video card, install the latest drivers and run a benchmark program?

I uninstalled the previous drivers before installing drivers for the next card. Everything was done properly I can assure you.

Im surprised you guys are so ready to flame before asking questions and gathering facts. I guess this is what an e-linch mob looks like..


Im not aiming my statements at you guys, shadow and cranbers. You responded as I would expect and its appreciated...

He does have one thing right, manufacture brand does matter. It really stands out in PSU selection and Memory these days. As well as motherboard manufacturing as well. There are a few low budget motherboard makers out there that your really taking a gamble with, this really started showing itself when motherboard manufacturing moved from
Tiawan to the mainland. But our OP has to remember that even the best manufacturers let a bum unit slip out from time to time. But its there support that makes them stand out as best in there fields.

ESD when cards were being swapped, perhps?http://www.esda.org/basics/part1.cfm

The facts are that your one card samplings will not result in a statistical model from which you can draw any reliable conclusion. The number of review sites that disagree with your findings is compelling.

As a software engineer, I sometimes gauge the professional maturity of other engineers by their willingness/tendancy to first look deeply to their own work when things go wrong before blaming someone else's code.

Spongebob,

I understand one sampling doesnt make the standard rule of thumb. But I am only suggesting that you can buy a card and it function but not as you expect.

What other way will you know other than comparing?

Lol, good post. Some people blame everyone but themselves.

As for the "difference," keep in mind Nvidia makes the gpus, then sell
them to different companies listed. Those "manufacturers" don't
actually "manufacture." It's identical gpu on identical board layout
required by Nvidia. It's a matter of pasting on labels.