Can't get 8600M GT to perform

I tried using forceware drivers on my new Inspiron e1720 and I just can't get the scores for the 8600M I see posted on other sites. They were boasting 6500 and higher with factory settings in 3DMark05 and I was getting a dismal 5333 with factory defaults. What can I do to speed things up besides driver updates?

 

1. they probably are using a lower rez
2. they probably have a laptop like the g2s which has ddr3 instead of ddr2 like your dell.
Check these two things and see if they account for the higher scores

 

how much ram did they use compared to you?
what kind of hard drive, and CPU did the other sites used compared to you?

operating system?


if you match all your hardware and software up to theirs im sure you'll have very similar score.

 

What's the factory clock setting for an 8600GT in the new Dells? It's possible they're underclocking it versus ASUS.

 

I suggest overclocking your video card / cpu / memory. Getting a higher score on a synthetic benchmark test is by far more important then your computer's functionality.

 

Drivers are very important. THe best ones are probably the 158.45 from the laptopvideo2go site.

 

would those work for my 1520?

 

They should work. If not just roll back the driver.
After downloading the correct drivers, read this on how to install them using the modded INF.
http://www.laptopvideo2go.com/forum/index.php?showtopic=11997

 

Alright, so I've done some tweaking to my system and I'm now able to run Oblivion respectably by following some tweaking guides.

As for the system compared to, I just googled the 8600M GT benchmarks and looked over several pages from different forums and saw this: http://www.notebookcheck.net/NVIDIA-GeForce-8600M-GT.3986.0.html

I would have assumed that the new Santa Rosa chipset with 2gb of RAM would have performed about as well.

But here's the thing: after all that tweaking after reading the entire TweakGuides.com Tweaking Companion manual for Vista, I'm not exactly sure which settings might have caused my performance to plummet any more. While games run undeniably smoother than before tweaking, 3DMark05 now clocks in at a dismal 3200 - over 2000 less than my original scorings. My drive is absolutely defragged, spyware/virus free, and I have very little running in the background.

I tried allocation 4096MB to pagefile instead of 1024, and I also reverted back to the original drivers all to no avail. So this prompts two questions:

1. Is 3DMark05 an accurate meter for my performance? I know people can get screwy results, but losing 2000 points from my own initial reports with the exact same set-up is kind of upsetting.
2. Is there a certain setting that'll completely debilitate a machine when tweaking HDDs?

 

Your 3dmark05 score is accurate for Dell's incarnation of the 8600m GT. You can overclock it to regain some lost performance, though. I've gotten mine to score over 7200 in 3dmark05.

Er, this post was typed on the assumption of ~5500 for 3dmark05 stock, not your latest 3200.

 

The site you linked appears to use an 8600GT with DDR3. The 8600GT used in the new Dells is DDR2. Hence your original 3D mark score was lower.

 

Any clue why I get the terrible score compared to my own system though? Is there a way to restore to factory defaults on a system WITHOUT having to format? I have all of my PC games installed on this thing and updated... which was a 2-day affair and I leave for New Mexico tomorrow for two weeks.

 

Well, I installed the 158.45 drivers for Vista 32-Bit on my 1520 and I am going to test them now.

 

Does the 3dmark score really matter to you that much? It isn't the end of the world if you don't get that fixed before your trip, is it? I'm guessing that one of the tweaks disabled something with the drivers that the card makers use to "cheat" with 3dmark.

 

I think I found the problem.

I enabled the "Performance" setting under HDD options in the BIOS and apparently it makes the drive not only louder, but half-speed as well.

 

It isn't the 3DMark score that matters to me, but it's one of the few "reliable" benchmarks out there. I'm just trying to get the absolute most bang for my buck out of my purchase.

I'm going to be staying with my cousins in New Mexico, which means there's going to be a hell of a lot of gaming and LAN parties, and I want my system to perform properly. After fixing it, my framerates in Oblivion went from 25 to 45.

 

3DMark05 is not an accurate tool for gaging real life gaming performance, especially with the newer DirectX 10 video cards. However it's a good indicator of whether or not your system is performing as well as machines with the same video card. If you are getting 2000 points lower than someone else with similar specifications and the same video card, then there is something wrong.

What kind of power settings are you using? If you are using Vista, change it to "High Performance" (click the battery/power icon in the system tray and change it).

 

You're missing my point. A great 3dmark score doesn't always translate to great in-game performance. In short it isn't a good benchmark at all.

EDIT: not you Chaz

 

I know, but it's still a great diagnostics tool.

 

True. As Chaz said you can use it to compare your 1720 to other 1720s. Beyond that it isn't worth much. I don't use it myself.