*update* new rig, alarmingly bad 3dmark06 score

[james1701]

Did you try resetting your Bios? I dont know if you have a reset button, or you need to cycle the power 3 times or so. That should clear you up. Just right down all your settings first. That way it will be easy to fix it back when you are done.

 

[jeffjones66]

Thanks for all the replies.

Woodbutcher, no, I hadn't tried reseating the vid card yet.

Taltamir and Louisss, yes, I did a clean install of Windows. The HD I used is new and had nothing on it.

James, yes, I did a bios reset. One of the voltage configurations I tried wouldn't post, so I was forced to

Now get this ... I woke up this morning, booted it up, and lo and behold the Link Width is now x16! I'm not going to get too excited, it could fall back to x1 on the next boot, but still, it's something. I'm going to re-run 3dmark now and see what happens.

 

[swtethan]

You probably needed a windows update.

 

[badnewcastle]

It's got to be the PCI-E showing 1x now 16x or something like that, I was getting 11,200+ with my system at stock, specs below... Now I'm getting 12,400+ with the OC listed below.

Oh also check your settings in 3dMark06, are you running a high resolution with AA and AF turned up? That will lower your score.

YOU DO NOT NEED A QUAD CORE TO GET ABOVE 10,000 MARKS~!

 

[jeffjones66]

old score with Link Width at x1: 4277

new score with Link Width at x16: 11,189

Not THAT is more like it! :thumbsup:

 

[chizow]

Nice! Would be good to post how you fixed your PCI-E setting problem in case someone else stumbled across the same problem.

 

[jeffjones66]

Originally posted by: chizow
Nice! Would be good to post how you fixed your PCI-E setting problem in case someone else stumbled across the same problem.

I would love to ... if I only knew! :laugh:

If anything, I set my BIOS stuff pretty much back to standard before giving up last night. I woke up this morning, and booted up, and voila ... 16x. I just tested Crysis, and the difference is amazing. VERY fluid at 1680x1050, all settings on High.

I'm very happy right now, but I just hope this sticks till my new PSU arrives. My instincts tell me that my underpowered 420w PSU might be the culprit.

 

[stapuft]

Originally posted by: jeffjones66

Originally posted by: chizow
Nice! Would be good to post how you fixed your PCI-E setting problem in case someone else stumbled across the same problem.

I would love to ... if I only knew! :laugh:

If anything, I set my BIOS stuff pretty much back to standard before giving up last night. I woke up this morning, and booted up, and voila ... 16x. I just tested Crysis, and the difference is amazing. VERY fluid at 1680x1050, all settings on High.

I'm very happy right now, but I just hope this sticks till my new PSU arrives. My instincts tell me that my underpowered 420w PSU might be the culprit.

Your current PSU is 18A on the 12V, single rail, correct? Not enough. Your PSU may be overloading and the good nights rest and cool-down helped you recover some Amps. If that's the case, it will happen again. If you have multiple HDD's and optical drives you can reduce power consumption in the short term by unplugging the nonessentials. Remove any CPU/GPU overclock (and downclock any Factory OC on the GPU). Open the side of your case and unplug any nonessential 12V case fans. Use a traditional fan blowing in if you're worried about that 8800 egg fryer.

BTW, I just pulled my Thermaltake 420W because it was being drawn below spec at load using the following setup (prior to my 8800GT install):

Athlon X2 4200+ OC to 2.6 (HTT 236)
Biostar Tforce6100-939 board, all devices enabled
2x1G 2.6V DDR400 (set to 166Mhz, HTT OC to 196MHz)
6800GS PCIe
1 HDD
1 DVD
2 case fans


I installed a CoolerMaster 650W (3 rail) from Newegg and the new EVGA 8800GT 650/1900, 3DMark score is now ~9500.

 

[jeffjones66]

Stapuft, yes, it's 18A on the 12v. Thanks a lot for the suggestions. You reminded me to disable my second HD, and a few case fans, at least until the new PSU arrives. Came home from work for lunch, and booted up and the Link Width is still at 16x. Hopefully it'll hold up till I get the new one.

 

[chonga]

I am having this same issue. Has anyone found the root cause?

I have a gigabyte ds3r with 8800GT, e6750, 600w gameXstream PSU. Link width is stuck at x1 in CPU-Z. I have tried re-seating the card, i have re-installed windows, tried several different nvidia drivers, and i've flashed my motherboard BIOS to the latest.

Also this is kind of weird. I tried my roommates 8600GT in my rig and it worked perfect, registered at x16. So i try my 8800GT in his rig and he only gets x8 instead of x16. Not as bad as my x1, but still not optimal.

I'm wondering if there is some sort of compatibility issue with my motherboard and this 8800GT or if it is the card and needs an RMA?

Brand is eVGA

 

[JustaGeek]

I bet it is the PCIe 2.0 vs. 1.1 vs. 1.0a issue...

 

[SLU Aequitas]

For the record, the GA-P35-DS3R is PCI-E 1.1 compliant; I don't know what the exact problem is with the PCIe 2.0/1.1/1.0a issues surrounding the 8800GT though.

Edit: I have the board myself, and have one of the Dell MSI 8800GTs lined up...should I be expecting problems?

 

[Zap]

I've noticed that some BIOSes have an option for "PEG force 1X enable/disable." Perhaps that may be the culprit?

 

[myocardia]

Originally posted by: SLU Aequitas
For the record, the GA-P35-DS3R is PCI-E 1.1 compliant; I don't know what the exact problem is with the PCIe 2.0/1.1/1.0a issues surrounding the 8800GT though.

Edit: I have the board myself, and have one of the Dell MSI 8800GTs lined up...should I be expecting problems?

Nope, I have the GA-P35C-DS3R and an 8800GT, and it's performing rather well. Also, I didn't have to do anything but install it. Of course, my psu has 49A continuous on the 12v rail, so it's definitely not being underpowered.

 

[chonga]

I think I may RMA the card that way i'll know if it is the mobo or card at fault. If it is the mobo, i may have to buy a new one as i would like to keep the 8800GT.

It has to be something with the pci-e 2.0 thing thought because my roommates 8600GT works fine at x16 on my board as well as a 7900gs in my board.

Unless i got a lemon, this must be the initial pains with pci-e 2.0. sigh.

 

[Munchiselleh]

Well, I've had the exact same problem...with an 8800gt SSC. I have a 375w psu, which I think might be the problem. Or else it's a bad card or it's my mobo. If anyone could supply me with confirmation on which piece of hardware is at fault, it would be great, because I don't want to waste time or money ;]

 

[chonga]

If you wait a week I can tell you. I'm RMA'ing my 8800GT tomorow so I should get the replacement by middle to later next week.

I'm guessing hardware isn't faulty but rather there may be some sort of compatibility issue with these newer pcie 2.0 cards.

I'll keep everyone updated.

 

[Munchiselleh]

alright sweet, I hope this issue is fixable via a vga bios that they're developing or something easy...

 

[jeffjones66]

As an update, my new PSU hasn't arrived yet, and I'm still having the problem off & on. What's odd is that it booted about 12 straight times to 16x, then lately booted to 1x about 7 straight times (regardless of case temp, how long the psu was off, etc).

From the sound of it, my new PSU isn't going to solve the problem, since one of you stated that you had the same hardware except with a 600w OCZ, and still had this issue.

If I'm understanding the most recent posts, the 8800GT is PCI-e 2.0, and the DS3L isn't? Can someone recommend a mobo that is fully 2.0 compliant, or at least give me a combo that I can count on to have none of these issues.

edit: this board looks like a winner:
http://www.newegg.com/Product/...x?Item=N82E16813127030

Any comments?

 

[nullpointerus]

Still, IMO, the problem is caused by your PSU.

I doubt that a configuration or compatibility problem would exhibit randomly working/broken changes based solely on what happens when the power button is applied.

IIRC, PCIe 2.0 motherboards are relatively rare and expensive at the moment. The P35 chipset (and the board you linked to) does not have PCIe 2.0 slots. So nearly everyone who is running these new PCIe 2.0 video cards is running them in a PCIe 1.x compatibility mode of some kind (although I do not believe there is any performance loss from that).

My advice is to try the new PSU. You don't have to gut and rewire your case. Just sit it on your desk, connect the power cables as needed, and then boot a few times to verify the PCIe link speed is fine.

A faulty motherboard or video card could conceivably cause this problem, but IMO the PSU should first be eliminated as a factor.

 

[jeffjones66]

Null, I agree, and definitely plan to try the new PSU first. I really hope you're right, because I don't look forward to replacing the motherboard.

 

[chonga]

I would agree except my 600 watt SLI OCZ GameXstream is having the same link width issue and this PSU is more than ample to supply power to a single 8800GT. I still think it is a motherboard/8800gt compatibility issue with pcie 2.0/1.1/1.0a

I will post results once I get my RMA from eVGA..... just shipped mine today 2 day shipping so I should get the new card by the end of next week. If it is the motherboard, I may just have to purchase a new one or RMA that as well. I've read some people having no problems with the Gigabyte p35-ds3r and the 8800GT, so it's still a mystery.

I just want to enjoy Crysis and WiC !!!! I've been waiting weeks!

At least I will have a definitive answer for everyone later in the week.

 

[jeffjones66]

Chonga, i hope we can figure this thing out. I ordered an OCZ GameXstream 700w, and I'm hopeful that will do it, but I suppose it could be the motherboard/card combo. The next step for me will be to buy a new mobo (different brand/model) and try that, and if that fixes it I'll RMA the DS3L. Maybe together we can solve this issue. Seems others have had fixes by adjusting various voltages, I had no such luck.

 

[Munchiselleh]

I don't know about you guys, but I'm absolutely 100% it's my card, and I'm RMA'ing monday. Is there anyway I can RMA on the weekends? I tried the card in five different computers, three of which had 700+w PSU's, and many different mobos. It's the card. At least for me.

 

[chonga]

Ok, so I just got a reply back from Gigabyte stating that they have tested the 8800GT with their DS3 line of boards with no problems. They believe the issue lies within a certain BIOS that the 8800GT may ship with. They have a few inquiries about it but had no problems themselves testing the card with their boards...... My 8800GT RMA should be getting back to me later in the week. And just to make sure my board wasn't crappin out, I tested my roommates 8600GT again and it ran perfectly at x16.

So if this new card is still operating at x1, i'm going to RMA the board and if the new board is still not cutting it, guess that means i'll have to buy a new board. At least the newer p35 boards are pretty solid and cheap.

Also, read this in another forum:

"Some makers cards only work with a 2.0 pci-e bus (using stupid nVidia bios). Other companies modded their bios so it works properly with pci-e 1. You can flash to one of those. (Cards are all nVidia reference)
There have been a lot of silly compatibility problems with that card as well, as noted."