View Full Version : New P35-DS4 rev2.1 + E8400 rig....my list of issues.

02-16-2008, 10:01 PM
Nearly all are already known on here, but thought I would throw them out there since a lot seem like BIOS issues thus the more known, the more likely an update will fix them.

I'll start with the CPU itself since it is unrelated to this forum. I got an excellent oc'er but one with the flakey temp sensing. I have played with it to the extremes to get the full scope of the temp problems....everything from undervolted stock with the HS fan blowing hard for the lowest possible temps, to oc'd and a little overvolted with the fan off to push near max temps. End result....undervolted to normal voltage, weather stock or modest oc, core temps are about 10 degrees off. Push a little more till the cores are into the 60's, they are about 5 off. Make the cpu run hot, right near the 72 cpu die temp, the core temps even out in the high 70's+ (using coretemp 0.96.1 for reference). Many other reports of similar situations on the net, so after re-seating the cooler many times and even trying a different one, I have come to the conclusion I have gotten one of the bum chips as far as temp reading goes.

For the motherboard...

1) cpu fan in F11 runs high. The lowest it ever goes is about at 40% of the fan's range, even though temps are fine and not necessary (unless somehow the 10 degree high core is somehow effecting them PWM control). Also, voltage/auto settings in bios seem to do nothing, and there is no option to control the fan by temp.

2) #2 sys fan header is flakey. I havent put a multimeter to it to see exactly what it is putting out, but fans attached to this do not run full speed. I've tried 3 completely different fans with same result...all considerably slower.

3) Even under ctrl + F1 advanced mode, BIOS (F11) appears to be missing a couple options....namely temp control for the cpu fan, and multithreading enable/disable. Granted options can come and go with updates, and it is possible that temp control was taken out although I dont see why it would be taken out, but the multithreading option, since it was a change thus def supposed to be in there, I dont get that one. Other new item like loadline calibration is there though.

4) CPU temp.....I dunno if I buy the reading given by BIOS on this. It seems awfully high, and is awfully high compared to various utilities in windows. I'm more inclined to believe the various utils in windows for CPU temp as the readings are more in line with the core temps and respective delta between those and die temp then BIOS is. Usually not a big deal, but if it affects cpu fan speed, it is just creating more unnecessary noise with the higher temp readings.

5) The boot cycling....OMG how annoying! It has happened to me SEVERAL times! Seems to be very sensitive to OC/voltage settings....too much so! I can be 24hr prime stable on various oc and over or under volted settings...go in and out of standby fine...basiclly a rock stable system while running....untill it tries to reboot and does the cycling. It seems pretty random too. Same situation, can reboot fine like 3 or 4 times, then even though nothing at all has changed, the next reboot cycles. Also seems to happen at 100% default no oc no voltage adjusted settings every now and then at random too, albeit no where near as often. Seems like the only way to 100% make it not do it is run 100% stock, and turning off even the modest default performance enhancements in BIOS.

6) Last big issue....I dont think it's a fault of either piece individually, but the motherboard and creative xfi extreme gamer PCI sound card do not play nice together. At first, the machine wouldnt even get through post with it in. Once I figured out that was the cause, it took disabling EVERYTHING I possible could in BIOS to get it to post with that card in there. Just using a different slot didnt change anything. Now, with onboard audio off (obviously), com port off, parallel port off, it posts. turn on either one of those 2 ports much less the audio, nothing. Even that is not 100% as I have noticed a modified beta driver that changes some I/O aspects of the soundcard is the only driver package that wont hang the system when the device tries to come online. Again...I'm pretty sure this is just one of those random times where 2 things just dont play nice, but thought I would mention this as well since I have seen motherboard BIOS updates in the past that address similar situations.

In the end....this is probably the most problematic build I have done....EVER! I've been putting together my own machines once every year or 2 since the 386SX-16 days too. Hopefully most those issues will be addressed with future BIOS updates. Other then those items, I am very happy with the motherboard. The CPU core temps bug me though. I have already gone as far as lapping the heatsink as well as the cpu heatshield to see if it resulted in any change at all. It did lower temps another few degrees, but the difference is still there. I have a feeling that will get the better of me and the CPU is gunna be ditched for either a conroe or more likely one of the new quads when they come out.

02-17-2008, 01:55 AM
Just for the sake of it....I have a GA-P35-DS4 v2.0 ( Bios F11 ) with a Q6600 @ 3.6, 2Gb of D9GMH and have no issues whatsoever.

In fact this is probably the most straightforward build I have ever done ;)

02-17-2008, 02:46 AM
Just for the sake of it....I have a GA-P35-DS4 v2.0 ( Bios F11 ) with a Q6600 @ 3.6, 2Gb of D9GMH and have no issues whatsoever.

In fact this is probably the most straightforward build I have ever done ;)

I was counting on this being straight forward as usual, but wasnt the case.

CPU is kind of a bummer deal...wasnt expecting that at all, but what can you do.

The rest of it seems to be all common minor BIOS issues, so hopefully that all comes out in the wash with future BIOS updates. Nothing that affects normal system operation except that random boot cycling. Would be nice if the sound card played better too I guess, but it works at least.

Rest of the specs for anyone interested....

Antec P180 case (carried over)
Corsair 750w PSU
2gb (2x1) DDR2-800 Crucial Ballistix
4ea Samsung spinpoint 160JJ's in raid 0 on ICH9 (carried over. possibly will be changed for a new 32mb cache 500gb or 2 soon)
BenQ DW1640 IDE DVDRW (carried over)
EVGA 7900GT CO (carried over for now... new vid coming soon, 8800GT or equivilant or better)
Hauppauge PVR250 (carried over)
Creative X-fi Extreme Gamer sound card (carried over)
Mitsumi 7 in 1 3.5" floppy + card reader (carried over)

Running XP MCE2005. CPU running 3.6ghz (400fsb) @ 1.15v load. Ram 4-4-4-12 800mhz (2x multiplier) @ 2.16v. Vid card running default 500mhz gpu and 1500mhz ram due to already being RMA'd a couple times and these cards tendancies to artifact and corrupt. I'm afraid to bump it up since this 3rd card actually works ok. I dont want to mess that up. haha.

02-17-2008, 05:03 AM
For the temps...if it is around 70+ in coretemp and CPU-z I would say you need a better heatsink, or reapply less paste

What voltage are you using for 3.6?

02-17-2008, 05:43 AM
For the temps...if it is around 70+ in coretemp and CPU-z I would say you need a better heatsink, or reapply less paste

What voltage are you using for 3.6?

It's set on auto still in BIOS, and it has been giving it 1.06 idle and 1.15 load. Just threw it there and left it for now since about all these chips do 3.6 on default voltage fine.

I dont have an issue with heat, rather just 1 of the core temp sensors. CPU temp is fine. I'm up in the 1.3v's and 4.2ghz+ before the die temp gets in the higher 60's, getting near the 72.4 max. Start getting up that high and the core temps even out and come together in the mid to upper 70's. Anything less then that though, the cores are significantly different. All default, 10-12 degrees different between core 0 and core 1. 3.6ghz same volts, idle is about the same with load 1 or 2 degrees higher with the same spread across the cores. Start getting 3.8 on up + a little bump in voltage where both cores start getting into the mid to upper 60's, then the spread starts going going down as the gap narrows to about 5 degrees different...then as mentioned with even more speed and voltage, by the time the cores reach mid to upper 70's, they even out and read the same.

It's something with these cpu's though. Search for E8400 temp problems or anything similar and you find tons of similar threads. I've reseated and switched heatsinks numerous times, and even lapped both the better heatsink and the cpu's IHS, and while it lowered temps across the board a few degrees, the difference is still the exact same. It's something internal like a miscalibrated diode or something.

oh yeah...70's for the core temp isnt that bad either. Higher then anyone would like to see of course, but the throttle temp is based off the die temp, or cpu temp in other words, which should be 10-15 degrees under the core temps where the heat is generated. Core temp just needs to be under the Tjunction....which is commonly thought to be 105 for these, but some evidence is out there that it may actually be 95. For my own taste, I dont want to see my cores higher then 20 degrees below tjunction...so with the most commonly given 105 tjunction, no higher then very low 80's....which the cpu/die temp is about 70 even at that point. Too close to that 72.4.

02-17-2008, 07:09 AM
1.06 IS NOW WHERE Near enough for 3.6 on a quad. I dont know where you are getting your info but that is not enough. I thought maybe you was using too much is why your temps where high, but I guess not. that low of a voltage is most likely your reason for the reboot to default cycles you are running into.

I use everest for core temps and it is about the same as coretemp itself, it is speedfan that is 12-15 C off not coretemp

02-17-2008, 07:17 AM
1.06 IS NOW WHERE Near enough for 3.6 on a quad.

He doesn't have a quad ;)

02-17-2008, 07:29 AM
Ahh, just seen that! I musta looked at the quote above

02-17-2008, 07:55 AM
And also, once again, temps as in heat is fine. It's the core temps that are different! Thats my whole issue with the cpu, not heat that it is putting out.

Like right now at idle it looks like this.

cpu = 28
core 0 = 35
core 1 = 46

the cores should be pretty even...within a few degrees of eachother, not 10+!

02-17-2008, 08:07 AM
Yeah you're right of course and I think you are also right when you assume a faulty sensor/bios hook is the cause :)