Testing / comparing : Intel D975XBX2 / Asus P5B DX ***56K WARNING***

[aamsel]

The last two batches of GSkill's that I bought were super-high dollar for 2 sets of: F2-8000PHU2-2GBHZ which is supposed to be DDR2-1000 @ 4-4-4-5. This is supposed to be their "super HZ" stuff. I can't fully put my finger on what is wrong, but my son had 2 sticks of it die, first one stick, then the other one at or below the rated 2.3v on a P5W64 WS Pro. I have been having weirdness with my systems since I started using it, freezing in CMOS, etc. About to try this same memory on a BX2 this weekend. This was after having some GSkill DDR-400 die last month.

I personally don't see any reason why memory rated higher wouldn't run lower.
That just makes NO sense to me.

I got the above GSkill memory to run at 950 on the P5W64, but not 1000. Of course, the memory manufacturer will tell you that it will get to 1000+ but at a much lower FSB, and only at certain timings.

I would set timings to SPD and run memtest.
If it fails then, I would consider it bad memory.

Andrew

Anyone seen memory errors on this board with DDR2-1000 that is clocked too low? Like in the DDR2500-800 range? G.Skill is telling me this RAM works but only if I increase the FSB to get it to DDR2-1000 but I think the chips are bad.

[nealh]

There's no sensor related to the NB. You're either reading the motherboard or memory. Check the guide for their locations.

ok...I assumed the mobo temp was NB temp..my bad..sorry...

That is really odd that there is no NB sensor..well still impressed that after 25hr of P95 blend with no fan the NB temp were not causing errors

BTW...no all board with overvolt or undervolt the same..get a DMM and measure the voltage...mine does not seem to overvolt much..0.01v

Lome's revision 505 ..seems like my 505..has bad bios readings and poor sensor which do not match DMM

[m4sopmod]

Ok i just flashed to bios 2333. Everything is stock. What i did was i disabled C1E, enabled powerslope in bios and installed no C1E file via XP.

Once i clicked on core temp the whole thing just reboots. Happened twice already. First time managed to spot a BSOD and the second time it just reboots by itself.

What's going on ???

[nealh]

Ok i just flashed to bios 2333. Everything is stock. What i did was i disabled C1E, enabled powerslope in bios and installed no C1E file via XP.

Once i clicked on core temp the whole thing just reboots. Happened twice already. First time managed to spot a BSOD and the second time it just reboots by itself.

What's going on ???

I owuld jumper back to maintenance mode..ie 2-3 jumper position and enter bios and check what the settings are...

I would load optima default...reboot..reset jumper and just load up windows/OS and make sure everything is OK

the nreboot and retry entering the bios

I bet the ram timings or vdmm is off

EDIT: OOPs..which coretemp are you using I have heard issues with 0.95...if your using this get 0.94(if you need I can send it to you)

[hotchoc]

Ok i just flashed to bios 2333. Everything is stock. What i did was i disabled C1E, enabled powerslope in bios and installed no C1E file via XP.

Once i clicked on core temp the whole thing just reboots. Happened twice already. First time managed to spot a BSOD and the second time it just reboots by itself.

What's going on ???

I found out that after i opened CPU-Z followed by core temp, the whole thing crashes. Saw another BSOD as well. It was ok in bios 2663 before.

Btw, do i leave 'Clear trusted platform module' as NO in the maintaince bios?

[MuttonFlanks]

Have you ran it again since your issues though?

No I haven't, I guess I'll re-run it and see just to be safe though.

I think I figured out my problem after reading gtj's guide. I'm a victim of the motherboard not saving voltage settings unless I completely unplug the power source and plug it back in. I noticed this after a few times of the computer not POSTing. For some reason I decided to unplug my PSU and let it sit for a few seconds and then plug it back in. After I did that the computer would boot without me having to change any settings Will have to try for stability later...

Guess I have a lot more testing to do tonight after work. Stayed up till almost 3am last night messing with it and had to get up at 7 for work. This can't be good for my health... lol

[MuttonFlanks]

As a reference, I did 9x400 with vfsb 1.3 and vmch 1.625.

mem i try to keep below 2.2 as the board overvolts about 0.03 (measured with DMM).

Also be sure yo haven't maxed your chip out. The F batches aren't known to be excellent o/c-ers

This is on stock cooling right? I only have the Tuniq on the CPU and the rest is stock. No fans on them or anything.

[hotchoc]

I owuld jumper back to maintenance mode..ie 2-3 jumper position and enter bios and check what the settings are...

I would load optima default...reboot..reset jumper and just load up windows/OS and make sure everything is OK

the nreboot and retry entering the bios

I bet the ram timings or vdmm is off

EDIT: OOPs..which coretemp are you using I have heard issues with 0.95...if your using this get 0.94(if you need I can send it to you)

nah i checked the ram timings n vdimm after goign thru the maintainence bios. it's all good 4-4-4-12 @ 2.04vdimm. windows xp will be ok running as logn i don't open cpu z or coretemp together. usign everest for nwo to monitor my temp. weird thing was, this never happened when i was in bios 2663.

yeah am using 0.95

do pm me the 0.94 link if possible

thks

[nealh]

Coretemp 0.94
http://www.thecoolest.zerobrains.com/CoreTemp/

[nealh]

Is there a way to measure chipset temp on this board?

Is there in option in everest to control fans speeds..I have been looking but cant seem to find it?

[gtj]

Is there a way to measure chipset temp on this board?

Is there in option in everest to control fans speeds..I have been looking but cant seem to find it?

Only with a temp probe.

Everest only monitors. Speedfan is the only program available to change fan speeds other than the IDCC.

[MuttonFlanks]

Well I am pretty happy now Now that I figured out my mobo wasn't saving my voltage settings every time, I tested a bunch of settings by unplugging the PSU each time to make sure the voltages stay.

Although I can't get 3.6GHz to stay stable with 1.525 vcore 1.40 vfsb and 1.60 vmch, I can do 3.51 GHz (390 FSB) and 1.42 (380 FSB) stable for 1 hour. Gonna leave it on overnight to make sure.

At 3.51 GHz my load temps are 63-64 which seems pretty safe to me. I've got it at 1.465 vcore 1.375 vfsb 1.575 vmch. I'm pretty sure my chip will be able to do 3.6 but I don't have all the proper cooling in place yet.

I'll probably pick up the Thermaltake Ultra 120 Extreme and the Corsair or OCZ RAM cooler. I also have a HR-05 and HR-05 SLI laying around that I can't install due to the size of the Tuniq. Once I get all those on I'll give 3.6 GHz or higher another try. But right now I'm happy at 3.51. Will try to lower vcore too and see if I can still get it stable. Will keep this thread updated

[hotchoc]

Well I am pretty happy now Now that I figured out my mobo wasn't saving my voltage settings every time, I tested a bunch of settings by unplugging the PSU each time to make sure the voltages stay.

Mind to elaborate more? So what you did was change the voltage say from 1.325 to 1.35 and then save the settings via F10. After that you manually off the pc via the psu plug or do you load into windows, shut down then continue with the psu thingy.

[MuttonFlanks]

Mind to elaborate more? So what you did was change the voltage say from 1.325 to 1.35 and then save the settings via F10. After that you manually off the pc via the psu plug or do you load into windows, shut down then continue with the psu thingy.

I up the voltage, press F10, press Y, turn off power, then unplug from PSU or from wall, whichever is closer to you. Then plug back in and go from there.

Just curious, if my comp is 8 hours Orthos stable but crashes when I open Everest, what does that mean? Right now I'm at 2.5 hours Orthos blend stable and still going, but before that I tried to open Everest a few times and it crashed every single time.

Also, running MemTest in Windows I pass 100% coverage no problem and 2 passes of MemTest+ from a bootable CD... but Everest crashes every time.

[shimmishim]

I up the voltage, press F10, press Y, turn off power, then unplug from PSU or from wall, whichever is closer to you. Then plug back in and go from there.

Just curious, if my comp is 8 hours Orthos stable but crashes when I open Everest, what does that mean? Right now I'm at 2.5 hours Orthos blend stable and still going, but before that I tried to open Everest a few times and it crashed every single time.

what version of everest are you trying to use? there are some older ones that crash the bad axe 2 even if you're 10000000 hours orthos stable

[MuttonFlanks]

what version of everest are you trying to use? there are some older ones that crash the bad axe 2 even if you're 10000000 hours orthos stable

Ultimate Edition 2006. It works at 3.24 GHz on 4-3-3-4 but crashes at 3.51 on 5-5-5-16

[nealh]

Memory testing...I found I could run P95 blend on an RD600 at 450mhz..burned a dvd and it was choppy as hell

reran...test 5...after 30 passes it threw errors..BTW this passed 1000% on win memtest....re did memory settings and reran memtest 86+ 1.70 test 5 for 50 passes....

DVD burns were fine

My point is 2 passes of memtest is no good....I would do no less than 10 passes of all the tests and still run test 5 for 50 passes

For me if I can pass test 5 ..50 times, I will not get memory related issues at a given timings ...

so i set mem first, run 50 passes on test 5(at given overclock)...then run 3DMark05, if this passes(those 2 cpu test are huge stress on the cpu) then i will run Blend P95

I used to accept 10-12hrs of P95...I kinda like 24 hrs of P95 now

But the memory has to go 50 passes or I really worry

Muttonflanks....make sure powerslope is on and this may allow a lower vcore setting

also what did you see with voltage in windows prior to figuring out the bios was not taking the settings

[MuttonFlanks]

Thanks for the tips Neal. 50 passes... damn I'll give it a try one night before I sleep or something.

I always have power slope enabled. What is P95?

I can't see voltages in Windows. Some reason CPU-Z and Core Temp both read my voltages at the default voltages no matter what I set it to in BIOS.

The thing that doesn't make sense about the RAM to me is at 3.24 GHz with the RAM at 4-3-3-4 233/566 and 2.16v, I can refresh the latency test in Everest non stop and no problem at all.

At 3.46 or 3.51 I can't even refresh it once at 5-5-5-18 233/566 and 2.16v.

[Console]

Thanks for the tips Neal. 50 passes... damn I'll give it a try one night before I sleep or something.

I always have power slope enabled. What is P95?

I can't see voltages in Windows. Some reason CPU-Z and Core Temp both read my voltages at the default voltages no matter what I set it to in BIOS.

The thing that doesn't make sense about the RAM to me is at 3.24 GHz with the RAM at 4-3-3-4 233/566 and 2.16v, I can refresh the latency test in Everest non stop and no problem at all.

At 3.46 or 3.51 I can't even refresh it once at 5-5-5-18 233/566 and 2.16v.

Prime 95 it is what ppl used before Orthos became the norm ( the new verion on p95 works on dual cores now )

[Skerlnik]

I've got an odd one. I am using Vista 32 Ultimate, and was using my Logitech Quickcam Vision when my computer blue screened and went to restart itself. When it went to load windows it claimed one of the loader files had a bad header and that I would need to do a repair from the Vista DVD. Ya right!

I power it off again, but then it failed to POST! I have a minor OC of 333fsb, 1.38vcore, 266/800 4-4-4-12 @ 2.32v memory (4x 1gb ocz titanium alpha). I move the jumper to maintenance mode and power back on and it gives me memory error beep codes! I had to pull 3 sticks to make it into the bios! I reset the memory voltage back to 2.32 in the bios and save but it still won't POST in normal mode.

I had to change my memory divider to 266/533 for it to boot in normal mode! Vista loaded just fine (I knew the files couldn't have been corrupt).

What gives? Any suggestion why my memory will no longer run at 266/800 on 333fsb? Is it possible the MB is damaging the modules?

Regards,
Sker

[nealh]

Thanks for the tips Neal. 50 passes... damn I'll give it a try one night before I sleep or something.

I always have power slope enabled. What is P95?

I can't see voltages in Windows. Some reason CPU-Z and Core Temp both read my voltages at the default voltages no matter what I set it to in BIOS.

The thing that doesn't make sense about the RAM to me is at 3.24 GHz with the RAM at 4-3-3-4 233/566 and 2.16v, I can refresh the latency test in Everest non stop and no problem at all.

At 3.46 or 3.51 I can't even refresh it once at 5-5-5-18 233/566 and 2.16v.

not 50 passes of all the tests..ugh

50 passes on test 5...this about 1 hr -1.5 hr max

test 5 ..always seem to be the one that failed for me if I pushed too much and if I dont get 50 passes(like i said once i hit like 38 passes amd then got memory errors), i get issues..ie dvd burns for example..yet Prime 95 passed fine

I personally like Prime 95 version 2.51....but orthos is fine but uses older version of prime, FWIW

when i set 3500..390x9 with ram at 487mhz(266:533) 2.16v 5-5-5-18...I can pass 50 passes on test5 but I could not get 3Dmark05 to run without pushing vcore up...it seems the cpu is getting stressed but I thought the memory controller was on the nB and not on the die of the cpu
with the opterons/A64 dual cores...if you pushed ram speed/timings this was common problem and I would need more vcore
I was surprised this happened on my C2D

I did not push past 1.6v on mch to see if I could get 266:533 stable at 5-5-5-18(also it maybe the 18 should be lowered ...sometimes it is weird how a lower timing may help too)

[gtj]

Thanks for the tips Neal. 50 passes... damn I'll give it a try one night before I sleep or something.

I always have power slope enabled. What is P95?

I can't see voltages in Windows. Some reason CPU-Z and Core Temp both read my voltages at the default voltages no matter what I set it to in BIOS.

The thing that doesn't make sense about the RAM to me is at 3.24 GHz with the RAM at 4-3-3-4 233/566 and 2.16v, I can refresh the latency test in Everest non stop and no problem at all.

At 3.46 or 3.51 I can't even refresh it once at 5-5-5-18 233/566 and 2.16v.

cpu-z and coretemp do NOT read voltages. They read the current VID. See the "Overclock/Voltages" section in my guide to see what that means. Everest Ultimate 2006 will cause a crash. You have to have a build number greater than 879 or something close. Get the latest beta.

[gtj]

I've got an odd one. I am using Vista 32 Ultimate, and was using my Logitech Quickcam Vision when my computer blue screened and went to restart itself. When it went to load windows it claimed one of the loader files had a bad header and that I would need to do a repair from the Vista DVD. Ya right!

I power it off again, but then it failed to POST! I have a minor OC of 333fsb, 1.38vcore, 266/800 4-4-4-12 @ 2.32v memory (4x 1gb ocz titanium alpha). I move the jumper to maintenance mode and power back on and it gives me memory error beep codes! I had to pull 3 sticks to make it into the bios! I reset the memory voltage back to 2.32 in the bios and save but it still won't POST in normal mode.

I had to change my memory divider to 266/533 for it to boot in normal mode! Vista loaded just fine (I knew the files couldn't have been corrupt).

What gives? Any suggestion why my memory will no longer run at 266/800 on 333fsb? Is it possible the MB is damaging the modules?

Regards,
Sker

You may have to increase the vMCH with 4 sticks. I had to.

You might also want to update your signature with your specs. It'll help the rest of us help you.

[ArcticOC]

GTJ, thanx again for your nice guide

Just one tip, could you please add a line where you date the last update/revision on your page?

[gtj]

GTJ, thanx again for your nice guide

Just one tip, could you please add a line where you date the last update/revision on your page?

That is going to be in my next revision.
This weekend.