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

[shimmishim]

yo ucant' use nTunes with an intel board.

ntune is for nvidia boards.

[hotchoc]

Found out that my cpu is kinda vcore hungry. This is my current settings (stable in OCCT, Orthos, SuperPi, XP):

Vcore =1.4125
Vfsb = 1.350
Vmch/ich =1.600
FSB = 355 X 9 @3195
Temps = 42C idle/58C load
Ram = 5-5-5-15 (1:1) 2.2vdimm

Am starting to have doubts as to achieving 3.6Ghz under 1.5 vcore

Will increasing the vmch and vfsb affect the ocing in my case?

[gtj]

Thanks gave this a try..the Corsair HX620..has cables are black..so I found a molex that was unused from a fan in my case

BIOS setting 1.525v bios reads 1.535...DMM check 3 of the squares..1.39-1.42v!!!..still not sure if this correct(in winxp at idle...not I have RMclock with C1e disabled and disabled in bios C1E and EIST..using bios 2333

Blah...I reset bios to optimal defaults and turned off C1E/EIST...stock idle vcore is 1.16v....load was the same

I checked my DMM on some batteries..AAA 1.5v..reads them fine

other than 2333 bios what would you guys recommend ..2507?

Make sure that you're measuring on the CPU side of the blue or black power chokes. Measuring on the side closest to the blue heatsinks will give you inconsistent results because you're measuring before the filter. The 1 in the upper left corner is the only one that's safe to use because the rest all have caps right next to the outputs. Check the voltage section of the guide for pics.

In Everest the core voltage is clearly marked CPU Core on the Computer/Sensor page not to mention the OSD. If you don't see voltages or fans then you're using an older version.

[nealh]

Make sure that you're measuring on the CPU side of the blue or black power chokes. Measuring on the side closest to the blue heatsinks will give you inconsistent results because you're measuring before the filter. The 1 in the upper left corner is the only one that's safe to use because the rest all have caps right next to the outputs. Check the voltage section of the guide for pics.

In Everest the core voltage is clearly marked CPU Core on the Computer/Sensor page not to mention the OSD. If you don't see voltages or fans then you're using an older version.

Thanks as always..I was check the one closest to HS, not closet to cpu
the one in the pic is blocked by my storm and I can get to it...I will try again

the odd thing is everyone is reporting 1.5v..reads 1.48v....in my case it seems the vcore bios setting is lower than what bios reads(ie 1.525v=1.535v bios reading..so it seems my board is potentially overvolting)

Most have reported in the thread..better clocks at lower vcore on their cpu vs other boards...not for me..looks like badaxe2 need more vcore for same speeds

Gtj...if you set 333 ref freq..do you see 1333 strap in bios...if I do this I still see 1066 strap

[Reinvented]

Do you perhaps have Enhanced Powerslope on? Because that might overvolt it.

[gtj]

Thanks as always..I was check the one closest to HS, not closet to cpu
the one in the pic is blocked by my storm and I can get to it...I will try again

the odd thing is everyone is reporting 1.5v..reads 1.48v....in my case it seems the vcore bios setting is lower than what bios reads(ie 1.525v=1.535v bios reading..so it seems my board is potentially overvolting)

Most have reported in the thread..better clocks at lower vcore on their cpu vs other boards...not for me..looks like badaxe2 need more vcore for same speeds

Gtj...if you set 333 ref freq..do you see 1333 strap in bios...if I do this I still see 1066 strap

The "System Bus Speed" never changes from 1066 if that's what you're looking at.

Do you perhaps have Enhanced Powerslope on? Because that might overvolt it.

That's possible. It's also possible you need to recalibrate your DMM.

[nealh]

The "System Bus Speed" never changes from 1066 if that's what you're looking at.



That's possible. It's also possible you need to recalibrate your DMM.

Ok..then that is normal to 1066 strap

I doubled checked my DMM on batteries so it is fine....reads there voltage perfectly

Reinvented: I do have power slope enabled..I thought everyone did when overclocking?
The Badaxe2 seems to need more vcore than my RD600 for same speed..although it will actually run 3600 but I cant get 3mark05 cpu tests to run..these tend to be a good quick sense to whether I am going to be stable(although I got 3mark05 to run and P95 crashed ..first time i had that happen..usually if 3dmark05 runs P95 usually is ok...usually the both fail equally)

EDIT:..thanks GTJ...by using the upper left power choke and measuring on the cpu side..I could more appropriate results

I tested with power slope off....wanted to see what this was doing
Power slope on:
1.525v bios setting=1.535v bios reading
Power slope off
1.525v bios setting=1.511-1.523v bios reading

DMM..power slope off
1.525v bios setting= speedfan reports 1.52v idle= DMM reads 1.47v
speedfan reports 1.51v load=DMM reads 1.46v

so on my board looks like bios idle vcore is 0.0525v lower and load is 0.0625v lower than what is bios setting and windows reports .04-.05 off
It appears that the winxp programs (speedfan and everest) are matching what the bios is reading)

I feel much better that my vcore is really not 1.53v as measured by my DMM

BTW are others finding Power slope to be a big help with overclocking....from what I could see by my load vcore DMM reading the vdroop is excellent...and powerslope is giving a bit more juice for a given vcore setting

[Speederlander]

Well, after getting the new (replacement) board and overclocking it for the last few weeks, ended up with the following:

QX6700 @ 3.3 GHz, 330 (x10) FSB with 1.3625 vCore set in BIOS on air cooling.
FSB voltage is lowest possible setting in BIOS (1.2v I think).
MCH voltage is lowest possible setting in BIOS (1.5v I think).

Prime95 with all four cores running small fft gets to about 70C after 24 hours of stressing it in typical indoor temps. So far have stressed it for 24 hours of small fft and 24 hours of blend with no errors.

Memory is 4 gig Mushkin 8500.
running 4:5, 825 with timings 4/4/4/12 and vDIMM of 2v.

I suspect I can get more out of the memory if I want to push it but I don't see much need atm.

Not bad for a 24/7 air set-up.

Can't get it to high FSB and stable no matter what I do. Interestingly I booted into 401 FSB on my P5WDG2 WS Pro by mistake with the QX6700 and it posted no problem, though I am sure Windows would have been impossible.

[MuttonFlanks]

Well I guess my overclock wasn't stable as Orthos stopped after 2 hours running overnight. I'll try upping the vcore and see if I can get it to do better.

[MuttonFlanks]

I just upped my vCore to 1.4000 and the readings in CPU-Z and Core Temp didn't change. That's odd, could be a bug.

Anyways I was curious, what are the temperatures supposed to be like with a setup like mine at 3.42GHz? I'm running Orthos right now and with 1.4000 vCore it's about 36/39 idle in Core Temp/TAT. Under load it's about 61/62. That seems kinda high to me with a Tuniq cooling it?

I'm running it at room temp with no air con on just to be safe.

[MuttonFlanks]

Hmm, guess I got a little over-excited in my first post :P

I can't seem to get Orthos to run for more than 20 min on anything over 3.4GHz. I can get into Windows fine and play games and everything all the way up to 3.60GHz, but not Orthos stable. I think I will just run on Priority 1 instead of Priority 9 to test for now.

At 3.42GHz with 1.4500 vCore I actually saw my temps touch 70 degrees. The Tuniq could be seated improperly, but I'm too lazy to take everything apart again and check, haha.

At stock speeds though with the Tuniq and Intel HSF I saw a difference of about 9-10 degrees idle and 15-16 degrees load, so that has me thinking my Tuniq is working fine.

Right now I'm just doing 3.24GHz (360 FSB) at 1.350 vCore and everything else stock. The memory I have at 4-4-4-10 at 2.04 vDimm and on Orthos I'm at 2 hours. Will let it keep running overnight and see if it's stable. The temperatures are under 40 idle and under 60 load and I'm perfectly happy running at this speed.

Maybe when the TR Ultra 120 comes out I'll try it and see if I can get better temps

[nealh]

Just a quick followup

Power slope adds a small bump in vcore at a given bios setting..sort a in between value...does nothing for vdroop

Values tested with DMM

No power slope
Bios setting.....Bios reading......winxp idle/load........DMM idle/load
1.525v ........1.511-1.523.....1.51-1.52v/1.51......1.47/1.46

Power slope on
Bios Setting......Bios reading......winxp idle/load.......DMM idle/load
1.5375 ............1.54v ...........1.55-1.54..........1.49v
1.5625 ............1.57v ...........1.57v/1.56v.......1.52v/1.51v

I used Speedfan to report idle/load winxp pro values(everest is the same)

@1.5375 P95 blend failed immed and I could not get a load temp..I was using blend....
@1.5625 P95 blend generally gave a load 1.52 with a few at 1.51v..small fft gave 1.56v

So Powerslope seems to give a bit more juice and vdroop ON MY board with bios 2333 is excellent at .01(better than Intel specs? of 0.02-0.05)

But my bios reports vcore much higher than actually..typically .05v close to what Rick fx saw way back in the thread

Also the software programs seem to pull data very accurately from the bios ..they read the same

So if you have a DMM measure the vcore if you can...others have reported only .01-.02 difference between bios and actually values

[Lome]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah


warning: noC1E + x64 os = BSOD.

[gtj]

Just a quick followup

Power slope adds a small bump in vcore at a given bios setting..sort a in between value...does nothing for vdroop

Values tested with DMM

No power slope
Bios setting.....Bios reading......winxp idle/load........DMM idle/load
1.525v ........1.511-1.523.....1.51-1.52v/1.51......1.47/1.46

Power slope on
Bios Setting......Bios reading......winxp idle/load.......DMM idle/load
1.5375 ............1.54v ...........1.55-1.54..........1.49v
1.5625 ............1.57v ...........1.57v/1.56v.......1.52v/1.51v

I used Speedfan to report idle/load winxp pro values(everest is the same)

@1.5375 P95 blend failed immed and I could not get a load temp..I was using blend....
@1.5625 P95 blend generally gave a load 1.52 with a few at 1.51v..small fft gave 1.56v

So Powerslope seems to give a bit more juice and vdroop ON MY board with bios 2333 is excellent at .01(better than Intel specs? of 0.02-0.05)

But my bios reports vcore much higher than actually..typically .05v close to what Rick fx saw way back in the thread

Also the software programs seem to pull data very accurately from the bios ..they read the same

So if you have a DMM measure the vcore if you can...others have reported only .01-.02 difference between bios and actually values

Your DMM reading are unusual. Try measuring your +5v and +12v with your DMM then compare it to what the BIOS or other programs read.

[gtj]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah

Cool. Thanks.

[theonlybabyface]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah

Thanks. Works like a charm.

[Speederlander]

Oddly, my BIOS says my CPU temp is 0C. And I'm not on phase.

[nealh]

Your DMM reading are unusual. Try measuring your +5v and +12v with your DMM then compare it to what the BIOS or other programs read.

why would you say that...
Rickfx got very similar result

my board is revision 505....
PSU= Corsair HX620W..brand new for this build


EDIT:

Lome...thank you works like a charm

Gtj

Well I tested everything with my DMM

BIOS Readings
12v= 12.375
5v= 5.058
3.3v= 3.353
1.5v= 1.607
Vcore= 1.535

DMM Reading
12v= 12.10v
5v= 5.00v
vcore= 1.49v

[ChaosMinionX]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah

Great Stuff, worked for me

[Chosen.]

Well, after getting the new (replacement) board and overclocking it for the last few weeks, ended up with the following:

QX6700 @ 3.3 GHz, 330 (x10) FSB with 1.3625 vCore set in BIOS on air cooling.
FSB voltage is lowest possible setting in BIOS (1.2v I think).
MCH voltage is lowest possible setting in BIOS (1.5v I think).

Prime95 with all four cores running small fft gets to about 70C after 24 hours of stressing it in typical indoor temps. So far have stressed it for 24 hours of small fft and 24 hours of blend with no errors.

Memory is 4 gig Mushkin 8500.
running 4:5, 825 with timings 4/4/4/12 and vDIMM of 2v.

I suspect I can get more out of the memory if I want to push it but I don't see much need atm.

Not bad for a 24/7 air set-up.

Can't get it to high FSB and stable no matter what I do. Interestingly I booted into 401 FSB on my P5WDG2 WS Pro by mistake with the QX6700 and it posted no problem, though I am sure Windows would have been impossible.

Same situation here. Very weird behaviour by this board. I was running 389x9 for like a week, no problems whatsoever. Then I couldn't cold-boot it at all.

Also, there seems to be a hole between 333-350 for me. 333x10 is my everyday setting. I completed all 3d benches and whatnot at 375x10 and some lighter benchmarks at 385x10. However, 334x10 -> no POST.

Not sure what is causing it, I am suspecting my Gskills, so I ordered some OCZ Platinum Rev2 to see if there are any differences.

[gtj]

why would you say that...
Rickfx got very similar result

my board is revision 505....
PSU= Corsair HX620W..brand new for this build


EDIT:

Lome...thank you works like a charm

Gtj

Well I tested everything with my DMM

BIOS Readings
12v= 12.375
5v= 5.058
3.3v= 3.353
1.5v= 1.607
Vcore= 1.535

DMM Reading
12v= 12.10v
5v= 5.00v
vcore= 1.49v

Bingo. Your DMM reads low or your sensor chip reads high. I'm betting it's the DMM.

[nealh]

Bingo. Your DMM reads low or your sensor chip reads high. I'm betting it's the DMM.

why would you say the DMM is wrong..I tested numerous batteries and they all read correct voltages not low

Hmm..this is not good if your correct
if my DMM is wrong and the bios is correct..then the BAdaxe2 needs alot more vcore vs the RD600

ok..I guess I will go buy another DMM and see waht happens

If my vcore really is idling at 1.53/1.54v and loads at 1.52v...then an idle temp of 40/42C with ambient of 24.5C is amazing on e6600

[hotchoc]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah

how do u use it ?

[azzido]

how do u use it ?

read noC1E.txt file (right click on noC1E.inf and select install; reboot your computer)

Thanks Lome it works great

[Chosen.]

this is my c1e fix
based on info I found here http://www.xtremesystems.org/forums/showthread.php?t=104811
its just a driver (very small) that sets the propper msr to disable c1e (end then exits)
only tested on my 6600 and vista but i dont see why it wouldnt work with anything else (core2/x32 WINNTOS/other boards besides xbx2)


got sick of waiting for intel to fix it
I included source so you can check it out and whatnot
at your own risk blah blah

One of those little things that make your day. Many thanks, mate, for providing something that a company the size of Intel thought was unimportant to fix.