[18th Anniversary] Asus Black Pearl - P5k-Premium Testing Report

[stepinrazor]

I was reading the P5K-E thread and people on there say that at the lowest voltage setttings for VMCH, VPLL, VFSB it overvolts. Is this the case with the P5K-Premium too? If I had a multimeter I'd test it myself if someone would give me the measuring points.


Per Barbornet on P5K-E Thread

VMCH BUG
Bios ---------------- Tester

1,25v(or AUTO) --- 1,55v
1,40v -------------- 1,40v
1,55v -------------- 1,55v
1,70v -------------- 1,70v


VPLL BUG
Bios ---------------- Tester

1,5v(or AUTO) --- 1,8v
1,6v -------------- 1,6v
1,7v -------------- 1,7v
1,8v -------------- 1,8v


VFSB BUG
Bios ---------------- Tester

1,2v(or AUTO) --- 1,4v
1,3v -------------- 1,3v
1,4v -------------- 1,4v
1,5v -------------- 1,5v

[jonny_ftm]

Already have, did blend for over 4 hours with no errors.

Prime95 Blend won't reveal the NB-RAM unstability as F@H or OCCT in RAM mode will do. OCCT is though faster than F@H, the last one needing more than 12h to reveal the problem. Look at the system temperature when you OCCT in RAM mode or F@H, it really cranks

What I mean, try to loosen your OC. If the problem disappears, than, like it or not, your OC is unstable. If the problem persists at stock or when you loosen your memory and CPU OC, than it could be a board/bios issue.

We don't have all the same testing rigour for defining stability. Most XS members will go with a SuperPi and 1h of priming. For me, this is just good for world records or gaming, not to be called stable for 7/24 use. It is a personal choise, no need to launch that hot debate anyway.

Thanks all for the info..
Im now running as below
cpu volt 1.3625... is this about right for 3.4Ghz?
cpu pll 1.5
fsb volt 1.4...... was 1.3 but does it matter ?
nb volt 1.4
sb volt 1.05
clock over-charging volt ???? what is this????
cpu gtl 0.63x
nb gtl 0.67x

This is with the 266 strap and 380 fsb and this gives me 953Mhz ram. Now when i use the 333 strap then it gives me 913 Mhz ram but this also causes some blue screens for me when booting in to windows also ive not had one yet using the 266 strap " touch wood" but i may run memtest just to be sure.

Its just passed the occt 30 min test so i think prime is in order now..

cpu volt 1.3625... is this about right for 3.4Ghz?
Will depend on the CPU, no way to know. You have to run Prime95 Small FFT which is quiete only sensible to the vcore/vPLL and vFSB, ruling out memory and NB OC issues (providing they have a minimum stability)

cpu pll 1.5, fsb volt 1.4...... was 1.3 but does it matter ?, nb volt 1.4
sb volt 1.05
At the 333 strap, look at the link in my signature, these voltages took me rock stable up to 475FSB on my Q6600. My board is very well cooled. I removed all stock cooling though. This suerely helps.

clock over-charging volt ???? what is this????
No need for it for now. Just set it to 0.8v. For a slight 3.4GHz, no need to touch vPLL or this setting

cpu gtl 0.63x
nb gtl 0.67x
No harm here, but for 3.4GHz only, you can set them to Auto for now. Once stable, look at them maybe, even if useless at 3.4GHz


Its just passed the occt 30 min test so i think prime is in order now..
This is only a quick test. I wil go only with OCCT in CPU mode or Prime95 Small FFT for overclocking the CPU. Set you RAM timings and voltage on Auto and don't overclock it till your CPU is stable (so lower your memory frequency to DDR2-800 even if it is rated at 1200MHz). If you keep like you do overclocking the RAM and CPU, you can't know where lies your problem. Once you're stable on the CPU, go to the memory and NB settings and test with OCCT in RAM mode

I was reading the P5K-E thread and people on there say that at the lowest voltage setttings for VMCH, VPLL, VFSB it overvolts. Is this the case with the P5K-Premium too? If I had a multimeter I'd test it myself if someone would give me the measuring points.


Per Barbornet on P5K-E Thread

VMCH BUG
Bios ---------------- Tester

1,25v(or AUTO) --- 1,55v
1,40v -------------- 1,40v
1,55v -------------- 1,55v
1,70v -------------- 1,70v


VPLL BUG
Bios ---------------- Tester

1,5v(or AUTO) --- 1,8v
1,6v -------------- 1,6v
1,7v -------------- 1,7v
1,8v -------------- 1,8v


VFSB BUG
Bios ---------------- Tester

1,2v(or AUTO) --- 1,4v
1,3v -------------- 1,3v
1,4v -------------- 1,4v
1,5v -------------- 1,5v

Humm, the vPLL issue is wired. But the X38 ASUS boards are all known to have overvolting issues. I lowered my vPLL to 1.5 and I didn't see any temperature increase (or decrease) compared to 1.6v, so I don't think it overvolts that much. Also vFSB and vNB. I'm not stable at lowest values while I'm stable at respectively at 1.3 FSB and 1.4 NB. So, it shouldn't overvolt in manual settings. In Auto, ASUS boards overvolt when you OC, so I will never put a voltage on AUTO

[Shocky]

Prime95 Blend won't reveal the NB-RAM unstability as F@H or OCCT in RAM mode will do. OCCT is though faster than F@H, the last one needing more than 12h to reveal the problem. Look at the system temperature when you OCCT in RAM mode or F@H, it really cranks.

12 hours?

Thats impossible as I'm not around long enough for that and I'm not leaving it on well i'm out or well i'm sleeping, but i'll try it for as long as I can, so what if that also gives no errors?

Anyway, what would you recommend I set for the other voltage settings, I've only changed core and memory voltage, the rest are all still set too auto.

[jonny_ftm]

12 hours?

Thats impossible as I'm not around long enough for that and I'm not leaving it on well i'm out or well i'm sleeping, but i'll try it for as long as I can, so what if that also gives no errors?

Anyway, what would you recommend I set for the other voltage settings, I've only changed core and memory voltage, the rest are all still set too auto.

OCCT RAM in 2h to 4h is where most errors will occur, so should be a rather stable test for the NB-RAM. Small FFT in Prime95 is a long debate :-)
List your confing and bios settings, and we can see
Also, at stock, does your mb dual boot too?

[Shocky]

OCCT RAM in 2h to 4h is where most errors will occur, so should be a rather stable test for the NB-RAM. Small FFT in Prime95 is a long debate :-)
List your confing and bios settings, and we can see
Also, at stock, does your mb dual boot too?

Ok, 4 hours shouldn't be a problem, to be honest I havent tested it at stock but I don't think it does, i'll try it later.

System.
Q6600 G0 (425x8) + 0504
P5K Premium
8800GTX
3GB (2x1GB + 2x512MB) 4.4.4.12
X-FI XtremeMusic
3x500GB Seagate drives
Tagan TG700-U33

All i've changed in the bios is is core, memory voltage, multipler, FSB and also strap to 400, also disabled C1E, CPU TM, everything else I set to auto I think.

[jonny_ftm]

Ok, 4 hours shouldn't be a problem, to be honest I havent tested it at stock but I don't think it does, i'll try it later.

System.
Q6600 G0 (425x8) + 0504
P5K Premium
8800GTX
3GB (2x1GB + 2x512MB) 4.4.4.12
X-FI XtremeMusic
3x500GB Seagate drives
Tagan TG700-U33

All i've changed in the bios is is core, memory voltage, multipler, FSB and also strap to 400, also disabled C1E, CPU TM, everything else I set to auto I think.

Ok. I'd really try stock. If it doesn't do it, than you have really to enhance your OC.

What I see is a memory issues maybe. You have 4 slots of mixed DDR2 sticks.
- strap goes to 333 (400 was revealed to be really unstable by many)
- memory voltage: must meet the minimum needed by the most demanding two sticks
- Memory frequency: don't set it to Auto as you have mixed sticks. Go to manual and select the multiplier giving the closest memory frequency to the lowest performant sticks. In your case, if you setup a 1:1 multiplier, your memory will be at 850MHz, be sure all sticks can do it at those agressive timings you put. In your case, I'll begin with manual and set 5-5-5-18-5-42-12-12-12. Static read control and transaction booster disabled at relax 3.
- vPLL to 1.50 or 1.60 (to avoid crazy killing auto overclock by ASUS
- vNB to 1.40 or 1.55 as you have 4 sticks of memory
- vFSB to 1.30 or 1.45

Now see if still dual boot. If all is ok, you can now tighten your memory timings and lower voltage to check for stability

Let us know

[Shocky]

Ok. I'd really try stock. If it doesn't do it, than you have really to enhance your OC.

What I see is a memory issues maybe. You have 4 slots of mixed DDR2 sticks.
- strap goes to 333 (400 was revealed to be really unstable by many)
- memory voltage: must meet the minimum needed by the most demanding two sticks
- Memory frequency: don't set it to Auto as you have mixed sticks. Go to manual and select the multiplier giving the closest memory frequency to the lowest performant sticks. In your case, if you setup a 1:1 multiplier, your memory will be at 850MHz, be sure all sticks can do it at those agressive timings you put. In your case, I'll begin with manual and set 5-5-5-18-5-42-12-12-12. Static read control and transaction booster disabled at relax 3.
- vPLL to 1.50 or 1.60 (to avoid crazy killing auto overclock by ASUS
- vNB to 1.40 or 1.55 as you have 4 sticks of memory
- vFSB to 1.30 or 1.45

Now see if still dual boot. If all is ok, you can now tighten your memory timings and lower voltage to check for stability

Let us know

Well heres a screen showing 5hr run of OCCT, i'll check those settings tomorrow and see if they make any difference, also it doesnt reset on default settings but it only does when first powering up so i'll leave it on defaults tonight and see if it does it tomorrow.

Thanks for the help.