ASUS M4N82/955 problems....plz NO AMD chipset suggestions

[ghot]

I finally broke 3.7 stable.....AND it reboots lol

Thank you......"theoldtimer"

.and get this: one stick in channel A and one in channel B LOL Corsair Dominators (1066)...yet CPU-Z still shows DUAL lol

Ambient 20C
settings for this run:

ASUS M4N82 Deluxe [980a chipset]
AMI BIOS 0802
Win XP Pro 32bit
CPU Multi = 18.5x
FSB Freq. = 206
5-5-5-24-30-2T-2.20v
Dram freq. = 1099mhz
CPU/NB Freq. = 2472Mhz
HT Link Speed = 2.0Ghz
CPU Vcore = 1.450v
CPU/NB Voltage = 1.2125v
CPU/VDDA Voltage = Auto

HT Voltage = Auto
NB Voltage = 1.28v
nForce 200 Voltage = Auto

Validation:

[theoldtimer]

Way to go ghot. You stuck with it and made it work and IMO that MB is not that easy to work with.

[BeepBeep2]

.and get this: one stick in channel A and one in channel B LOL Corsair Dominators (1066)...yet CPU-Z still shows DUAL lol

Of course it does, you have one stick in A and one stick in B. That's two channels.

Grats on 3.8 stable.


Also, try to tighten the tRC timing a bit, it's the main bandwidth maker on nvidia chipsets. I don't know why you are running 5-5-5-24, if it's stable at 5-5-5-15 you will gain alot (ALOT) more bandwidth also, just make sure to keep the tRC equal or over to tRP + tRAS. From tRC 30 to 20 you will most likely gain a good 200 MB/s and tRAS down to 12-15 area around another 100 MB/s.

[ghot]

Of course it does, you have one stick in A and one stick in B. That's two channels.

Grats on 3.8 stable.


Also, try to tighten the tRC timing a bit, it's the main bandwidth maker on nvidia chipsets. I don't know why you are running 5-5-5-24, if it's stable at 5-5-5-15 you will gain alot (ALOT) more bandwidth also, just make sure to keep the tRC equal or over to tRP + tRAS. From tRC 30 to 20 you will most likely gain a good 200 MB/s and tRAS down to 12-15 area around another 100 MB/s.

well the motherboard manual says:

Due to CPU limitations.......DDR2 1066 Ram is supported for one DIMM per channel only...

and on the previous page....

the yellow slots are labeled DIMM_A1 and DIMM_B1...the black are labeled....DIMM_A2 and DIMM_B2....

....in either case with DDR2 1066 RAM in either 2 black slots OR a black and a yellow slot...I still have same issue....only now...and probably before...I can reboot ONLY on second try at anything over 3.7Ghz...its stable IN Windows...it's still just the reboot problem. At 3.7Ghz it reboots fine...EVERY time lol

The only change I've found is do NOT run Everest in the background when you are running Prime 95

.and bro I've long ago tried ALL the variations on the RAM timings....and done THAT on 3 different sets of RAM. From the screens above lowering any RAM timing just makes it BSOD at boot ...faster :/

[Evil]

Just got then sama board and cpu.
ghot, how´s the vdrop on your board?
I have 1,4v in bios, in windows cpu-z shows 1,376v idle and 1,344v-1,360v under load. Sucks a bit.

[ghot]

mine does same thing in CPU-Z, and I don't know why.....however....In Core temp it shows exactly what it is set in the BIOS and doesn't fluctuate???

[ghot]

OK, what would cause 2 cores to fall behind the other two.....it started when EITHER the 12K FFT's started or I started talking on CELL phone...both occurred at same time??





For this run I switched from Corsair 1066 2x2GB Dominators to OCZ 1066 2x 2GB (OCZ2P1066AM4GX, made for use with AMD Chips), which fit in the yellow slots as recommended by motherboard manual for best overclocking, AND fit under my CPU cooler

Validation:

[BeepBeep2]

Did you try FSB clocking? I'd much rather run OCCT for 2 hours on small FFT than Prime95 for 12 and a chance at killing the mobo...

[ghot]

yea I have and it works also NOW that I found the problem i was having with the 0x0000000A stop errors at every other reboot.....

in the 0802 AMI BIOS the Suspend Mode is set to auto by default. I don't even USE Suspend Mode......I set it to S1/POS and no more 0x0000000A stop errors at Windows Boot progress bar loading screen :/

The ONLY reason I thought of that, was that my older ASUS M2N32 SLI Deluxe 590 chipset board with the 1603 Award BIOS didnt have an Auto option....and since I don't use Suspend anyways...I set that one at S1/POS.

Best I never meet the BIOS engineer who put the Auto setting in the 0802 AMI BIOS.....2 months and 250$ in different memory companies and configs for nothing :/

[ghot]

OK what subsytem causes this???









Blue screened at 23:38

1st SS all cores still in sync...2nd and 3rd SS cores 2 & 4 lagging bad???

Cpu Multi = 19x
FSB = 200
Dram freq = 1066
CPU/NB = 2600
HT Link Speed = 2000
CPU Volt = 1.450
CPU/NB volt = 1.2125
CPU VDDA = Auto
HT Volt = Auto
NB Volt = 1.20

DRAM Timing: 5-5-5-15-22-2T-2.20V

Memory Config
MemClk Tristate C3/AltVid = Disabled
Mem Hole remapping = Enabled
DCT Unganged mode = Enabled
Power Down Enable = Disabled

[beta]

The lag is caused by other background processes, that's normal windows behaviour. Setting higher priority for the threads will help, but can also make your system unresponsive, so be careful.

[BeepBeep2]

...Blue screened at 23:38
...
Cpu Multi = 19x
FSB = 200
Dram freq = 1066
CPU/NB = 2600
HT Link Speed = 2000
CPU Volt = 1.450
CPU/NB volt = 1.2125
CPU VDDA = Auto
HT Volt = Auto
NB Volt = 1.20

DRAM Timing: 5-5-5-15-22-2T-2.20V

Memory Config
MemClk Tristate C3/AltVid = Disabled
Mem Hole remapping = Enabled
DCT Unganged mode = Enabled
Power Down Enable = Disabled

Is 1.45v all you are willing to put in it? 3 hours and a crash isn't stable. I like to use OCCT because it tells (tries) you what subsystem went bad, and which core errored. I've passed 2 hour prime blend then failed OCCT small fft in 19 minutes.

[ghot]

Is 1.45v all you are willing to put in it? 3 hours and a crash isn't stable. I like to use OCCT because it tells (tries) you what subsystem went bad, and which core errored. I've passed 2 hour prime blend then failed OCCT small fft in 19 minutes.

No....it doesnt crash...it just blue screen w/e I run a monitoring program.....like everest or Core temp..it would have run prime all night...btw you have to run prime for 3 hours minimum...either that or manually set the size range for the FFT's. As for OCCT it won't let me change the test...at least the latest version...it set on medium test, and I can't change it...its greyed out.

If I don't turn on any of those monitor apps above 3.7ghz, it runs fine no problems, as soon as i do....blue screen.

...and I've tried ALL voltages....that's just where it WANTS to be

at 3.7ghz its 24/7 stable with ALL monitoring apps running at only 1.425v.

[BeepBeep2]

...btw you have to run prime for 3 hours minimum...either that or manually set the size range for the FFT's.

I manually set it to blend, stopped after 2 hours , 40 minutes...complete pass, OCCT failed in 19 minutes and you could see the timer kind of unstabily counting (not in time)

As for OCCT it won't let me change the test...at least the latest version...it set on medium test, and I can't change it...its greyed out.

Because at the top it says "Auto 1 hr" with a dropdown box. You need to switch that to custom to switch FFT's. (Small, med, large)

If I don't turn on any of those monitor apps above 3.7ghz, it runs fine no problems, as soon as i do....blue screen.

Only over 3.7 Ghz? Try multitasking and running multiple tasks as you prime or whatever program you choose...OCCT has monitoring built-in, so I'm sure that would be a better stability test for you in this situation.

...and I've tried ALL voltages....that's just where it WANTS to be

at 3.7ghz its 24/7 stable with ALL monitoring apps running at only 1.425v.

[ghot]

I know a lotta ppl like OCCT....I don't....as far as I'm concerned Prime is the way to go....well Orthos was the best....but it wont run on four cores, even WITH 2 instances and the affinities set :/

OCCT really screws up your registry bro, even after you uninstall or remove it ^^

[BeepBeep2]

I know a lotta ppl like OCCT....I don't....as far as I'm concerned Prime is the way to go....well Orthos was the best....but it wont run on four cores, even WITH 2 instances and the affinities set :/

OCCT really screws up your registry bro, even after you uninstall or remove it ^^

That's something I've not heard before...it loads and unloads drivers every time it starts. It doesn't even make registry keys if I'm right...

[ghot]

get JV16 powertools and run registry finder for OCCT

[Tommie_lj]

the vcore you see in coretemp is the VID setting, wich is what you set in the BIOS, and not the actual vcore..

i have an ASUS M4N72-E board, with Nforce 750 chipset, and set at VID 1.475, i get 1.42V IDLE, 1.38V LOADED, monitored with CPU-Z wich actually shows very close to the real voltage on this MB. i have checked with MM.

[ghot]

Im lost...I know what the vcore is and about vdroop???

[Tommie_lj]

i reffered to this "mine does same thing in CPU-Z, and I don't know why.....however....In Core temp it shows exactly what it is set in the BIOS and doesn't fluctuate???"

[BeepBeep2]

the vcore you see in coretemp is the VID setting, wich is what you set in the BIOS, and not the actual vcore..

i have an ASUS M4N72-E board, with Nforce 750 chipset, and set at VID 1.475, i get 1.42V IDLE, 1.38V LOADED, monitored with CPU-Z wich actually shows very close to the real voltage on this MB. i have checked with MM.

Loadline Calibration?

[Tommie_lj]

Loadline Calibration?

what the h is that ?? ^^

i want to get rid of the voltage drop on LOAD, if that is possible.. at 1.55V the drop is 0.05 volts..

[BeepBeep2]

Nevermind...something ASUS puts in some BIOSes for some mobo's...gets rid of vdroop.

[Cheeseball]

Er, I think BeepBeep2 means to use it since you're experiencing vdroop, not to nevermind it.

[Tommie_lj]

i dont think i have this setting in my BIOS, but im going to check one more time..

time to go to bed.. its 02.00 at the night here in sweden