Audio stutter after OC attempt.

[LostBok]

I've been tinkering again - in an attempt to get a few extra Mhz out of my 3500+ Venice for HL2 - turns out it's gonna be another month before I can get the cut-price 4400+ that I was after.

Step one: take memory back to SPD settings (2-3-2-6)

Step two: fire up Ai Booster and (leaving the multiplyer static), I pushed the FSB speed up one Mhz at a time in the tuning panel and applied each time.

There was a slight audio stutter around 216Mhz (2.376Ghz), so I figured the memory might be struggling at this speed, so I pushed the memory voltage up from 2.6V to 2.65V and it seemed better after that. So I tried playing HL2 for a bit, slightly smoother, but not much.

Back to Ai Booster... and I tried pushing the speed up some more: 218Mhz (2.398Ghz) and the audio stutter was quite noticeable again, so I pushed the memory voltage up to 2.7V.

All this time the most the CPU had done was move from 34deg idle to 36deg (although it did climb to about 42degs when gaming).

When I then pushed the memory to 219Mhz the machine did a BSOD, so I went into the BIOS and wound the memory speed back to 2.65V and 214Mhz and rebooted, reboot was fine, but the audio stutter was bad!

I then reset ALL the memory speed, voltage, etc to default and rebooted. Audio still bad.

I ran MemTest86 and went out for the evening... 50 full loops. ZERO errors. The memory is fine.

What else could I have broken?




Rig spec.

3500+ Venice w/ Akasa EVO120
ASUS nF4-SLI-Premium
1Gb (2x512Mb) twinX @ 2-3-2-6-T1
Radeon X800XL 256Mb DDR3 w/Passive Heatpipe

2 x 74Gb Raptor RAID0 on SIL3114
200Gb SATA Maxtor Plus 10 on nV SATA
300Gb SATA MAxtor Plus 10 on nV SATA
+external 1394 hard drives

Pioneer DVR-108 on IDE Primary MAster
Lite-On CDRW 52246S on IDE Priamry Slave
NEC 4551 DVDRW on IDE Secondary Master

USB2 hub
USB2 multi card reader
Dell Multimedia USB keyboard
Logitech MX700

[vikeor]

after OCing run Prime95 to see if ur system is stable... the audio stuttering may be causing by instability

[LostBok]

okay - thanks for the tip - will do.

I thought the only thing I'd really pushed hard was the memory and in my book, 50 passes in memtest86 is a pretty darned stable - I'll hit it with prime95 this evening.

[JLK03F150]

Audio stutter (or loss of audio) is usually a sign of an overclocked PCI bus. Even onboard sound uses the PCI bus. Figure out how to set it to Fixed at 33 mhz in your BIOS. Your ide HDDs communicate on the same bus so you are also risking data corruption.

The PCI bus can be overclocked to some extent, but the dividing line is very distinct.
37.5mhz everything is wonderful - 38mhz crash & burn. lol

Also as you get your FSB higher you need to loosen your ram timings to help keep them stable.

[LostBok]

I did loosen the RAM timings - they were VERY tight before!

PCI was on auto - I thought it would keep it around 33Mhz, but I'll check that - probably ought to lock it to 33Mhz!

Any idea why it's still "pickled" after I've clocked it back down?

I've known a seriously overclocked PCI bus to corrupt IDE data and to cause distortions, but not to bvgger up hardware... 219Mhz is the highest I went... that's only a 9.5% OC, so that's only 36.5Mhz! I used to OC mine to 41.5Mhz in the pre-AGP days (man that Orchid Righteous used to fly at 83Mhz FSB!!).

[LostBok]

JLK you were on the right track - the initial audio stutter was the PCI bus... the persistent audio stutter was because I had corrupted data on one of my disks, so even once I clocked back down I still had the prob:

I ran 50 FULL loops in memtest86 (overnight) without a prob and then a full night of prime95, without a prob.

next attempt: try a new soundcard - still has stuttering sound!

SOLUTION: checkdisk - with auto-fix.

Machine now boots perfectly: login sound is fine, half a dozen songs, no probs, a DVD.... it's all fine!!!

Needless to say I jumped straight back into the BIOS, locked the PCI bus on 33Mhz and cranked the FSB up to 218Mhz (at stock memory and CPU voltage) and it's PERFECT! Not even a glimmer of sound being messed up. I tried 219Mhz and memtest is a big screen of red - 218Mhz has done it's first pass in memtest and it's all fine.... gonna leave it overnight again....

Happy days!!

[JLK03F150]

Woohoo! Glad you figured it out because I didn't know the corrupt data could cause the sound issues to continue. Which test was causing the errors in memtest - 5, 6, etc?

[LostBok]

He he he....when I had it clocked to stock voltages and the memory at 220Mhz, 2-3-2-10-1T.... ALL OF THEM.... you could scarely see anything recognisable!

When at 218Mhz, stock voltages, it's made a couple of loops without a prob - I need to format an IDE drive and then I'm gonna try some HL2 "stability testing"

[LostBok]

Although.... having said that, it's a 250Gb disk and I need to do a FULL low level format as this disk is for my Sky+ box, so I might have to leave the HL2 testing until tomorrow

[JLK03F150]

I need to format an IDE drive...

At stock speed I hope...

I never run overclocked while doing Windows updates, installing software, manual virus scans, etc. Just a safe idea IMO.

[LostBok]

but if I clocked it back to stock speeds, then that would be like admitting it wasn't stable!!

I used to clock back a bit becuase I was worried about IDE data corruptions, . but with the PCI bus locked to 33.33Mhz, that shouldn't happen!