I7 920 issues (i think)

[tcmike]

Hey guys,

So about 2 weeks ago i was doing some benchmarking (nothing seriously overclocked... i was just seeing what all the hype was about in 3dmark) and i decided to go into bios and do an overclock on the CPU (i wanted to see the difference between 2.66 ghz vs 3.6 ghz) (also its watercooled with a corsair H50). Well after i applied the overclock to the cpu, the system ran extremely slow and so i had to turn it off and reset the clocks (i overvolted by a single increment when i overclocked)

well one week later now, my entire computer is crapping its pants (it was running a bit sluggish after the OC, but its reached a point now where everythings messing up). 1 fan and a LED light i had plugged in just died over night.

Basically ill boot up the computer, and everything will just start not responding frequently. I was at a LAN party last night and we were transferring games to eachother using external hard drives and i would notice everyone finished within 10 minutes and when it gets to me, itll take up to 3 hours for 4 gigs. (note: i was watching the rates carefully and it seems the i7 would randomly start performing fast on the transfer, and then it would just stop and sit there for X minutes and start going a bit faster and then going back down hill again. rinse n repeat.)

I was tired of waiting for the download so i decided to start playing warcraft 3 dota with the others and i started noticing lag spikes happening every 5-10 seconds (hiccups).. when the download finished, this was no longer an issue, and the CPU usage was low (below 15% when both were running).


ALSO, ive reason to believe one of my hard drives could be going out, but im trying to see if its something else first because with RAID 5 enabled i cant just remove hard drives for testing.

Is there a program of some sort on the net that can allow me to test individual hard drives or a program that will let me know its my hard drives for sure? lots of things could have gone wrong and i dont want to replace things that dont need replacing
Just for reference reasons my build is:
ASUS RAMPAGE II GENE
2x EVGA GTX 280
Core I7-920 @ 2.66 ghz (watercooled with corsair H50)
3x Caviar black 500 gb HDDs in raid 5
1x dvd burner
910 watt PC power and cooling PSU
3x 2 gig sticks of ocz gold ram @ 1600 mhz

 

[tcmike]

update:

ok so i went further into my dilemma and i decided i would take apart my computer and put it back together to see if anything was loosened and my results are as following:

using different fans, i found that the CPU fan plugin slot on the mobo is gone (i dont see any sign of physical wear or burn marks so im not sure whats going on there). This plugin slot was powering my radiator's fan, but thankfully my case has so many fans in it that it didnt ever overheat my CPU last night or today (stayed below 45C).

Upon removing my 280s, i found that one of my hard drive cables was half in the connector and half out (the asus rampage ii gene has very loose plugins for the sata cables), so i decided to replace the cable itself and plugged it in again and i put a small amount of wood glue to hold them together so it doesnt happen again.

nothing else out of the ordinary.


booted computer up, noticed a lag logging into the computer (sat at the welcome screen for minutes on end) and programs started responding faster, but they were still having issues.. but im going to just go on a limb and say that my RAID is probably recovering at the moment due to that flimsy cable.. so maybe it fixes itself but if it doesnt then im not sure what ill do and i DEFINATELY NEED SUGGESTIONS ON WHAT TO DO.



next test: replacing the hard drive that had the cable half out with an old one and letting the other hard drives ghost over to see if its the hard drive..

not sure how to test if CPU is doing crappy

 

[jpmucha]

Maybe something shook loose on the way to the LAN party.

Would entering your RAID bios give you a health report for the RAID? It should tell you if you're running in a degraded state. If the one drive dropped out of the RAID, it might be trying to rebuild the 'lost' drive with the data/parity info from the other drives. RAID5 should be able to survive the single drive loss... although the rebuild will take hours (or days)... if that's what its doing. It might just be a matter of letting the thing do its thing overnight.

 

[tcmike]

yeah upon boot up it shows the health of RAID, and its been healthy all this time, which is why i posted this under CPU, but it could be a drive going out.. im going to leave it running overnight incase though!

Though i really hope its not the i7.. thats just too much money haha !

 

[Pc Guru_07]

Slow boot and file transfer means that hdd is failing.How many hdd are connected to your pc?Start testing one hdd at a time.Download hd tune and check smart health.Did you apply crazy voltage when you overclocked?When you overclock you should always increase the speed slowly step by step.

 

[tcmike]

Yeah, the file transfer went up from going 5 mb/s to 10 mb/s since i plugged it back it, but normally its around 4x that.

Very nice program, thanks for giving me the info!

and no it was just one tick of(which i believe is .05 volt increase?) volt increase and its been watercooled all this time so i think it would be fine because i looked up guide to overclocking it and applied it and ive overclocked it before quite a few times just to play with it (enthusiast haha).

HD Tune Pro: Intel Raid 5 Volume Benchmark

Test capacity: full

Read transfer rate
Transfer Rate Minimum : 0.8 MB/s
Transfer Rate Maximum : 106.3 MB/s
Transfer Rate Average : 49.2 MB/s
Access Time : 22.3 ms
Burst Rate : 114.4 MB/s
CPU Usage : 3.2%

HD Tune Pro: Intel Raid 5 Volume Error Scan

Scanned data : 1000 gB
Damaged Blocks : 0.0 %
Elapsed Time : 1:26


AVERAGE WRITE SPEED : 2.8 mb/s (both of these are odd because my laptop HDD that isnt raided out performed my raid on all categories.. which tells me that one or more are definately dying.. but itll take me ages to find out which one it is because RAID 5 doesnt allow me to remove hard drives and test them on their own without having to repartition each individual one and install windows before i can find out.(unless theres a way i dont know about which would be much appreciated to know)

AVERAGE READ SPEED: 52.2 mb/s

for some unknown reason, my hard drives arent SMART featured (though they are only like 4 years old) so i cant tell the future failure.. is there another way to tell?

 

[Pc Guru_07]

Woh! 4 years old hdd.Don't you think its about time that you change them.Cavair black supports smart.Which model you have?Try testing one hdd at a time.

 

[The_OGS]

for some unknown reason, my hard drives arent SMART featured

Well of course, your BIOS will default to SMART=disabled because that's the way Asus (and my mobo, Gigabyte) think it should be.
I always manually enable SMART in BIOS, usually chuckling and shaking my head at the time...
Apparently SMART can cause a tiny, almost undetectable performance penalty in some circumstances.
And because the world is benchmark-crazy and mobo manufacturers are ultra-competitive, SMART is disabled by default.
I run a Seagate 500GB and two Caviar Blacks (500 and 620GB) with my own i7-920.
You should probably use the ASUS RAID hardware SATA connectors (not the Intel ICH10R) anyway.
I suspect many answers could be found within your BIOS setup...
My performance is totally fine without a RAID stripe, and I have zero urge to run one.
A RAID mirror will also boost performance (which most people seem to forget) while of course providing data security.
Maybe you should mirror two of your 500s, and simply leave the third one alone to install your OS.
Then of course, relocate your My Documents (WinXP) or what Win7 calls your Libraries, and anything else important, to the mirrored volume.
I'm certain your performance in this configuration would be sufficiently brilliant... and much more reliable!
Good luck,
Regards

 

[tcmike]

ok so i was going insane for the past hour trying to find SMART in the bios and found nothing so i checked raid controller and other things for options and found nothing. searched manual 3x and finally found something showing SMART under [Main] on bios but its only appliable for single hard drives (in raid, smart isnt reachable). So i decided to boot up the PC, go to western Digitals website and found a hard drive diagnostic tool and ran it. The test ended after 30 seconds of running (didnt even reach 3% complete) and an error popped up saying ' too many bad sectors'. So it looks like im going to be tearing this computer apart and finding out who the culprit it. Its too bad, because i really liked raid!

thanks you guys for the help on my issue, i appreciate it! all good tips and definately helped expand my knowledge some!

 

[jpmucha]

I was going to follow up suggesting the WD diag tool
I haven't run it on a RAID system, but with single drive it will let you pick a drive at a time to run the non-destructive tests on. Does it let you test a drive at a time, or is it just the complete RAID volume?

Anyway, with the 3x500GB drive RAID5, you can pick up a 1 or 2TB drive for $60-80 and do a bulk back-up of the whole array. RAID5 can survive a single drive failure and run in a degraded mode, but I read statistical studies that describes a low (but still significant) probability that the rebuild could have a problem, especially as the volume sizes increase. A backup of your data as soon as you can probably would be wise.