Jump to content

only one core fails? my attempt @ 3GHz...


Recommended Posts

well, since i got my flat panel in i thought i would shoot for an OCDB entry (since i can run 3dmark 06 at its native res now :D). well, im shooting for a 3GHz entry, so far its not looking good.

 

*Disclaimer*

In no way do i intend to remove my IHS, i have no way in crap a way to replace anything if it goes south

*End Disclaimer*

 

im doing 334x9 1.5v with my ram sitting on known good settings, so i have rooted that out. also, core temp shows one core WAY hotter than the other, here are some before and after images:

 

temps before (stable core0 loaded)

http://img.photobucket.com/albums/v206/cha...2911/before.jpg

 

temps after (unstable core1 right at fail)

http://img.photobucket.com/albums/v206/cha...22911/after.jpg

 

**took a couple of tries to catch the temp of core1 before core temp updated :D**

 

now i have a thermistor glued to the side of my IHS with AS adhesive, max temperature it has recorded is 41°C... a big difference from core temp, and also only 2°C off from bios readings. so guess which once i believe ;). anyhow, its weird the failing core also has a higher temp then the stable core, i have seen the failing core go as high as 78°C before, yet my trusty thermistor records a max of 41°C. so i took off the heatsink and double check everything, cleaned and reapplied AS:C... same results.

 

so to get to the point, do i have a lazy core or you think mabye something is up with the IHS, or maybe core temp is crap?

 

BTW, i realize my PSU is scraping by (plan to upgrade soon), but at the settings in my sig, im sure a fully loaded x1950xt @ 660MHz 1.475v core and 1008MHz ram, and a 2.88Ghz dualcore @ ~1.4v draws more than a 3Ghz dualcore @ 1.5v. but then maybe im delusional because its 2:16am and need to get my butt in the bed :D.

 

thanks for the time you take to reply guys :).

 

Charlie.

 

EDIT:

 

BTW, i just looked at my temps, thermistor reads 35°C while im idle, browsing in firefox @ 334x9 1.5v...

Share this post


Link to post
Share on other sites

If its failing super PI its mega miles away from being stable I could pass 32M with my Opty 165 @ 3.1 but as soon as you start Stress Prime 2004 Orthos edition it would crap out.

 

I never got my 165 to 3GHZ stable for long enough to get it in the OCDB I must have done 40+ runs & the best I could do was 6 hours 50ish Minutes & as you can see from my SIG I ended up running 2925 24hour stable then gave up & went C2D.

 

I think you need to slow her down a bit & do some heavy stability testing increasing FSB & voltage as you go until you find the CPU's 10-12 hour stable wall.

 

After 10 hours of Orthos both cores will be bang on with each other its just when they are cold/cool they are 10-14oC out.

 

I believed Smartguardian with my Opty because it read lower temps now I chose to believe Core temp with my C2D rig for the same reason LOL.

 

I think 50-55oC in Smartguardian was the sweet spot for my 165 I might of made 3GHZ 8 hours stable if I was using a better cooler or water but because my Opty 144 did 3GHZ on the stock cooler I figured the 165 should too.

 

BTW

 

I was getting 55oC full load @ 2925MHZ with 1.5v on the stock cooler.

Share this post


Link to post
Share on other sites

I still think it is your PSU holding you back at this point... Especially with the 1950xt sucking down juice.

 

Although, looking at your temps in Coretemp, it is more then the PSU holding you back, it is also your cooling. What are you using to cool that chip anyway?

 

What are you temps as reported by smartgaurdian?

 

I agree with Chump, if you are failing SuperPi, you are nowhere near stable.

Share this post


Link to post
Share on other sites

I still think it is your PSU holding you back at this point... Especially with the 1950xt sucking down juice.

 

Although, looking at your temps in Coretemp, it is more then the PSU holding you back, it is also your cooling. What are you using to cool that chip anyway?

 

What are you temps as reported by smartgaurdian?

 

I agree with Chump, if you are failing SuperPi, you are nowhere near stable.

 

im on the stock heatpipe (same as FX line) cooler modded with an 80mm fan (from 70mm).

 

i lowered my OC to 324x9 1.45v and she went 3 hours in orthos before crapping out. would secondary timings play a role in this still? since i got this new ram (uses hynx CTP-D43 chips) im not sure what to set, so its on auto. anyhow, i bumped the vcore to 1.475v and am reruning orthos (about 15mins through at post time).

 

about the temps, here is Smart Gurdian wile going through orthos:

http://img.photobucket.com/albums/v206/charlie22911/sg.jpg

 

will post some more as soon as i get back from work... thanks guys

Share this post


Link to post
Share on other sites

Well, I definitely think it is your temps that prevent you from getting the higher clocks stable.

 

A new HSF will probably let you get a higher OC before a new PSU. I'm not saying you don't NEED a new PSU - you definitely do, but with your temps where they are, you won't get the OC stabilized even with a new PSU.

 

I say get a new PSU (600w+) & HSF - Turniq tower, TT BigTyphoon, etc.)

 

 

Seeing your Coretemps & SG temps gives you a really good idea of how far off your thermistor is...

Share this post


Link to post
Share on other sites

im on the stock heatpipe (same as FX line) cooler modded with an 80mm fan (from 70mm).

 

i lowered my OC to 324x9 1.45v and she went 3 hours in orthos before crapping out. would secondary timings play a role in this still? since i got this new ram (uses hynx CTP-D43 chips) im not sure what to set, so its on auto. anyhow, i bumped the vcore to 1.475v and am reruning orthos (about 15mins through at post time).

 

about the temps, here is Smart Gurdian wile going through orthos:

http://img.photobucket.com/albums/v206/charlie22911/sg.jpg

 

will post some more as soon as i get back from work... thanks guys

 

I think the 3 hour crap out is CPU voltage which you have just upped the problem is the more volts you give it the hotter it gets & the less it wants to overclock.

 

My PWMIC used to get hot like that when I ran with the side off my case I would put it back on to let the air flow through your case.

 

If the side is on get some case cooling on it it might bring down the CPU temps.

 

Optys are a little funny with RAM but I don’t think they are holding you back yet.

 

I always find the max fsb of my RAM before I start the CPU overclock.

Share this post


Link to post
Share on other sites

I didn't even notice your PWMIC or chipset temps.... They are out of control too.

 

Well, not so much your chipset temp, that is pretty normal for the stock chipset HSF.

 

 

But the PWMIC @ 68c :eek2:

 

What happens when you take the door off your case?

Share this post


Link to post
Share on other sites

Charlie what's the steeping of that opty??

 

its a CCBBE 0616XPMW :D

 

 

Well, I definitely think it is your temps that prevent you from getting the higher clocks stable.

 

A new HSF will probably let you get a higher OC before a new PSU. I'm not saying you don't NEED a new PSU - you definitely do, but with your temps where they are, you won't get the OC stabilized even with a new PSU.

 

I say get a new PSU (600w+) & HSF - Turniq tower, TT BigTyphoon, etc.)

 

 

Seeing your Coretemps & SG temps gives you a really good idea of how far off your thermistor is...

 

well, guess i installed that thermistor for nothing :(. so you do recomend a HSF then? is so i will be ordering that big typhoon today. and mabye a PSU next month if my pay is big enough.

 

I think the 3 hour crap out is CPU voltage which you have just upped the problem is the more volts you give it the hotter it gets & the less it wants to overclock.

 

My PWMIC used to get hot like that when I ran with the side off my case I would put it back on to let the air flow through your case.

 

If the side is on get some case cooling on it it might bring down the CPU temps.

 

Optys are a little funny with RAM but I don’t think they are holding you back yet.

 

I always find the max fsb of my RAM before I start the CPU overclock.

 

i have the max overclock of my ram (225MHz @ 2.5-3-3-6 @ 2.6v). timings, or voltage and any combination of the two dont increase its OC either. the ram is 2x1GB Hynix CTP-D43 BTW.

 

i have a small fan that fits the PWMIC area perfectly, just looking for a way of attaching it... my side is on BTW.

 

I didn't even notice your PWMIC or chipset temps.... They are out of control too.

 

Well, not so much your chipset temp, that is pretty normal for the stock chipset HSF.

 

 

But the PWMIC @ 68c :eek2:

 

What happens when you take the door off your case?

 

the evercool has been out of stock for a while now... any other recommendations?

but about the PWMIC, she runs a little hotter with the side off. the entire front of this coolermaster centurion5 case is like a big vent, with a 120mm exaust... it works VERY well :). i have a small fan im going to attach to the PWMIC heatsinks, did a test with bread ties and it works well, just need some superglue to make it "offical" :D.

 

im still testing some stuff, im thinking i just may drop back down to the settings at my sig for an OCDB entry until i get a better PSU and HSF to work with. hopefully i have been blessed with a good overclocker :).

 

 

thanks for the replies guys! keep'em coming :D.

 

Charlie

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...