Heads up: ASUS Suite's Fan Xpert 2 decided not to kick in...high CPU temp ;o(

[Noel 4,501]

I thought I had the back up BIOS fan settings in a safe zone, but apparently not.  For some odd reason it decided not to take command of fan control, the profile changed all my itself, and in the midst of a long flight session I notice very warm air coming out of my typically very well cooled HAFx case.  I have been able to run my SB-E at 4.55Ghz at 133v prior to this event, where core temps got up in the 80-85 range for I don't know how long.  The CPU fan was coasting along at the Quiet mode setting of 200-300rpms!!  When I finally noticed this I looked at Fan Xpert, selected my former Custom mode, and temps dropped right back down.  Now if I try to run my CPU at 4.55 she freezes at 1.33v, but fortunately runs steady in Prime95 torture at 1.335v, but I have to think its 'heat treatment' did some damage.  Hopefully it isn't on a more rapid path to degradation as it's been a phenomenal CPU so far.  Normal temps at 4.55Ghz is 52-58C.

 

Just a heads up for you to make sure you are aware it's possible for Fan Xpert NOT to use it's selected profile, somehow some way.  I think maybe what happened was I had recently done a slight BIOS setting change, and this must have interacted w/ Fan Xpert to reset its default profile, unbeknownst to me.  

[doog79 43]

Happened to me after updating the BIOS and re doing my overclock.  Everything was fine until I tried a flight in FSX then it blue screened.  Turned out Fan Xpert needed to be set back to how I had it. 

 

Now I leave FanXpert in Turbo mode which controls my Noctua NHD14 cooler and 2 of my exhaust fans.  The other 5 case fans 3 intake on the front and 2 large exhaust fans on the top are controlled by a NZXT digital fan controller I installed a few months ago.  This made a quite a difference with noise and temps. 

 

My box stays way cooler than it used to which is an AntecDF85.  I also upgraded to Antec true quiets on the front intakes and Noctua fans n the exhaust while installing the controller.

 

Doogie

[Noel 4,501]

 

 

Happened to me after updating the BIOS and re doing my overclock. 

 

OK, that nails it pretty much.  I have had a totally stable, cool overclock since last June at 4.3Ghz, then 4.423@1.315v a month or so ago, then to 4.55@1.33 and completed several long flights w/ no problems.  These over clocks were done thru Asus Suite II.  I decided to play w/  a different approach to overclocking in the BIOS, and bam, didn't notice the blithering CPU fan control didn't kick in.  If I go to 4.55v@1.335v now Prime95 is error free though I haven't tried P3DV2 yet, but this is very clearly a new development and must have had some degradation from that [$(*%U@#$()%U@(%*U] session where I didn't realize how a BIOS change would affect Fan Xpert profile loading.  I hope others pay attention here as it's rather disheartening.  I did complete a 1h flight at 4.42@1.315 which was fine, so hopefully the degradation wasn't too bad, but still!  I know I've had a very friendly chip to work with.

[martin-w 3,235]

where core temps got up in the 80-85 range for I don't know how long.

To be honest Noel, if you are saying that 80-85 worries you, and may have caused degradation... I very much doubt it, 85 is still well below the max. The CPU won't even be throttling back at that temperature.

 

That's if you are referring to core temperature of course, and not CPU temperature.

 

If you used Ai Suite to overclock, and it was stable, and are now unstable with your own UEFI settings, then I would say that Ai Suite was using different settings to you, and this is where the issue lies.

[Noel 4,501]

To be honest Noel, if you are saying that 80-85 worries you, and may have caused degradation... I very much doubt it, 85 is still well below the max. The CPU won't even be throttling back at that temperature.

 

That's if you are referring to core temperature of course, and not CPU temperature.

 

If you used Ai Suite to overclock, and it was stable, and are now unstable with your own UEFI settings, then I would say that Ai Suite was using different settings to you, and this is where the issue lies.

Hi Martin!

 

Yes I guess you're right--I think TJmax is 105 on this chip, but used to under 60 under any conditions.   But I am using the identical method to overclock so this does not explain why it won't go now do 4.55Gh at 1.330.  Perhaps this is a red herring then and perhaps it was just on the edge of not making that clock speed at that voltage anyway. It is odd though in that when this occurred it was immediately after that, and the the next day when I attempted again, to no longer be doable.  However...adding a 5/1000th more and I think it's stable once again, i.e. 1.335 though haven't flown at the speed yet, just got thru Prime95 for about 15min.  4.42 ver 4.55 is not worth the fretting though so either way we're good to go!

 

The discussion about DX12 & 64 bit is sounding like FS Nirvana so it will be fun to see how that ensues.  It appears DX12 has the GPU exploiting multi-core much better--yeah!  P3DV2 continues to improve...

[martin-w 3,235]

To be honest Noel, the behaviour of out silicone chips is frequently baffling.

 

I was having an overclock issue. Thought I'd found the issue, thought it was the XMP setting that was excessive, Increased RAM voltage, tweaked the timings, all seemed well.

 

Nope, now getting BSOD's, Event ID 19, which I believe is voltage. That's despite the fact that my voltage is high.

 

This has been a nightmare of a chip to overclock. I'm coming to the conclusion that anything over a small overclock is impossible for my dodgy 3770K.

 

The chip doesn't overclock and coupled with that the IMC isn't good.

[Noel 4,501]

To be honest Noel, the behaviour of out silicone chips is frequently baffling.

 

I was having an overclock issue. Thought I'd found the issue, thought it was the XMP setting that was excessive, Increased RAM voltage, tweaked the timings, all seemed well.

 

Nope, now getting BSOD's, Event ID 19, which I believe is voltage. That's despite the fact that my voltage is high.

 

This has been a nightmare of a chip to overclock. I'm coming to the conclusion that anything over a small overclock is impossible for my dodgy 3770K.

 

The chip doesn't overclock and coupled with that the IMC isn't good.

That's too bad Martin.  OTOH it gives you some years to get ready for what seems surely to be something to build for, and that would be DirectX 12 w/ 64-bit in P3D or XP.  It sounds like there is significant exploitation of multicore CPU as well by the DX12 API that the best of all worlds will be hexacore or more Intel CPU AND...save up for Titan Black!  The benefit there is I believe you're looking at 12Gb of video ram along side dual GPU potency versus SLI'd use you only get to half of the VRAM total from both boards.  Coming in at a mere $3000 I'm sure we'll all go right out and get one!  I'd do it if the performance exploited 12gb of video ram properly and there was a 64bit simulator to use it in.  It could in the P3D world since they are building for commercial application so they may build in that sort of scalability that might not be considered affordable for the commoners.  Seems best to always build in maximum scalability unless it's too costly to code for.

[martin-w 3,235]

Interesting stuff Noel.

 

As an update, I'm now clocked back to 4.2 and all seems okay...for now, time will tell.

 

I don't mind if that's all it will do, as we have said before it makes a miniscule difference in frame rate.