Brand new Haswell 4790k not booting at once bootable voltages

Hey all, ive surfed the forum a few times, and im part of the team over at hwbot. Im having a weird thing happening with my brand new cpu. When i say new, 2 days new... but anyway on to the question. So for instance, I've stablized and benched my cpu at 4.7 Ghz 1.285v with XTU, so within XTU i would bump the clock up another multiplier to ensure that its not stable at 4.8 with the same voltage and as you guessed it BSOD. So the computer goes to cycle on again but now i cannot boot at 4.7 1.285...the benchmark and stress test both passed, but if i bump to 1.295 it comes back up. Then i tried to lower it back to 1.285 via XTU and it crashes. So my once stable sweet voltage has become a thing of the past. And it is consistantly happening. the 4.7 was finally stable again at 1.32, Its like they are one time use. 4.9 @1.39v was a thing, now it has bsod'd until it is finally stable at 1.44. With adia64 i noticed that my cache was always the cause of a stressing failure within like first 8 seconds until i did the multi auto in bios and 1.2v trick. I finally got tired of bsods and moved to tweaking that.

I haven't done a CMOS reset yet because its a pain on my board. Ive got the maximus formula vii, water cooled both the VRM and CPU, so i would have to drain the loop to remove the plastic cover of the board to access the battery. Im not positive of the bios version ill check when i get home. Any ideas on why im losing these milestones? Or how to fix them? Thanks everyone.

My full build is:
Intel i7 4790k
Asus maximus vii formula
Coolermaster glacer v2 240l (expanded loop)
Asus strix gtx 970
16gb corsair dominator platinum 1866 cl9
Corsair ax860
120 pny ssd

Im on mobile right now so sorry for any misspellings or formating errors.

Make sure other bios settings didn't change on the original BSOD, such as RAM frequency and voltage, cache ratio, cache voltage, system agent voltage, etc.

You also might consider resetting the CMOS and starting from scratch.

trents said:

Make sure other bios settings didn't change on the original BSOD, such as RAM frequency and voltage, cache ratio, cache voltage, system agent voltage, etc.

You also might consider resetting the CMOS and starting from scratch.

Click to expand...

The other settings did stay the same, i probably will just drain and start fresh, im fairly certain a bios update is in order as well. Atleast i wrote them down so i can try at my lower voltages again. Thanks.

TuKr said:

The other settings did stay the same, i probably will just drain and start fresh, im fairly certain a bios update is in order as well. Atleast i wrote them down so i can try at my lower voltages again. Thanks.

Click to expand...

First off if you update the BIOS don't be surprised if previous "stable" Overclocks are not stable with the new BIOS.

I'm just trying to get a clear picture of how you are going about the overclocking process. Are you using the XTU utility and raising the multiplier or does XTU have some level up feature to Oc the Cpu? Also what are your temps like when stress testing? These I7 47xx chips can run really hot when Oced. Are you leaving the voltage on Auto or is it set manually?

Mandrake4565 said:

First off if you update the BIOS don't be surprised if previous "stable" Overclocks are not stable with the new BIOS.

I'm just trying to get a clear picture of how you are going about the overclocking process. Are you using the XTU utility and raising the multiplier or does XTU have some level up feature to Oc the Cpu? Also what are your temps like when stress testing? These I7 47xx chips can run really hot when Oced. Are you leaving the voltage on Auto or is it set manually?

Click to expand...

When I start I go in to BIOS and set XMP and turn off the Powersaving features(Cstates, VR efficiency, and speedstep) then let it boot normally to windows, I open up XTU and see what the core states, Voltage, multiplier, temp. i run the XTU stress for 5-15 minutes, if it passes and temps are within reasonable, i run the benchmark, if it passes i save the profile with the settings as the name(44-1.20). then i up the mulitplier by one, and repeat. Stress, if good, bench, save, up multi again. if it fails i usually just boot back in, load the most recent profile, up the voltage by a tick or 2 depending on how quickly it bsod'd and stress again. what usually ends up happening is i get tired of not being able to reach the next multiplier with reliable scaling, but with this new 4790k, it wouldnt load the most recent successful pass. all of the Bios settings were saved and while i was there i figured hey why not try to load the last pass into bios and start there. well that didnt work either. the temps really change with the voltage but usually below 1.35 i stay below 70c. I have never seen this chip hit 75c. I set voltage manually through XTU, static, no offset.

I know that the BIOS can effect the stability of any overclock that why im wondering if my current bios has anything to do with it, the board is only about a week or 2 old. I had a PSU, drop available voltage durring benches and it created instabilities and cause an obnoxios coil whine for my PCI lane. So i got a new 970, still had it, swapped PSU's it got better but was still present, then i couldnt keep my GPU stable on benches and the coil whine was still present, so because i had tried everything but a new mobo and cpu, i swapped the mobo, then felt as though the CPU had degraded incomparison to previous benches and just swapped that. The feeling of degradation could have been caused by the BIOS not being updated now that i think about it. I didnt even think to look at the Bios with the new board. maybe thats why i dont have offset features in bios either

How about flashing the bios back to the old version that was working better. Also, there usually is no need to turn off the power saving stuff in bios.

trents said:

How about flashing the bios back to the old version that was working better. Also, there usually is no need to turn off the power saving stuff in bios.

Click to expand...

just updated my previous post. See above, it will explain why i had not tried that lol.
Also i turn off the Power saving stuff because the edge of stable benches can be failed benches if the Cpu fails to respond everywhere fast enough.

Well if you guys can think of anything else, let me know. Im going to clear cmos and update the bios when i get home in about an hour. So maybe in about 2 to 3 hours ill drop am update.

Pour some tobasco sauce on the CPU while you are at it.

trents said:

Pour some tobasco sauce on the CPU while you are at it.

Click to expand...

But i dont wanna get it too hot

Turns out a bios update cured it. 4.8 @ 1.329v

Good to hear.