Problems overclocking 4770k

C

Cam

n00b
Joined
Dec 7, 2005
Messages
17
Hi,

I seem to be having issues getting a mild overclock on my 4770k. Currently I have it running stable at 4.2GHz with the following changes done to BIOS. Everything else in BIOS are set to default.

4.2GHz Settings:

Multiplier 42x
Vcore: 1.18v
XMP profile Enabled

With the above settings, everything is stable (18+ hours AIDA)

Now, when I try pushing it to 4.4GHz everything is stable under load (100 rounds IBT, 12 hours AIDA), but once my computer starts idling (usually 15 minutes or more), or even when I am using it for light activity like browsing the Internet or watching YouTube, it will shut off completely randomly, and when my computer turns back on, BIOS gives me the following message:

“Previous overclock settings have failed”.

4.4GHz Settings:
Multiplier 44x
Vcore: 1.26v
VCCIN: 1.9

I have tried getting closer to a 1:1 clock/ring ratio and that didn't help. Nor did increasing the ring voltage. I have also tried adjusting the Vcore by increments of 0.01v all the way up to 1.30v, as well as increasing and decreasing the VCCIN voltage. Adjusting System Agent voltage also seems to have no effect, nor does disabling all C-States via BIOS. I am fairly certain that my Windows install is fine and is not causing the problem. Sometimes if I turn on the computer from a cold boot I get a black screen for 5 seconds before that message appears. I am running out of ideas. Do you guys have any ideas or on what it could be?


Thanks.

System specs:

Intel Core i7 4770k
MSI Z87-GD65 v1.9 BIOS.
8GB 1600 G.SKILL Sniper RAM (Cas 9)
Samsung 840 EVO 250GB
ASUS R9 280x
Corsair H100i
Thermaltake Chaser MK-1
Windows 8.1
 
What voltage mode are you using in BIOS?

I gave up on trying to get Intel's Adaptive and Offset modes working on an overclocked 4770k (forget what MSI calls them in BIOS). I use manual 24/7 and turn the power configuration in windows to High Performance, and drop the Minimum Processor state to 1%. Still drops vcore when idle, still drops clockspeed.
 
I'm using Override mode. Haven't attempted using adaptive mode or using offsets yet.
 
I guess try turning off speedstep? Kinda shooting in the dark a little but if it's having problems only under small loads... and it isn't c-states... try a small bump to Digital I/o and analog I/o along with the system agent bump, as well.
 
I tried 3 different sets of RAM and put them all in different slots. Idle reboot still happened.

I just disabled Speedstep. EIST is also now disabled as a result and I just had the same thing happen again, however, I just gave a slight bump in voltage to the Digital and Analog I/O so we'll see what happens.

[EDIT] Happened again. Even with the latest beta BIOS and with memory being configured manually instead of via an XMP profile.

I am also troubleshooting this issue with another friend. We are going to see if it's a power issue tomorrow by measuring the voltages with a multimeter, having a camera record the readings, and review the readings before it shuts down. I want to rule out if it is a power issue or not.
 
Last edited:
is it the case you tring to fiddle with too many things, haswell is very basic to overclock.

its just adjusting the vcore and turbo voltage.

According to asus the best way to go about it is adjust vcore manually whilst memory is on SPD settings, to find the stable cpu clock, then after that point use adaptive for that voltage and increase RAM clock.

There is 2 adaptive voltage settings the 2nd one was a bit confusion I experimented with it, got bSOD's at the windows login screen and went back to offset :D

The adaptive offset means when cpu is not in turbo mode it is using stock voltages (and should be also on stock clocks if only adjusting turbo multiplers). The 2nd adaptive voltage setting I believe is an actual value not an offset but again for turbo only mode, I leave the 2nd alone and only set the offset, I managed to get currently 4.3 clocks on a 1.13v voltage (which is 0.05 offset since default is 1.08).

My rig is defenitly stable, aisuite 3 annoyed me as it forced my rig to high performance on windows power settings (so the thing is changing my windows settings behind my back), I will uninstall aisuite3 asap when I find a monitoring app that can monitor my voltages/fans properly. If you have aisuite installed one has to wonder if its fiddling with o/c voltages etc behind your back. I was running at my 4.3 turbo clocks even when idle as aisuite had disabled all power saving stuff.

So I suggest uninstall aisuite.
Reset bios to defaults.
Bump bios vcore to 1.15v whether its via offset or standard
Try 42 turbo multipler
See if stable

Also do you have a haswell compatible psu? haswell drops to very low voltages on idle and some psu's cannot handle it.
 
Yeah, Haswell seems pretty simple to overclock especially for smaller overclocks. I only started to mess around with other settings once these reboot issues started happening...

I'm going to try messing around with using voltage offsets rather than setting voltages directly. Hopefully it will make a difference. I installed multiple BIOS's today to see if it was BIOS related but it appears it isn't. I even tried a different Power Supply (A SeaSonic 860 Platinum).

That is interesting regarding AIDA though. I'll keep that in mind. I was wondering why my voltages were running higher with it running.

Regarding your settings suggestion, I tried that and it's stable with a multiplier of 42x with a voltage of 1.17v (can't go any lower without BSOD).

SeaSonic also lists my PSU as fully Haswell ready.

Thanks for your ideas and suggestions. It's much appreciated and I plan on solving this mystery. :)
 
I tried using both, but unfortunately I have the same problem.
 
us guys seem to own cpu's that the community doesnt seem to like much :(

on forums across the net it seems about 80% use either sandy bridge or ivy bridge and about 10% haswell.
 
Bluesun311 said:
I've had several 4770k's that could do no better than 42x core no matter what I tried.
Click to expand...

Really? What motherboards have you used? Were you getting similar idle restart behavior?

I ask because I found This thread over on the MSI forum. This guy has the exact same issue as me and he fixed the issue with a motherboard swap. I am not exactly sure it's a CPU problem specifically especially since my CPU seems to be a (slightly below average, but not horrible) overclocker.

chrcoluk said:
us guys seem to own cpu's that the community doesnt seem to like much :(

on forums across the net it seems about 80% use either sandy bridge or ivy bridge and about 10% haswell.
Click to expand...

Yeah. Haswell is a bit wonky. It's definitely a silicon lottery with Haswell CPU's. But I suspect many people don't like it because of the heat (garbage TIM, that's why I de-lidded.) :D
 
Cam said:
Really? What motherboards have you used? Were you getting similar idle restart behavior?

I ask because I found This thread over on the MSI forum. This guy has the exact same issue as me and he fixed the issue with a motherboard swap. I am not exactly sure it's a CPU problem specifically especially since my CPU seems to be a (slightly below average, but not horrible) overclocker.
Click to expand...

On Haswell I've used an AsRock Extreme4 and a couple Asus Hero boards. I had 3 out of 8 4770k's I tested that topped out at 4.2GHz. Since then I've learned a few tricks that help get memory stable but I don't think those chips would have benefited. The idle reboots are suspicious about your settings, you don't mention your Cache voltage or multiplier. It sounds like when your CPU is idle the board isn't giving it enough juice. I'd be surprised if motherboard was at fault and not just a motherboard setting, since it works fine at 4.2GHz.
 
Cam said:
Really? What motherboards have you used? Were you getting similar idle restart behavior?

I ask because I found This thread over on the MSI forum. This guy has the exact same issue as me and he fixed the issue with a motherboard swap. I am not exactly sure it's a CPU problem specifically especially since my CPU seems to be a (slightly below average, but not horrible) overclocker.



Yeah. Haswell is a bit wonky. It's definitely a silicon lottery with Haswell CPU's. But I suspect many people don't like it because of the heat (garbage TIM, that's why I de-lidded.) :D
Click to expand...

doesnt surprise me on the board, z87 boards seem flaky.
 
Bluesun311 said:
On Haswell I've used an AsRock Extreme4 and a couple Asus Hero boards. I had 3 out of 8 4770k's I tested that topped out at 4.2GHz. Since then I've learned a few tricks that help get memory stable but I don't think those chips would have benefited. The idle reboots are suspicious about your settings, you don't mention your Cache voltage or multiplier. It sounds like when your CPU is idle the board isn't giving it enough juice. I'd be surprised if motherboard was at fault and not just a motherboard setting, since it works fine at 4.2GHz.
Click to expand...

well if overclocking via turbo multiplier then it means when idle the cpu is at stock meaning its instability at stock.

Also if using adaptive voltage then it means both voltage and clocks are stock when idle. Adaptive + turbo multiplier should mean zero added risk of instability when idle or low utilisation.

Voltage fluctuations can cause instability if the power supply is unstable or the board is faulty.
 
Update on my situation:

Okay, I seem to be stable. I've been running 4.5GHz @ 1.34v and +0.05 SA voltage for the past few days and I've had no issues so far. Currently I am baffled as to why it's stable, but it seems like upping the VCCIN to 1.9v was causing the instability issues? I read that it's a good idea to bump VCCIN after a certain point to help stabilize overclocks, which I was doing beyond 4.2GHz. I think that was the cause of my overclocking issues. Why would a VCCIN bump cause that?
 
I've read that VCCIN is somewhat connected with stability of the ringbus or cache multiplier.
 
You must log in or register to reply here.
Back
Top