[Techspot] BCLK overclocking non-K Intel Skylake CPUs is now possible, tested here

[VirtualLarry]

Newest BIOS (2.90) on my ASRock Extreme4 seems to have broken or disabled BCLK OCing...

It says it at the download screen..
1. Update CPU microcode to 0x76.
2. Remove SKY OC function.

Seems up to the user if they want to download/use it, but you can always go back if you wish to bclk overclock.

That's a little scary. Just the other night, I Internet-Flash updated one of my Z170 Pro4S boards, to 3.00. Which was listed on their site as "Microcode Update x74", I believe.

But when it boots, it says "Press X for SKY OC Boost function", and BLCK OC is active.

I'll have to give their web site a new look-see today, and see if that version has been pulled too.



Edit: Thankfully, 3.00 was still the newest. I updated my other rig, and then after it booted, went into the UEFI and did the BCLK OC again. Still working. Woot!

 

[VirtualLarry]

I'm sure Intel has done the math and sees no upside to letting BCLK OC to continue. But for my part, this little maneuver has greatly diminished the desire to spend money on new PC hardware anymore.

The newest BIOS is a poison pill. Reverting to an earlier version does not restore BCLK OC functionality.

It's certainly reduced my desire to spend money on Intel hardware anymore, that's for certain.

I can't wait for AM4 and eventually, Zen!

This is a prime example of Intel killing innovation. (Innovation by the mobo makers.)

 

[crashtech]

I've switched between Gigabyte's P95 fix bios and back to the prior OC bios on my z170x-UD3 no problem (and OC'd my i3 6100 afterwards).

Lucky you, on the ASRock Z170 Extreme+, installing BIOS rev. 2.90, then reverting to either 2.33 or 2.40 (both successful OCing BIOSes) results in boot failure when BCLK adjustments are attempted.

 

[crashtech]

It's certainly reduced my desire to spend money on Intel hardware anymore, that's for certain.

I can't wait for AM4 and eventually, Zen!

This is a prime example of Intel killing innovation. (Innovation by the mobo makers.)

Sure, I'll be keeping my eye on AMD, if we are lucky they might be able to bring some excitement back to the PC hobby. As it stands, my enthusiasm is at an all time low.

 

[deasd]

Finally, got blocked, instead of improving BCLK OC. Giving a reason to me to wait for AMD......

 

[Dufus]

Lucky you, on the ASRock Z170 Extreme+, installing BIOS rev. 2.90, then reverting to either 2.33 or 2.40 (both successful OCing BIOSes) results in boot failure when BCLK adjustments are attempted.

Do you still show microcode update version 0x76 ? Can be checked with HWiNFO for example.

 

[DrMrLordX]

This is a prime example of Intel killing innovation. (Innovation by the mobo makers.)

*sigh* gonna have to agree with you. The Haifa team rather-obviously went to the trouble of setting up Skylake to be bclk-overclockable, and Intel played along . . . for the k-parts only. So mobo manufacturers found an (admittedly somewhat-flawed) way to get the non-k parts to bclk-overclock, and what do you know, Intel killed it, despite all the assurances that Intel was behind it 100%, at least for Z170.

Sad, really. Overclocked i3-6100s were a great idea.

Here's hoping AM4's Promontory can hold up to bclk overclocking.

 

[VirtualLarry]

Here's hoping AM4's Promontory can hold up to bclk overclocking.

I'm honestly surprised that given the issues with AM1 BCLK OC, that AMD hasn't released an unlocked-multiplier part for AM1.

 

[DrMrLordX]

AM1 will never get the love it could have gotten, like a full raft of Mullins/Carrizo-L products. Yeah they wouldn't have been much different than the 5350, but at least they would have had better clockspeeds etc.

I'm surprised AMD keeps releasing new platforms without proper bus locks so bclk won't hose SATA and PCIe functionality so much. Are they really that scared of non-k overclocks?

I can understand Intel's position, even if I don't like it very much. Skylake OC probably opened up a whole can o' worms wrt the shady SE Asian OEMs. Er well maybe it did, I don't know. Just a guess.

 

[crashtech]

Do you still show microcode update version 0x76 ? Can be checked with HWiNFO for example.

It says microcode revision 49 (0x31). Yet, BCLK OC is broken nonetheless, even though I rolled back the BIOS revision to 2.40.

 

[Burpo]

Maybe sell the board on ebay & replace, only not do latest flash with new board? Cost is minimal & get you back to where you were..

 

[crashtech]

Unless I get a really new board, and it has it 2.90 pre-installed. But it's looking more like BCLK OC is a pipe dream. Not only is Intel actively trying to quash it, there are things that it breaks, temp reporting chief among them, that make it a solution with many serious compromises. My current plan is to sell it all off in favor of a used X99 system.

 

[Madpacket]

Another reason to stick with 83xx FX cheap combo's for entry level systems. Intel's greed getting to them.

 

[hhhd1]

Not only is Intel actively trying to quash it, there are things that it breaks, temp reporting chief among them, that make it a solution with many serious compromises.

Looks like it is not stable enough when overclocked ?
Maybe there are more long term issues ?

It maybe Intel/OEMs are trying to protect themselves by avoiding warranty claims.

 

[crashtech]

Looks like it is not stable enough when overclocked ?
Maybe there are more long term issues ?

It maybe Intel/OEMs are trying to protect themselves by avoiding warranty claims.

I didn't have any problems with stability at all, the issue is that users with inadequate cooling could conceivably have some serious problems.

 

[Dufus]

Not only is Intel actively trying to quash it, there are things that it breaks, temp reporting chief among them, that make it a solution with many serious compromises. My current plan is to sell it all off in favor of a used X99 system.

Doesn't PECI temperature still work? Question is does thermal management still work.

 

[crashtech]

I can't test anything now, since I can't OC.

 

[Dufus]

Maybe I'll have a look on Haswell. I don't have Skylake but one can disable PM to probably achieve almost the same. Of course no huge bclks without coarse ratio's.

It's strange reverting didn't bring back overclocking. If you have a proper backup of a previously working BIOS version maybe try writing that, preferably externally.

 

[The Stilt]

I'm surprised AMD keeps releasing new platforms without proper bus locks so bclk won't hose SATA and PCIe functionality so much. Are they really that scared of non-k overclocks?

It simplifies the design significantly, thats the only reason.

 

[Dufus]

@crashtech Here's a G3258 with PM disabled resulting in loss of power readings, core temperature readings, EIST, C-States higher than 1 and so on however PECI temperatures still work.

The RealTemp readings are actually invalid as reported by the CPU registers used by it.

Unfortunately another side effect was being stuck at the LFM (8x), will have to take another look at that sometime and maybe use the opportunity to take some external IHS temp readings as they should be pretty close to core temp at low power. With the heatsink pulled off temperatures took a little while to build up but seemed to accelerate once hotter. Usually ACPI would kick in and shut things down. I didn't see any throttling flags before the computer shutdown with the last reading showing 125C. No reason in the Windows event log, just an unexpected shutdown.

My best guess at this time would be thermal management control doesn't work and one would have to rely on catastrophic temperature shutdown at ~125C-130C and hope thermal runaway doesn't occur.

 

[crashtech]

So, almost worst case then, but presumably it will shut down before permanent harm is done, though 130°C is scary hot.

 

[Dufus]

Hopefully, I know it wasn't guaranteed in the older processors but maybe the last several generations are better but would one want to be running at just under that, say 120C, for any length of time.

FYI HWiNFO will report those PECI temperatures while correctly not showing the DTS ones. This pic below shows both DTS and PECI temperatures but with power management not initialized in Skylake non-k OC DTS will not be shown as those CPU registers don't have the valid data flag set.

 

[cainn24]

Unfortunately another side effect was being stuck at the LFM (8x)

That happens to me if I have the "Boot Performance Mode" option in the BIOS set to "Max Battery", which is only supposed to enforce the LFM until the OS has booted. But since it gets stuck there (presumably because there is no longer any facility for the CPU to dynamically adjust the multiplier since that would be controlled by the disabled PCU) I can only assume that this is another "bug". The problem goes away if I use either of the other settings (Turbo, Max Non-turbo).

 

[inf64]

Also as we saw from hardware.fr testing, blck OCed Skylake has some throttling issues under some workloads so not all is well when you do OC the locked models(apart from breaking temp reporting).

Just buy the K model, it is not that expensive. Funny thing is the Haswell has gone up in price recently, not by much but still has.

 

[crashtech]

...Just buy the K model, it is not that expensive...

At the time I bought an i7-6700, the "K" model was $100 more. I found the difference to be significant.