Can't overclock i7-5775C on ASRock Z97 Extreme9

[Galatian]

Hey all,

I'm having problems overclocking the i7-5775C on my ASRock Z97 Extreme9. I ran the i7-4790K @ 4,6 GHz before, so I don't think I'm doing something obvious completely wrong. The overclock is actually applied, but when running prime I can see that the processors tried to downclock to stay within it'S 65W TDP envelope. I have never seen such a behavior on an unlocked processor. ASRock hasn't been very helpful so far. I have tried out 3 different BIOS version, one of which has been sent to me from ASRock directly, which doesn't even show up on their website. Does anybody know what I'm doing wrong?

Best

 

[Burpo]

"Make sure to disable C-states as power throttling will kick in aggressively and lower clocks dynamically even if temperatures are far from thermal throttling level."

"Getting 4.3GHz even for validation takes a bit more effort than in Haswell and I found that additional VRin voltage is necessary once over 4GHz,"

http://www.modders-inc.com/intel-core-i7-5775c-review-meets-eye/6/

BTW, AsRock has been VERY helpful if they're sending you BIOS updates for better overclocking. The rest is up to you..

 

[Galatian]

"Make sure to disable C-states as power throttling will kick in aggressively and lower clocks dynamically even if temperatures are far from thermal throttling level."

"Getting 4.3GHz even for validation takes a bit more effort than in Haswell and I found that additional VRin voltage is necessary once over 4GHz,"

http://www.modders-inc.com/intel-core-i7-5775c-review-meets-eye/6/

BTW, AsRock has been VERY helpful if they're sending you BIOS updates for better overclocking. The rest is up to you..

That didn't work at all (nor should it, since the C-States are for lower power operation and have nothing to with the CPU throttling in order for it to stay within a given power limit). What needs no be adjusted is said power limit as explained here at MaximumPC.com:

"The problems with overclocking continued, however, as the BIOS doesn’t properly override certain safety features. Intel CPUs have the ability to exceed their TDP by a small amount for a period of time. Not surprisingly, 4.2GHz at 1.36V ends up requiring more than 65W—around 80W according to our monitoring software—and after 10–15 seconds, the power protection would kick in and the clocks would drop to around 3.3–3.6GHz. Normally, you can adjust the power limit, current limit, and even the amount of time you’re able to exceed the power limit, but the BIOS on this motherboard doesn’t appear to have the right microcode updates yet, so that didn’t work. Good news, however: We found a “current offset” feature in the BIOS that appears to trick the CPU into thinking it’s using about one third as much power, so we’re back in business. As with the first board, the manufacturer is working on a BIOS update that should fix the problems we encountered with overclocking."

I have such a "current offset" as well:

Unfortunately changing them to something reasonable like 150W or the highest possible value of 1000W and 1000A respectively does not seem to work either. The Long Duration maintained can only be set up to Auto or 58s at most. But alas none seem to work anyway; it does not override the settings provided by the microcode from the processor.

BTW: ASRock sending me some "fishy" BIOS that they don't even show on their support page, which could potentially brick my motherboard, so that I have to completely remove it (which is a days work with all the water-cooling installed) is not superb service in my book.

 

[Galatian]

Oh and a little update: using the Intel Extreme Tuning utility I can actually raise the wattage and current limit, so it would appear that the option ASRock implemented is at fault, because it really doesn't override the settings like the XTU does.

 

[Ansau]

ASRock is the only brand that hasn't updated the first bios supporting Broadwell. Asus, Gigabyte and MSI have all launched updated bios with improved support.
So this must be a purely issue with the bios, specially when IETU allows you to overclock without that problem.

Btw, I'm also interested in this chip. How far have you oced it?

PD : Most MB brands recommend to install latest intel igpu drivers for full support.

 

[Galatian]

Well just started...But it seems like it is doing 4,0 GHz at it standard 1,25V right away...will post later how much I can gain.

 

[ctk1981]

Good luck, 4ghz was easy for me to obtain but mine hits a brick wall around 4.3 and requires quite a bit of voltage to stabilize there.

 

[Galatian]

Well I was stable at 4.1 GHz with merely 1.3 V (that's 28.7/AVX stable though). For the love of god I can't seem to get 4.2 working though...any tips?

 

[Ansau]

Well I was stable at 4.1 GHz with merely 1.3 V (that's 28.7/AVX stable though). For the love of god I can't seem to get 4.2 working though...any tips?

Don't use prime95 to test overclock stability. It puts unnecessary pressure to the cpu.
Try OCCT Large Data Set, Realbench, x264 benchmark, rendering long video...

 

[Galatian]

Isn't that the general idea? I mean putting so much pressure on it and see if it can handle it?

 

[Kenmitch]

Isn't that the general idea? I mean putting so much pressure on it and see if it can handle it?

It would really depend on how your going to use your rig in the end.

What is your intended use for the rig?

 

[Galatian]

Does it matter? I want it stable (at least in the sense stable under the highest pressure I know off). I know the processors passes Prime at stock so I want it to pass over clocked as well. Am I perhaps giving up performance in games like that? Perhaps...but I get some peace of mind

 

[Ansau]

There are other ways to make an overclock stable without using prime95, which pushes the cpu beyond its limits.
There's no real use that reaches the amount of stress put by prime95. So why limiting the overclock (because prime95 requirements are much higher than even most other stressed tests) and stressing the cpu in such a way you're risking with its degradation, for the sake of the glorious stability in prime95?
Do you know that even passing 24h of prime95, the cpu still has chances to fail? It's impossible to have a 100% stable chip, since that would require the validation of running Prime95 during its whole life.

I never used prime95 and my OC of 4.5GHz at 1.36v in a 4690k is solid (by solid I mean it passes 3 hours of rendering or 6 hours of Gw2, which it can put all cores at +80%).

 

[ctk1981]

From another forum, quick dirty haswell OC.

3. Set Fixed CPU Vcore to 1.25 -1.30v
4. Set CPU Cache Voltage also could be called Ring Bus voltage to 1.20v -1.25vv
5. Set Vrin also called "CPU Input Voltage" To 1.9v - 2.1v

See if that helps you get past the wall. I gave up and "settled" on 4ghz with 1.25v vcore.

So far the nice thing about 5775c has been the temps, much better and easier to keep in check than haswell.

1.36V for 4.5ghz on a 4690K? Mine only took 1.225v to run 4.5ghz with 4ghz uncore.

 

[VirtualLarry]

It's impossible to have a 100% stable chip, since that would require the validation of running Prime95 during its whole life.

And some of us distributed computing people, do just that.

For me, I had to be satisfied with 4.0Ghz with my G3258, because anything higher would BSOD eventually, running BOINC projects, even if I boosted the vcore +0.100V. Currently at 1.200V, and it has crashed once so far at those settings.

 

[Galatian]

Again, stress testing implies just that: stressing your CPU as much as possible. Chances are that when your CPU survives Prime it will survive other things you throw at it. On the other hand I thought my CPU is stable after it passed 24h XTU just to figure out it the next game it would crash. That has never happend to be with prime. So what do I gain by not using prime? Perhaps 200 MHz more for benchmarking, but that's about it.

 

[ShintaiDK]

I agree. If it failed at any test, including Prime95 28.7. I would stay its unstable without question. The hard part is to actually get stable, since these tests only use a fraction of the instructions. So plenty of tests is needed just to get somewhere near.

 

[Ansau]

New bios for your MB with the new microcode included.
http://www.asrock.com/mb/Intel/Z97 Extreme9/?cat=Download&os=BIOS

 

[ShintaiDK]

Almost like Asrock was watching this thread

 

[Galatian]

Well it seems there really is a hard wall at 4,2 GHz. I can get it somewhat stable with using LLC 1 (off) and pumping quite a bit more voltage into the processor, but it seems the sweet spot is at 4.1 GHz.

EDIT: The BIOS doesn't solve my problem.

 

[dsplover]

I read the 128MB cache causes troubles when overclocking unless you either disable it or use the UEFI from Asus that allows underclocking it to keep @ 1800MHz-2000MHz...?

Lots of articles and reviews about success from "reviewers" but only 1 or 2 where they explain why they ran into troubles overclocking.

Not sure who to believe as most never mention whether they disabled the Cache eDRAM or undeclocked it, etc.

Love to hear more on this as I want an ASRock not an Asus board.

My current 750 and 951 run like bats out of hell, don't like the idea of losing one of those just to get a less M.2 featured Asus.