My i5 4670k doesn't fit the standard OC profile

[marxses]

Hey guys,

No doubt you get a lot of temp related threads but I'm struggling to find after much searching and scouring through forums any Haswell chips that have similar thermals to my chip.

I am actually quite content with my OC, just curious as to why my chip is behaving so differently.

I built this system (my first) about 3 months ago and have been doing a lot of reading and tweaking and this is what I'm come too.

* CPU: i5 4670k 4.4Ghz @ 1.3v
*Gigabyte Geforce gtx 760 oc edition
* Cooler: Noctua NH-D14
* Motherboard: Asus Gryphon Z87
* RAM: G Skill 8G(2x4G) DDR3 1600Mhz PC12800 RL 9-9-9-24(CL9D-8GBRL)
* Case: Corsair 350D Micro/Mini ATX Case
* Boot Disk: Samsung 250GB SSD 840 EVO
* Seagate SATA3 1TB 7200RPM 64mb Cache
* Multi Monitor: AOC i2757Fh 27" IPS Monitor + Old HP 19" Monitor
* Samsung SATA Black Internal DVD/RW Drive
* Power Supply: Corsair RM 650
* Windows 8.1 64bit



NOTE: I set VCORE in the bios at 1.3v but as far as I can tell using HWmonitor, what it shows as VIN4 is the actual VCORE. The reason i think VIN4 is the VCORE is it shows the voltage steps depending on load you'd expect with C states enabled. When fully stressed though VIN4 maxes out at 1.328v

With that OC and voltage, running all fans on silent Prime95 after an hour can't push the temps past 65°C. With ambient temps around 18°C my idle temps are 22°C.

quick test just to show you the readings i get.



I was playing around with voltages of 1.39v at 4.6Ghz which isn't stable. Prime95 would cause a BSOD after the 10th or 11th test and the temps hadn't passed 72°C.

So unlike most Haswell chips I've got plenty of thermal headroom but to get the chip stable at higher frequencies requires big jumps in voltage. 4.3Ghz is stable at 1.25v and anything below if I remember correctly is stable at stock voltage.

When doing the easy chip test, 4.6Ghz @ 1.2v I get to the boot screen which as I understand should mean a chip in the upper 50th percentile?

Anyways just curious to know if anyone else has got a similar chip or perhaps I'm missing something??

___________________________________________________________________________________________________

Also am i right about VIN4 being the actual VCORE?

CPU-Z reads VCORE as 1.302 V but there is no stepping, just static at 1.302V.

I read somewhere I think in this forum that suggested VIN4 in HWmonitor is in some cases VCORE. Would rather not be running CPU 24/7 @ 1.3 V but the only way to get what CPU-Z reads as VCORE to step is to use adaptive or offset and when video editing or encoding at 1.3 V I get spikes to 1.395 V.

What readings/software should I go by?

______________________________________________________________________________________________________________________

There is a whole lot of conflicting info out there.

Came across an ASUS overclocking guide provided by a PolRoger a few comments down in the comments section after this article

http://www.overclockers.com/3step-guide-to-overclock-intel-haswell

The best guide I've come across for specifically using an ASUS motherboard.

It states that the only way to get voltage stepping is to use offset or adaptive mode and recommends it to avoid cpu degradation..... which goes against the main Haswell OC thread in this forum that states that adaptive mode is useless..... that guide did have to cover more motherboards.... just not sure what best fits my situation.


So I've just switched back to adaptive..... in the bios I've got +0.12 V turbo an additional 0.01V.

CPU-Z shows Core Voltage adjusting according to load. When fully stressed Core Voltage is 1.3, I mentioned video editing previously but I think it's just handbrake that causes the voltage spikes which frequently goes up to 1.387 V

I probably only use handbrake for a about an hour each week. So I'm thinking those spikes are pretty harmless compared to sitting at 1.3 V 24/7 ??

Wondering if I've done some damage to my chip. Can't even boot into OS when taking it up to 4.6Ghz now, even at 1.4 V manual. 4.5Ghz at 1.35 manual is sweet though.

 

[marxses]

No thoughts from anyone at all? Put basically the same post about a month ago in Tom's hardware forum which was met with complete silence.

Hell even "use Google you moron" would be something

If there is too much waffle in my post all I really want to know; is my chip different and has anyone seen the same chip behave in a similar way? or have I got things mixed up and my chip is in fact completely "normal"?

 

[wingman99]

Every chip is different when they make them, some don't even pass any specifications so they have to discard some and some become lesser chips in the product line.

 

[marxses]

Cheers for providing some input mate!

I know that chips made from the same wafers on the same line can be wildly different and as you said the are graded and placed on the product according top there quality.

I'm asking about one specific chip, that's already been deemed to sufficient for the i5 4670k product line.

I also appreciate that there is quite a large variation when it comes to the overclocking performance haswell chips. But one thing that has been very consistent is that haswell chips run very hot.

I can run voltages close to 1.4 V on air and not break the mid seventies when running prime. It just seems to contradict everything I've read on overclocking haswell chips.

I guess because I'm new and learning and don't fully understand what software like HW monitor telling me I also want to make sure, that it doesn't have anything to do my settings, faulty sensors, software. For software though apart from asus suite, 3 different programs give the same temp readings.

 

[Johan45]

Your voltages seem pretty much in line with my 4770K but mine does run hotter here's a link to a shot at 4.6 with 1.38v http://www.overclockers.com/forums/showpost.php?p=7612746&postcount=77.
Now here's some temp pics of Mandrakes at voltages below 1.25 he's showing air VS water. http://www.overclockers.com/forums/showpost.php?p=7612579&postcount=76
Now his CPU is a super hottie but he can also manage to bench higher than I can. These CPUs are very inconsistent. Then you get one like this that will run at 5.0 with less than 1.4v http://www.techpowerup.com/forums/threads/5ghz-club-ii.138608/page-2#post-3090010 that poster goes on to say that was tested with 10 runs of linx, but if you notice in the CPU-z validation it's an Engineering Sample.
You have one cool CPU there lucky for you with aircooling.

 

[marxses]

Thanks Johan!

It's a pity I can't make use of the thermal headroom and get really high clocks but voltages needed are too high.

Still this PC runs very quiet, so that's definitely a plus for having such a cool chip.

 

[Mandrake4565]

Thanks Johan!

It's a pity I can't make use of the thermal headroom and get really high clocks but voltages needed are too high.

Still this PC runs very quiet, so that's definitely a plus for having such a cool chip.

Just the nature of the beast with these Haswells, they are all over the place with temps and OC's. They still pretty darn fast chips but it would be nice if they were a little more consistent. BTW!

 

[marxses]

Just the nature of the beast with these Haswells, they are all over the place with temps and OC's. They still pretty darn fast chips but it would be nice if they were a little more consistent. BTW!

Thanks Mandrake! They are crazy fast chips as far as I'm concerned, massive step up from my old HP Pavillion

 

[Lvcoyote]

Have you plaid around with the LLC and Digi+ BIOS settings yet? Those can be very helpful when trying to get higher clocks. Some of them will cause higher temps though, so tread lightly.

 

[Mandrake4565]

Thanks Mandrake! They are crazy fast chips as far as I'm concerned, massive step up from my old HP Pavillion

Indeed, to be more specific towards my chip. Johan and I did a little testing, comparing ours. His will run 6-7c cooler with more voltage on lesser cooling. That said so far I have been able to overclock mine slightly higher on really good water, 5250 Mhz to be exact, not stable for daily usage though.

 

[marxses]

Have you plaid around with the LLC and Digi+ BIOS settings yet? Those can be very helpful when trying to get higher clocks. Some of them will cause higher temps though, so tread lightly.

Yeh definitely, I've played around with pretty much everything and in terms of stability they've made no difference. I've mostly being changing the settings in the bios but gave XTU a go and get exactly the same results.

The only thing I haven't really looked into is RAM, can "down clocking" if that's the right term help with stability at all?

Indeed, to be more specific towards my chip. Johan and I did a little testing, comparing ours. His will run 6-7c cooler with more voltage on lesser cooling. That said so far I have been able to overclock mine slightly higher on really good water, 5250 Mhz to be exact, not stable for daily usage though.

Dam that's an impressive OC!

So are you using manual voltages? What's your take on CPU degradation?

Not really keen to keep my voltages set above 1.3 V 24/7 so that limits me to 4.4 Ghz when using manual voltage.

But if I use adaptive and take the voltage up to 1.35 V + to get to 4.5 Ghz + I'm worried voltage spikes likely well over 1.4 V could cook the chip.

I don't need the performance increase but it would be nice to see how far I can take it...

 

[marxses]

Indeed, to be more specific towards my chip. Johan and I did a little testing, comparing ours. His will run 6-7c cooler with more voltage on lesser cooling. That said so far I have been able to overclock mine slightly higher on really good water, 5250 Mhz to be exact, not stable for daily usage though.

Also are you using BLCK to get that frequency? My PC doesn't seem to like BLCK at all.

 

[marxses]

Check it out the temps: 4.6 Ghz @ 1.4 V........ if I want my chip to survive I'd be stupid to stay at the those voltages right?

Only 10 P95 tests but gives you the gist.



I updated to the latest Asus Bios (2003) and that has actually improved things a little. Can get 4.4 Ghz stable at 1.29 instead of 1.3 V........ 4.3 Ghz is stable at 1.22 V so it obviously hits some kind of wall after that frequency.

 

[Mandrake4565]

Hey, went to bed after the last post hence the lack of responses.

I'm not one to speak to on Cpu degradation, especially my 4770k. I mainly use it for benching so it has seen voltage most people should not use. I really haven't paid much attention to what are the recommended safe voltages on the haswells. I would think if your temps are good staying 1.4 and below for daily usage would be fine, though don't hold me to that.

I use manual voltage when benching and pushing for max overclocks and offset for my daily Oc of 4.4 Mhz I can do 4.4 at 1.23 v.


I do use BCLK to oc it when really pushing to do 5250 it was 51 X 103 though I didn't mess with any other voltage other then Cpu V and Cache Voltage.

Updating the BIOS does change things, some will OC better some worse. This is where playing around with the chip helps a lot. I know on certain BIOS with my AMD chip I can run the ram higher and another I can OC the Chip further, only found this out by playing though.

Lastly, yes running the ram at high Mhz can hinder the OC but I believe you need to be running it really really high, above 2400 mhz I believe. I was running my sticks at 2200 with 8-10-6 timings and was able to bench at 5200 Mhz on the chip.

 

[turbobooster196]

Hey guys,

No doubt you get a lot of temp related threads but I'm struggling to find after much searching and scouring through forums any Haswell chips that have similar thermals to my chip.

I am actually quite content with my OC, just curious as to why my chip is behaving so differently.

I built this system (my first) about 3 months ago and have been doing a lot of reading and tweaking and this is what I'm come too.

* CPU: i5 4670k 4.4Ghz @ 1.3v
*Gigabyte Geforce gtx 760 oc edition
* Cooler: Noctua NH-D14
* Motherboard: Asus Gryphon Z87
* RAM: G Skill 8G(2x4G) DDR3 1600Mhz PC12800 RL 9-9-9-24(CL9D-8GBRL)
* Case: Corsair 350D Micro/Mini ATX Case
* Boot Disk: Samsung 250GB SSD 840 EVO
* Seagate SATA3 1TB 7200RPM 64mb Cache
* Multi Monitor: AOC i2757Fh 27" IPS Monitor + Old HP 19" Monitor
* Samsung SATA Black Internal DVD/RW Drive
* Power Supply: Corsair RM 650
* Windows 8.1 64bit



NOTE: I set VCORE in the bios at 1.3v but as far as I can tell using HWmonitor, what it shows as VIN4 is the actual VCORE. The reason i think VIN4 is the VCORE is it shows the voltage steps depending on load you'd expect with C states enabled. When fully stressed though VIN4 maxes out at 1.328v

With that OC and voltage, running all fans on silent Prime95 after an hour can't push the temps past 65°C. With ambient temps around 18°C my idle temps are 22°C.

quick test just to show you the readings i get.



I was playing around with voltages of 1.39v at 4.6Ghz which isn't stable. Prime95 would cause a BSOD after the 10th or 11th test and the temps hadn't passed 72°C.

So unlike most Haswell chips I've got plenty of thermal headroom but to get the chip stable at higher frequencies requires big jumps in voltage. 4.3Ghz is stable at 1.25v and anything below if I remember correctly is stable at stock voltage.

When doing the easy chip test, 4.6Ghz @ 1.2v I get to the boot screen which as I understand should mean a chip in the upper 50th percentile?

Anyways just curious to know if anyone else has got a similar chip or perhaps I'm missing something??

___________________________________________________________________________________________________

Also am i right about VIN4 being the actual VCORE?

CPU-Z reads VCORE as 1.302 V but there is no stepping, just static at 1.302V.

I read somewhere I think in this forum that suggested VIN4 in HWmonitor is in some cases VCORE. Would rather not be running CPU 24/7 @ 1.3 V but the only way to get what CPU-Z reads as VCORE to step is to use adaptive or offset and when video editing or encoding at 1.3 V I get spikes to 1.395 V.

What readings/software should I go by?

______________________________________________________________________________________________________________________

There is a whole lot of conflicting info out there.

Came across an ASUS overclocking guide provided by a PolRoger a few comments down in the comments section after this article

http://www.overclockers.com/3step-guide-to-overclock-intel-haswell

The best guide I've come across for specifically using an ASUS motherboard.

It states that the only way to get voltage stepping is to use offset or adaptive mode and recommends it to avoid cpu degradation..... which goes against the main Haswell OC thread in this forum that states that adaptive mode is useless..... that guide did have to cover more motherboards.... just not sure what best fits my situation.


So I've just switched back to adaptive..... in the bios I've got +0.12 V turbo an additional 0.01V.

CPU-Z shows Core Voltage adjusting according to load. When fully stressed Core Voltage is 1.3, I mentioned video editing previously but I think it's just handbrake that causes the voltage spikes which frequently goes up to 1.387 V

I probably only use handbrake for a about an hour each week. So I'm thinking those spikes are pretty harmless compared to sitting at 1.3 V 24/7 ??

Wondering if I've done some damage to my chip. Can't even boot into OS when taking it up to 4.6Ghz now, even at 1.4 V manual. 4.5Ghz at 1.35 manual is sweet though.

Why is your psu that way with the fan up???

 

[marxses]

Hey, went to bed after the last post hence the lack of responses.

I'm not one to speak to on Cpu degradation, especially my 4770k. I mainly use it for benching so it has seen voltage most people should not use. I really haven't paid much attention to what are the recommended safe voltages on the haswells. I would think if your temps are good staying 1.4 and below for daily usage would be fine, though don't hold me to that.

I use manual voltage when benching and pushing for max overclocks and offset for my daily Oc of 4.4 Mhz I can do 4.4 at 1.23 v.


I do use BCLK to oc it when really pushing to do 5250 it was 51 X 103 though I didn't mess with any other voltage other then Cpu V and Cache Voltage.

Updating the BIOS does change things, some will OC better some worse. This is where playing around with the chip helps a lot. I know on certain BIOS with my AMD chip I can run the ram higher and another I can OC the Chip further, only found this out by playing though.

Lastly, yes running the ram at high Mhz can hinder the OC but I believe you need to be running it really really high, above 2400 mhz I believe. I was running my sticks at 2200 with 8-10-6 timings and was able to bench at 5200 Mhz on the chip.

Yeh think I'll just keep it at 4.4 Ghz. I use the PC for work, hoping it will give me 4-5 years. No point in burning it just to satisfy my curiosity.

Cheers for the info mate, really appreciate it.

 

[marxses]

Why is your psu that way with the fan up???

If that's the wrong way then purely because I don't what I'm doing

Is it supposed to be the other way?

Think I just figured the brand name shouldn't be upside down. With that particular PSU the fan rarely turns on anyway, Corsair states that because of it gold rating there is less heat being generated due to higher efficiency, therefore the fan is only needed under high loads.

Should i flip it? Thinking about sleeving my cables anyway, so can do it then.

 

[Mandrake4565]

If that's the wrong way then purely because I don't what I'm doing

Is it supposed to be the other way?

Think I just figured the brand name shouldn't be upside down. With that particular PSU the fan rarely turns on anyway, Corsair states that because of it gold rating there is less heat being generated due to higher efficiency, therefore the fan is only needed under high loads.

Should i flip it? Thinking about sleeving my cables anyway, so can do it then.

You can have it either way. The way it is, it is drawing air form the case if it's flipped it draws it from outside. The only issue with it the way it is, is if the case is really how hot or you drop something into it.

 

[marxses]

You can have it either way. The way it is, it is drawing air form the case if it's flipped it draws it from outside. The only issue with it the way it is, is if the case is really how or you drop something into it.

On impulse I've actually just started shortening and sleeving my cables now. As much as I admire some of the pc's out there this will be very basic.

The cables I've got are too long, they are stiff and their flat form makes them hard to manage, putting quite a lot of pressure some components. The back cover is actually bulging from cables.

I'll flip the psu.

 

[marxses]

Yeh think I'll just keep it at 4.4 Ghz. I use the PC for work, hoping it will give me 4-5 years. No point in burning it just to satisfy my curiosity.

Cheers for the info mate, really appreciate it.

Ha just flipped it and the writing is still the right way up

Thanks for pointing that out!