Overclock.net › Forums › Intel › Intel CPUs › Haswell Overclocking Guide [With Statistics]
New Posts  All Forums:Forum Nav:

Haswell Overclocking Guide [With Statistics] - Page 975

post #9741 of 19540
Quote:
Originally Posted by Wakizashis View Post

Damn, which ones? All in the test folder?

I don't remember, but check on the website of avisynth64, it says you have to put some DLLs in the Windows system 32 folder.
PC
(14 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus Impact Asus GTX770 DCII Gskill TridentX 2400 C10 
Hard DriveHard DriveCoolingOS
Samsung 840 Pro Synology DS413 Noctua NH D14 Windows 7 Ultimate 
MonitorKeyboardPowerCase
IIyama 82776HDS Corsair K70 Seasonic X-750 Bitfenix Prodigy Black 
MouseAudio
Logitech Performance MX Logitech Z5500 
  hide details  
Reply
PC
(14 items)
 
  
CPUMotherboardGraphicsRAM
4770k Asus Maximus Impact Asus GTX770 DCII Gskill TridentX 2400 C10 
Hard DriveHard DriveCoolingOS
Samsung 840 Pro Synology DS413 Noctua NH D14 Windows 7 Ultimate 
MonitorKeyboardPowerCase
IIyama 82776HDS Corsair K70 Seasonic X-750 Bitfenix Prodigy Black 
MouseAudio
Logitech Performance MX Logitech Z5500 
  hide details  
Reply
post #9742 of 19540
Quote:
Originally Posted by Gugus03 View Post

I don't remember, but check on the website of avisynth64, it says you have to put some DLLs in the Windows system 32 folder.

Well tried to copy Devil and avisynth.dlls to system32 also for system folder. Tired even dlls with a x264 and nothing worked. Same thing about redirecting to get Avisynth, which is installed....
post #9743 of 19540
Anyone know why my cpu isn't downvolting?



It downclocks just fine, but the voltage is stuck at my manual value.
I thought it downvolts with c states enabled?
post #9744 of 19540
Quote:
Originally Posted by 7ranslucen7 View Post

Anyone know why my cpu isn't downvolting?



It downclocks just fine, but the voltage is stuck at my manual value.
I thought it downvolts with c states enabled?
what kind of mode are you in? adaptive or manual?
post #9745 of 19540
Quote:
Originally Posted by paramazon View Post

what kind of mode are you in? adaptive or manual?

Manual
post #9746 of 19540
Put c-stats to Auto
Primo
(15 items)
 
  
CPUMotherboardGraphicsRAM
I7-7700k Asus Maximus Hero ix Evga GTX 1080 ACX 3.0 G-Skill F-43200C15D-16GTZKW 
Hard DriveHard DriveCoolingCooling
Samsung 850 Evo Corsair Neutron  Ek Supremacy CSQ clean OCOOL 360mm radiator 
CoolingCoolingOSKeyboard
OCOOL 360mm radiator Ek -FC1080 GTX  Windows 10 CM Storm 
PowerCaseMouse
Evga 750 watt Phanteks Enthoo Primo Roccat Kone Optical 
  hide details  
Reply
Primo
(15 items)
 
  
CPUMotherboardGraphicsRAM
I7-7700k Asus Maximus Hero ix Evga GTX 1080 ACX 3.0 G-Skill F-43200C15D-16GTZKW 
Hard DriveHard DriveCoolingCooling
Samsung 850 Evo Corsair Neutron  Ek Supremacy CSQ clean OCOOL 360mm radiator 
CoolingCoolingOSKeyboard
OCOOL 360mm radiator Ek -FC1080 GTX  Windows 10 CM Storm 
PowerCaseMouse
Evga 750 watt Phanteks Enthoo Primo Roccat Kone Optical 
  hide details  
Reply
post #9747 of 19540
Thread Starter 
This thread has gotten slow. By this thread's standards anyways. That's what happens when I leave. Hey look, the troll is gone!
Quote:
Originally Posted by fritzdis View Post


VID being lower than vCore is one reason why I noted how many entries were at or below 1.25v.

Certainly, you can question the stability of some of the entries. Let's throw out all of the 4.5 / <1.3v entries that depend on XTU or an unspecified version of Prime95, as well as the "some BF3" entry. We'll also ignore Hunter and Tobi's entries (I'm guessing Hunter would still be very stable under 1.25v, but Tobi might not be). We'll even throw out the reported batch 310 or 312 chips. After all that, we're still left with 12 of 25 entries having a VID of 1.275 or below.

The point is not that the "average" chip should be able to do 4.5 at less than 1.3v. As I said in the edit of my previous post (which got bottom-paged), we're not going to be able to draw that kind of conclusion from our limited data. I just wanted to point out that "almost no one using less than 1.30v for 4.5 and above" is probably not an accurate description of the data provided by the table on page 1.

There's only so much I can do to gather data. I've already been bugging people over and over in the past to try to get data. Some people are too lazy to tell me what they're running and I've even had to go out and PM people after they've submitted OC entries months later to ensure they are still stable.

I'll have a lot of fun trying to get random people who don't even post in this thread to cough up settings but telling them to run a specific number of prime hours and have the picture when they don't even know who I am, lol.

Quote:
Originally Posted by mav451 View Post


I'm actually a bit curious on how Wizzie put together the table. Mine is a Costa Rica, so if the very first number/letter indicates country of origin, my 3313A646 chip should really be a "313" and not a "331," correct?

So when you say 310 or 312, the *very first* number/letter was a L - indicating Malaysia.
Anyway, just realized that, but maybe there is something to be said of the earlier batches (Week 10, 12, or 13).

Hi, my name is Wizzie.
The batch number is taken as the first three digits of the number. So for you I would put 331. I believe the Malay people have batch numbers that read 312xxxxx (for example), so their batch would be listed as 312. So I don't think first number indicates location. But the new policy now is to display the entire batch number for less confusion and more complete data.

Quote:
Originally Posted by 7ranslucen7 View Post

Just got my 4770k+gene setup, I'm amazed how large the margin of cpu:voltage requirement is.

Currently using 1.22 for 4.6, testing for stability!

Good luck, please come back and report your results.

 

Quote:
Originally Posted by blaze2210 View Post


Apparently, I overshot the VID a bit the last time....So I increased the VRIN a bit, went from 1.945 to 1.956 and started gradually decreasing the VID - went from 1.430v to 1.38....Now, I'm running 4.7ghz with 1.41v....thumbsupsmiley.png

I tried messing with the BLCK before and just got a bunch of instability....Now that I know a bunch more about Haswell and my board, it might be worth giving it another shot....biggrin.gif

We've got another incident of Vrin affecting stability again. :buttkick:

 

Quote:
Originally Posted by fleetfeather View Post


It does. JJ is off his rocker. Half the crap he said was laughable.

"We've tested over 1000 chips now"... Yeah, maybe 310's and 312's. I mean hell, they probs weren't even retail chips if Asus had over 1000 of them...

 

That very line was used against my guide. I still recall that. JJ has tests thousands of CPUs, what does poor Darkwizzie have?

Quote:
Originally Posted by fleetfeather View Post


VID is 0.02 or 0.03 lower than Vcore.

Most of those people haven't been seen in months, with the exception of hunter and tobi who both recently reported they needed higher Vcore than originally reported (coincidently two people with two of the lowest vid entries for their respective multipliers).

Have a look at the stress tests some have used: XTU is laughable, Prime would've been okay if we knew what version it was (no doubt some ancient version with stupid easy parameters)

Also look at the time spent testing. 15 minutes? Yeah, that's rock solid.

Then take a look at batch numbers. Everyone who's been doing this for a while knows there were some superb 310 and 312 batches. Furthermore this is OCN, so binning will occur.


At the end of the day, I guess it was pointless to bring that table up, since I can easily read between the lines, whereas everyone else will take it at face value. Oh well.

I can spend time to maintain a chart but the effort is for naught if I cannot prove the validity of my data. So I guess now I have to back and PM many, many people. Fun!
I don't think people go back to like Prime v20 or something on purpose, I think almost all Prime users run 27.9, it's the standard. If people are willing to cheat at Prime to validate or claim validation, then they can just flat out make stuff up instead of wasting time with an ancient Prime version.

While there are factors that bring up the average overclock up, there are also people who refuse to go past a set amount of voltage. And that voltage may be low, and some people caught on Linpack or die or even Prime or Die mentalities. Which is their choice, but that sort of testing lowers their maximum OC. So there are forces that push the max OC up and things that bring it down. Binning isn't really that common and when I see it I can put up a double entry for the user for both CPUs.

You could even read my entry and see, look, crashes at x20 pass x264, the very stress test and length which I recommended and I started this thread! Yet, it's been months with tons of CPU usage from chess and I'm still fine with my stability. In fact, I actually upped my uncore by 100mhz with no side effects. Although, the people with less amounts of stress testing charted don't have as extensive notes on stability as my entry. A disproportionately high amount of users at the start reported good OCs, we're talking 4.7, 4.8 basically. That was the start of the chart. As time goes by I try to add more data and more followup calls and whatnot, picture verification, to try to get the validity of the data to go higher without forcing every charted person to run my stress tests - as nobody would bother doing that.

 

Quote:
Originally Posted by fritzdis View Post


For 4.5, 23 of 36 entries (64%) have a VID of less than 1.3v, and none of those even reported a load vCore of 1.3v or more. 15 of the entries (42%) actually have a VID of 1.25v or lower.

Even at 4.6, 16 of 25 entries (64%) have a VID of less than 1.3V, with only 1 reporting a load vCore of 1.3v or more. 11 of the entries (44%) have a VID of 1.25v or lower.

Edit: To be clear, this doesn't necessarily mean that the majority of chips can hit 4.5 at under 1.3v. The VID numbers at 4.4 are actually worse than at 4.5. If the people at 4.4 and below were to push their chips to 4.5, it would almost surely increase the average VID needed for 4.5. Likewise, if the people above 4.5 were to roll back their OCs to 4.5, it would likely lower the average VID at 4.5.

Figuring out what it would take for an average chip to hit 4.5 is probably not possible, given our limited data. Nonetheless, I don't think it's remotely correct to conclude that almost no one is using less than 1.3v to hit 4.5.

Yeah, I agree with the final conclusion.

 

Quote:
Originally Posted by Clockwerk View Post


Most of that probably stems from the three step guide and when haswell first came out an Asus rep said to start at 4.5 w/ 1.2 and see if it boots. If it did you supposedly had an average chip, stable was golden. I think since then we have figured out that 4.5 stable at 1.2 is really rare. Agree with you on coolers though, evo will probably not cut it for vast majority of people trying to hit 4.5 on haswell. Evo was a godsend on sandy bridge but not so much with haswell.

Again, it really depends on the stress test in question and the voltage required. And since voltage varies so much, and stress tests vary so much, the guidelines for what cooling solution is 'required' are very broad. Some are on HT, some are not. But I think starting with D14 or Silver Arrow is a good choice. If you're on 212 and you want to upgrade, now you've got a 212 you can't use and isn't worth the hassle of selling. 

Quote:
Originally Posted by JJFIVEOH View Post

I've got what may be a stupid question. I'm setting this computer up for somebody else so I won't be spending significant time on it. Currently I am at 4.3 @ 1.335 and it's stable. This is a 4670K with an MSI Z87-GD65. On my 2600K/Asrock OC Formula I have it set up so the voltage drops to around .96-1.00 when not under load, and 1.40 volts under load. This is a 4.6 setup. I don't see any settings in BIOS for this board/CPU combo that might allow for me to do with same thing, having the VCore drop when not under a load. Am I missing something?

On a side note, I am quite surprised at the PCMark7 scores I am getting from this computer at 4.3 with an SSD over the PC I have in my sig. Almost 20% higher.

C states enabled to C7. 

 

Quote:
Originally Posted by lilchronic View Post

4.8Ghz @ 1.48v biggrin.gif


 

Should I chart that or...?

Quote:
Originally Posted by Clockwerk View Post

OC Update followed by a question that some nice person will hopefully answer biggrin.gif

CPU: I7-4770k
Core Multiplier: 47
CPU VID: 1.32
Vcore: 1.324ish. Depends on stress test
Uncore multiplier: 44
Uncore Voltage: 1.2v
Cooling: Custom loop
Stability test: XTU/Cinebench
Ram Speed: 2133 9/11/10/27 1T
Input Voltage: 1.95
Mobo: MSI Mpower Max

Fairly happy with this OC and have been really struggling to stabilize 4.8. I found out yesterday while messing around that I am able to change the Bclk from 100 to 103 and it was still stable (so far). This increases my CPU to 4.842 and Ram to 2200. Vcore rises slightly when stressing but I have not had to change my CPU VID yet. Here's my question/issue. I read a lot of bad stuff about changing the base clock being able to torch other components (especially hard drives) since it changes the speed of just about everything including sata ports, how much of a risk am I taking by changing the bclk from 100 to 103? It was kind of a struggle to find anything related to haswell on the topic, most was SB or IB related. Thoughts/input? Thanks

Got it.

 

Quote:
Originally Posted by blaze2210 View Post


So you picked some random settings, got a BSOD, ignored the error code (which hints at what needs to be adjusted), and somehow it's the CPU's fault? Educate yourself on the workings of Haswell before you blame the instability on the CPU....

That's like not knowing how to work on cars, and blaming the car because you're not a mechanic....wink.gif

NOTE: I'm sorry if this comes across as rude, but it gets a bit irritating when people are constantly coming in this thread saying that their CPU sucks....If you read the 1st page of this thread, you will gain some knowledge....If you read the rest of the pages, you will gain even more....thumb.gif There were about 600 pages when I first found this thread, and I read every post in each page long before I even posted the 1st time in here. By the time I got to the last page, virtually EVERY question I had was already asked and answered....2cents.gif

How long did it take you? I changed the settings to 100 posts per page and now this thread only has 98 pages for me.

Quote:
Originally Posted by 7ranslucen7 View Post

Just got my 4770k+gene setup, I'm amazed how large the margin of cpu:voltage requirement is.

Currently using 1.22 for 4.6, testing for stability!

Do come back when you're done testing.

 

Quote:
Originally Posted by blaze2210 View Post


Apparently, I overshot the VID a bit the last time....So I increased the VRIN a bit, went from 1.945 to 1.956 and started gradually decreasing the VID - went from 1.430v to 1.38....Now, I'm running 4.7ghz with 1.41v....thumbsupsmiley.png

I tried messing with the BLCK before and just got a bunch of instability....Now that I know a bunch more about Haswell and my board, it might be worth giving it another shot....biggrin.gif

Maybe the best option from start to finish is, as you get towards the end is to not up the multiplier for a few days. Like if you're on x45, cool, you passed whatever stress test you feel is good. Now go game for a few days and don't come back until you've nolifed other games.

 

 

 

Quote:
Originally Posted by goodtobeking View Post

Dont you have to enable speed step for it to throttle?? I might be wrong on that.

 

You mean throttle the speed of the processor? Most likely yes. But with power settings, they vary a whole lot from mobo to mobo. Not kidding, testing it across mobos was hell because one mobo would act different with the same setting on another brand.

Quote:
Originally Posted by Leopard2lx View Post

Would a core voltage of 1.44 be too much for a 4670k @ 4.6 Ghz? It seems stable in Prime 95 so far although temps are about 90. LOL. However I know in real life gaming and apps I wont ever see that kind of temps.
I might be able to lower voltage a bit and see if still stable.

Define 'too much'.

 

Quote:
Originally Posted by Hanshin View Post

So, it's time for my entry!

Username: Hanshin
CPU Model: 4770k (Delidded)
Core Multiplier: 45
CPU VID: 1.23
Vcore: 1.248
Uncore Multiplier: 44
Uncore Voltage:1.22
Cooling Solution: H100i w/ Gelid Silent 120 PWM
Stability Test: x264 *20, OCCT Linpack 8 hours, OCCT CPU 7 hours, LinX 7 hours, Asus RealBench (benchmark) *10, gaming (Crysis 3, Planetside 2, FF XIV)
Batch Number: Malay L307B239
Ram Speed: Avexir Core Series 2000Mhz 9-11-9-27-36
Ram Voltage: 1.65
Input Voltage: 1.82
LLC Setting: LLC level 8
Motherboard: Maximus VI Formula 0804 Bios

And the picture!



It seems the sweet spot for my CPU is 4.4, as I can maintain it stable at this speed with 1.184v on all the benches I mentionned for my actual OC. I think I will revert to 4.4 as I can't really justify the voltage bump for only such a small gain.
The max stable tested is 4.8@1.295v but with so many horror stories about the chip degradation, I think I'll keep it "safe" and no push it this far.

I need to tweak the uncore and RAM but I'm satisfied with my overclock right now! Thanks to all for the numerous tips I read on this topic! thumb.gif

I'll chart you soon, I promise. Did you update your settings? Screw it, I'll pm you.

 

Quote:
Originally Posted by Wakizashis View Post

Hi again guys,

how should I get x264 bench to work? Downloaded it, installed AviSynth 2.5.8 and it still babbles about installing AviSynth first and redirecting me to the webpage when trying to run the test, pfffft. Running app as a admin + tried to DL version which should be working even without this step. Thanks.

 

 

Quote:
Originally Posted by mav451 View Post

I thought it was repackaged to be easier?
Is there still a 'test' folder created in the x264 bench folder? You need to follow the instructions listed in the text files - it's about three things you need to do manually (copying files, regedit, and something else I don't remember right now haha).

After that, you need to put this folder where it's easy for you to remember. This is why I recommended renaming it to just "HD Bench" and relocating the folder right off your boot drive, e.g. C:\HD bench. Then, running an elevated Command prompt, you change directory to the HD Bench folder and then run the looped bat file via the Command Prompt.

You cannot simply click the bat file, as that will not work.

 

 

Quote:
Originally Posted by Wakizashis View Post

Hi,
yeah, test folder is there. reg file of avi synth as well. Tried the bat file but also the 64bit exe, so maybe I miss the step with command line?

Please ask Forceman, he's the man to ask.

 

Quote:
Originally Posted by Nieeru View Post

Okay so, I thought I should go about getting the most out of my OC after fiddling around with my settings over the course of Christmas. (Been away for Work, yay..)

I set multiplier to x45 and ring bus to auto with XMP enabled (1333Mhz -> 1600Mhz) but the core voltage needed to be set to 1.3. I could probably have set it lower than that, mainly because I was running Adaptive mode while "testing" with Prime95... And an old model at that. (Don't hate! This is new to me.)

Anyway, I haven't been experiencing any BSODs or issues at all with 45:45 (1:1) @ 1.3v with XMP enabled, I'm not sure how good or effective that is. Take into consideration that my build is being used for gaming (BF4, WoW, Skyrim, FFXIV:ARR, ..)

What I'm thinking about doing is resetting my settings back to stock and work my way up from there as this is all still pretty confusing to me, overclocking that is. (The terminology and what some of the settings do mostly.)

Temps have never gone above 55c while running most games on that setup. What do you guys think?

 

You could try to just work with what you know to be stable. Go to 46:34. Up that Vcore and up that Vrin.

Quote:
Originally Posted by paramazon View Post

just a side question if i want to get my uncore to 38x multiplier would i need to up the uncore voltage or just leave it at default?

Probably won't have. If you do, probably won't have to up it by any sort of dangerous amount.

 

Quote:
Originally Posted by zeroofmhx View Post

core 42x 1.240v
uncore 34x 1.165v
LLC lvl 7
Initial voltage 1.170
Eventual Voltage 1.850
at 1333MHZ
using x264 HD 5.0 as stress test

should i set my ram to xmp now?



edit: forgot to post cpu-z

Still working on the OC? Don't leave me hanging.

 

 
Quote:
Originally Posted by fritzdis View Post


VID being lower than vCore is one reason why I noted how many entries were at or below 1.25v.

Certainly, you can question the stability of some of the entries. Let's throw out all of the 4.5 / <1.3v entries that depend on XTU or an unspecified version of Prime95, as well as the "some BF3" entry. We'll also ignore Hunter and Tobi's entries (I'm guessing Hunter would still be very stable under 1.25v, but Tobi might not be). We'll even throw out the reported batch 310 or 312 chips. After all that, we're still left with 12 of 25 entries having a VID of 1.275 or below.

The point is not that the "average" chip should be able to do 4.5 at less than 1.3v. As I said in the edit of my previous post (which got bottom-paged), we're not going to be able to draw that kind of conclusion from our limited data. I just wanted to point out that "almost no one using less than 1.30v for 4.5 and above" is probably not an accurate description of the data provided by the table on page 1.

There's only so much I can do to gather data. I've already been bugging people over and over in the past to try to get data. Some people are too lazy to tell me what they're running and I've even had to go out and PM people after they've submitted OC entries months later to ensure they are still stable.

I'll have a lot of fun trying to get random people who don't even post in this thread to cough up settings but telling them to run a specific number of prime hours and have the picture when they don't even know who I am, lol.

Quote:
Originally Posted by mav451 View Post


I'm actually a bit curious on how Wizzie put together the table. Mine is a Costa Rica, so if the very first number/letter indicates country of origin, my 3313A646 chip should really be a "313" and not a "331," correct?

So when you say 310 or 312, the *very first* number/letter was a L - indicating Malaysia.
Anyway, just realized that, but maybe there is something to be said of the earlier batches (Week 10, 12, or 13).

Hi, my name is Wizzie.
The batch number is taken as the first three digits of the number. So for you I would put 331. I believe the Malay people have batch numbers that read 312xxxxx (for example), so their batch would be listed as 312. So I don't think first number indicates location. But the new policy now is to display the entire batch number for less confusion and more complete data.

Quote:
Originally Posted by 7ranslucen7 View Post

Just got my 4770k+gene setup, I'm amazed how large the margin of cpu:voltage requirement is.

Currently using 1.22 for 4.6, testing for stability!

Good luck, please come back and report your results.

 

Quote:
Originally Posted by blaze2210 View Post


Apparently, I overshot the VID a bit the last time....So I increased the VRIN a bit, went from 1.945 to 1.956 and started gradually decreasing the VID - went from 1.430v to 1.38....Now, I'm running 4.7ghz with 1.41v....thumbsupsmiley.png

I tried messing with the BLCK before and just got a bunch of instability....Now that I know a bunch more about Haswell and my board, it might be worth giving it another shot....biggrin.gif

We've got another incident of Vrin affecting stability again. :buttkick:

 

Quote:
Originally Posted by fleetfeather View Post


It does. JJ is off his rocker. Half the crap he said was laughable.

"We've tested over 1000 chips now"... Yeah, maybe 310's and 312's. I mean hell, they probs weren't even retail chips if Asus had over 1000 of them...

 

That very line was used against my guide. I still recall that. JJ has tests thousands of CPUs, what does poor Darkwizzie have?

Quote:
Originally Posted by fleetfeather View Post


VID is 0.02 or 0.03 lower than Vcore.

Most of those people haven't been seen in months, with the exception of hunter and tobi who both recently reported they needed higher Vcore than originally reported (coincidently two people with two of the lowest vid entries for their respective multipliers).

Have a look at the stress tests some have used: XTU is laughable, Prime would've been okay if we knew what version it was (no doubt some ancient version with stupid easy parameters)

Also look at the time spent testing. 15 minutes? Yeah, that's rock solid.

Then take a look at batch numbers. Everyone who's been doing this for a while knows there were some superb 310 and 312 batches. Furthermore this is OCN, so binning will occur.


At the end of the day, I guess it was pointless to bring that table up, since I can easily read between the lines, whereas everyone else will take it at face value. Oh well.

I can spend time to maintain a chart but the effort is for naught if I cannot prove the validity of my data. So I guess now I have to back and PM many, many people. Fun!
I don't think people go back to like Prime v20 or something on purpose, I think almost all Prime users run 27.9, it's the standard. If people are willing to cheat at Prime to validate or claim validation, then they can just flat out make stuff up instead of wasting time with an ancient Prime version.

While there are factors that bring up the average overclock up, there are also people who refuse to go past a set amount of voltage. And that voltage may be low, and some people caught on Linpack or die or even Prime or Die mentalities. Which is their choice, but that sort of testing lowers their maximum OC. So there are forces that push the max OC up and things that bring it down. Binning isn't really that common and when I see it I can put up a double entry for the user for both CPUs.

You could even read my entry and see, look, crashes at x20 pass x264, the very stress test and length which I recommended and I started this thread! Yet, it's been months with tons of CPU usage from chess and I'm still fine with my stability. In fact, I actually upped my uncore by 100mhz with no side effects. Although, the people with less amounts of stress testing charted don't have as extensive notes on stability as my entry. A disproportionately high amount of users at the start reported good OCs, we're talking 4.7, 4.8 basically. That was the start of the chart. As time goes by I try to add more data and more followup calls and whatnot, picture verification, to try to get the validity of the data to go higher without forcing every charted person to run my stress tests - as nobody would bother doing that.

 

Quote:
Originally Posted by fritzdis View Post


For 4.5, 23 of 36 entries (64%) have a VID of less than 1.3v, and none of those even reported a load vCore of 1.3v or more. 15 of the entries (42%) actually have a VID of 1.25v or lower.

Even at 4.6, 16 of 25 entries (64%) have a VID of less than 1.3V, with only 1 reporting a load vCore of 1.3v or more. 11 of the entries (44%) have a VID of 1.25v or lower.

Edit: To be clear, this doesn't necessarily mean that the majority of chips can hit 4.5 at under 1.3v. The VID numbers at 4.4 are actually worse than at 4.5. If the people at 4.4 and below were to push their chips to 4.5, it would almost surely increase the average VID needed for 4.5. Likewise, if the people above 4.5 were to roll back their OCs to 4.5, it would likely lower the average VID at 4.5.

Figuring out what it would take for an average chip to hit 4.5 is probably not possible, given our limited data. Nonetheless, I don't think it's remotely correct to conclude that almost no one is using less than 1.3v to hit 4.5.

Yeah, I agree with the final conclusion.

 

Quote:
Originally Posted by Clockwerk View Post


Most of that probably stems from the three step guide and when haswell first came out an Asus rep said to start at 4.5 w/ 1.2 and see if it boots. If it did you supposedly had an average chip, stable was golden. I think since then we have figured out that 4.5 stable at 1.2 is really rare. Agree with you on coolers though, evo will probably not cut it for vast majority of people trying to hit 4.5 on haswell. Evo was a godsend on sandy bridge but not so much with haswell.

Again, it really depends on the stress test in question and the voltage required. And since voltage varies so much, and stress tests vary so much, the guidelines for what cooling solution is 'required' are very broad. Some are on HT, some are not. But I think starting with D14 or Silver Arrow is a good choice. If you're on 212 and you want to upgrade, now you've got a 212 you can't use and isn't worth the hassle of selling. 

Quote:
Originally Posted by JJFIVEOH View Post

I've got what may be a stupid question. I'm setting this computer up for somebody else so I won't be spending significant time on it. Currently I am at 4.3 @ 1.335 and it's stable. This is a 4670K with an MSI Z87-GD65. On my 2600K/Asrock OC Formula I have it set up so the voltage drops to around .96-1.00 when not under load, and 1.40 volts under load. This is a 4.6 setup. I don't see any settings in BIOS for this board/CPU combo that might allow for me to do with same thing, having the VCore drop when not under a load. Am I missing something?

On a side note, I am quite surprised at the PCMark7 scores I am getting from this computer at 4.3 with an SSD over the PC I have in my sig. Almost 20% higher.

C states enabled to C7. 

 

Quote:
Originally Posted by lilchronic View Post

4.8Ghz @ 1.48v biggrin.gif


 

Should I chart that or...?

Quote:
Originally Posted by Clockwerk View Post

OC Update followed by a question that some nice person will hopefully answer biggrin.gif

CPU: I7-4770k
Core Multiplier: 47
CPU VID: 1.32
Vcore: 1.324ish. Depends on stress test
Uncore multiplier: 44
Uncore Voltage: 1.2v
Cooling: Custom loop
Stability test: XTU/Cinebench
Ram Speed: 2133 9/11/10/27 1T
Input Voltage: 1.95
Mobo: MSI Mpower Max

Fairly happy with this OC and have been really struggling to stabilize 4.8. I found out yesterday while messing around that I am able to change the Bclk from 100 to 103 and it was still stable (so far). This increases my CPU to 4.842 and Ram to 2200. Vcore rises slightly when stressing but I have not had to change my CPU VID yet. Here's my question/issue. I read a lot of bad stuff about changing the base clock being able to torch other components (especially hard drives) since it changes the speed of just about everything including sata ports, how much of a risk am I taking by changing the bclk from 100 to 103? It was kind of a struggle to find anything related to haswell on the topic, most was SB or IB related. Thoughts/input? Thanks

Got it.

 

Quote:
Originally Posted by blaze2210 View Post


So you picked some random settings, got a BSOD, ignored the error code (which hints at what needs to be adjusted), and somehow it's the CPU's fault? Educate yourself on the workings of Haswell before you blame the instability on the CPU....

That's like not knowing how to work on cars, and blaming the car because you're not a mechanic....wink.gif

NOTE: I'm sorry if this comes across as rude, but it gets a bit irritating when people are constantly coming in this thread saying that their CPU sucks....If you read the 1st page of this thread, you will gain some knowledge....If you read the rest of the pages, you will gain even more....thumb.gif There were about 600 pages when I first found this thread, and I read every post in each page long before I even posted the 1st time in here. By the time I got to the last page, virtually EVERY question I had was already asked and answered....2cents.gif

How long did it take you? I changed the settings to 100 posts per page and now this thread only has 98 pages for me.


Edited by Darkwizzie - 2/8/14 at 11:05am
Undelwalt (2017)
(29 items)
 
 
Celapaleis (2013)
(16 items)
 
CPUMotherboardGraphicsRAM
7600k @ 5.197/1.392 1.392v z170 Asus Hero EVGA 1080ti SC Black (2100/6250) x2 8gb GSkill Trident Z 3600 @ 3804 15-15-15-32-2T 
Hard DriveHard DriveHard DriveCooling
Samsung 950 Pro 512gb (Undelwalt) WD Red 2tb (Pack Yak II) Seagate Expansion Drive 5tb (Phoenix Down II) 3x560mm HardwareLabs Nemesis GTX 
CoolingCoolingCoolingCooling
XPSC Raystorm Pro Watercool Heatkiller 1080ti Full Cover Waterblock EK XTOP Revo Dual D5 (Serial, PWM, v4) EK ZMT (1/2 - 3/4 ID OD) 
CoolingCoolingCoolingCooling
14xEK ACF (Compression Fittings) EK x4 250 (v2) Distilled Water + PT Nuke (Copper Sulphate) 24x140mm Silent Wings 3 
OSMonitorKeyboardPower
Windows 10 Pro 64bit Catleap 27 Inch 2560x1440 IPS 60hz Display Coolermaster Storm Trigger (Brown Switches) EVGA P2 1000w 
CaseMouseMouse PadAudio
Corsair Air 540 Logitech G Pro Tek Syndicate "Raze the World" Desk Mat O2 + Odac by Mayflower Electronics (Rev A) 
AudioAudioAudioAudio
Sennheiser HD 800 2xKrk Rokit 6 G2 Blue Yeti (Grey) Rode PSA-1 Arm 
Other
Aquaero 6 LT + Aluminum Heatsink 
CPUMotherboardGraphicsRAM
i5 6600k Delid @ 4.848/4.848 Asus z170 Hero MSI GTX 980ti @ 1499/4082 1.25v/134% Power 2 x 8gb Gskills Ripjaws V 3131 16-16-16-32 
Hard DriveHard DriveHard DriveHard Drive
Samsung Pro Series 840 256gb (Celapaleis) 1 TB Samsung Drive (Pack Yak) Samsung Pro Series 850 256gb (Celapaleis Reprise) Phoenix Down (External Backup) 
CoolingCoolingOSMonitor
Noctua D14 Kraken x61 Push-Pull for GPU Windows 10 Pro Catleap 2560x1440 @ 60hz 
KeyboardPowerCaseMouse
CM Storm Trigger Brown Switches EVGA Supernova 1000w P2 Corsair 540 Air Razer Abyssus 
Mouse PadAudioAudioAudio
Steelseries QcK Mini Blue Yeti Microphone Objective Dac + Amp Sescom AB Switch 
AudioAudioAudioAudio
Krk Rokit 6 G2 x2 Radioshack Ground Loop Isolator (For Rokits) Wooaudio Aluminum Headphone Stand Sennheiser HD 800 
AudioOther
Logitech c920 Webcam NZXT Sentry Mix 2 Fan Controller 
  hide details  
Reply
Undelwalt (2017)
(29 items)
 
 
Celapaleis (2013)
(16 items)
 
CPUMotherboardGraphicsRAM
7600k @ 5.197/1.392 1.392v z170 Asus Hero EVGA 1080ti SC Black (2100/6250) x2 8gb GSkill Trident Z 3600 @ 3804 15-15-15-32-2T 
Hard DriveHard DriveHard DriveCooling
Samsung 950 Pro 512gb (Undelwalt) WD Red 2tb (Pack Yak II) Seagate Expansion Drive 5tb (Phoenix Down II) 3x560mm HardwareLabs Nemesis GTX 
CoolingCoolingCoolingCooling
XPSC Raystorm Pro Watercool Heatkiller 1080ti Full Cover Waterblock EK XTOP Revo Dual D5 (Serial, PWM, v4) EK ZMT (1/2 - 3/4 ID OD) 
CoolingCoolingCoolingCooling
14xEK ACF (Compression Fittings) EK x4 250 (v2) Distilled Water + PT Nuke (Copper Sulphate) 24x140mm Silent Wings 3 
OSMonitorKeyboardPower
Windows 10 Pro 64bit Catleap 27 Inch 2560x1440 IPS 60hz Display Coolermaster Storm Trigger (Brown Switches) EVGA P2 1000w 
CaseMouseMouse PadAudio
Corsair Air 540 Logitech G Pro Tek Syndicate "Raze the World" Desk Mat O2 + Odac by Mayflower Electronics (Rev A) 
AudioAudioAudioAudio
Sennheiser HD 800 2xKrk Rokit 6 G2 Blue Yeti (Grey) Rode PSA-1 Arm 
Other
Aquaero 6 LT + Aluminum Heatsink 
CPUMotherboardGraphicsRAM
i5 6600k Delid @ 4.848/4.848 Asus z170 Hero MSI GTX 980ti @ 1499/4082 1.25v/134% Power 2 x 8gb Gskills Ripjaws V 3131 16-16-16-32 
Hard DriveHard DriveHard DriveHard Drive
Samsung Pro Series 840 256gb (Celapaleis) 1 TB Samsung Drive (Pack Yak) Samsung Pro Series 850 256gb (Celapaleis Reprise) Phoenix Down (External Backup) 
CoolingCoolingOSMonitor
Noctua D14 Kraken x61 Push-Pull for GPU Windows 10 Pro Catleap 2560x1440 @ 60hz 
KeyboardPowerCaseMouse
CM Storm Trigger Brown Switches EVGA Supernova 1000w P2 Corsair 540 Air Razer Abyssus 
Mouse PadAudioAudioAudio
Steelseries QcK Mini Blue Yeti Microphone Objective Dac + Amp Sescom AB Switch 
AudioAudioAudioAudio
Krk Rokit 6 G2 x2 Radioshack Ground Loop Isolator (For Rokits) Wooaudio Aluminum Headphone Stand Sennheiser HD 800 
AudioOther
Logitech c920 Webcam NZXT Sentry Mix 2 Fan Controller 
  hide details  
Reply
post #9748 of 19540
Quote:
Originally Posted by OutlawII View Post

Put c-stats to Auto

Same result

post #9749 of 19540
Quote:
Originally Posted by 7ranslucen7 View Post

Same result

how long have you waited it doesn't kick in right away
post #9750 of 19540
Quote:
Originally Posted by 7ranslucen7 View Post

Same result


You're using the wrong version of CPU-Z and looking at the VID sensor and not at your vcore.

VID does not drop, vcore does.

Isn't this in the guide?

I have not answered this question 20 times in this thread alone, it must be 30 by now - and that's only half of the times i've answered this on OCN.

Use 1.64.0 for Haswell only until they fix it (but i'm not counting on it since they used the wrong sensor for 6+ months)
Edited by Cyro999 - 2/8/14 at 11:08am
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
Insert Name Here
(14 items)
 
  
CPUMotherboardGraphicsRAM
6700k Asus Maximus VIII Hero Gigabyte Aorus Xtreme 1080ti Corsair LPX 2x8GB 3200c16 
Hard DriveHard DriveCoolingOS
Old Seagate HDD Samsung 850 EVO Thermalright Silver Arrow SB-E SE Windows 7 Home Premium 64 bit 
MonitorKeyboardPowerCase
Asus PG258Q (240hz + Gsync) WASDKeyboards.com v1 semi custom w/ mx browns, ... Superflower Golden Green HX550 Air540 
MouseMouse Pad
Logitech G Pro Qck+ 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel CPUs
Overclock.net › Forums › Intel › Intel CPUs › Haswell Overclocking Guide [With Statistics]