LTT Forums CPU Overclocking Database!

[Ryan Leech]

I see that my E8400 is in red but I did run prime95 for over an hour if you look at the verification

[MeltyMoon]

I see that my E8400 is in red but I did run prime95 for over an hour if you look at the verification

Fixed!

[Lays]

I wanna post my 5150 mhz 3930k, but I can't prime95 it or it'll TJmax (1.665v)

 

I don't think it's possible for my waterblock to get rid of the heat fast enough when I prime95, it hits ~72-75c while cinebenching at 5.1, but as soon as I open prime for even a second it hits 90+ celsius.

 

I've seen a few things saying right around 1.5v 5ghz for a GOOD 3930k chip they use around 300watts, so I'd imagine 1.65+ volts is upwards of 350 to 400 atleast

[ProKoN]

I wanna post my 5150 mhz 3930k, but I can't prime95 it or it'll TJmax (1.665v)

 

I don't think it's possible for my waterblock to get rid of the heat fast enough when I prime95, it hits ~72-75c while cinebenching at 5.1, but as soon as I open prime for even a second it hits 90+ celsius.

 

I've seen a few things saying right around 1.5v 5ghz for a GOOD 3930k chip they use around 300watts, so I'd imagine 1.65+ volts is upwards of 350 to 400 atleast

 

 

intel xtu will tell you how many watts your cpu is pulling under load. the option is called "cpu total tdp" at 1.7V the 4770k draws about 160Watts

 

you dont want to try and get stable @ 1.665v, that kind of voltage is for benchmarking, not everyday use.

[Lays]

intel xtu will tell you how many watts your cpu is pulling under load. the option is called "cpu total tdp" at 1.7V the 4770k draws about 160Watts

 

you dont want to try and get stable @ 1.665v, that kind of voltage is for benchmarking, not everyday use.

 

 

Yeah I know, trying to check if it was stable at 1.545v absolutely murdered it, it takes a whole .05 volts more to achieve 5ghz now because of the damage I did when I primed it

I did a test at 1.45v 4.8 with hw-info 64, and it said 280 watts when I primed it, so I'd imagine 1.65+ is well over 375 since the TDP goes out the window above 1.5 volts

[ProKoN]

Yeah I know, trying to check if it was stable at 1.545v absolutely murdered it, it takes a whole .05 volts more to achieve 5ghz now because of the damage I did when I primed it

I did a test at 1.45v 4.8 with hw-info 64, and it said 280 watts when I primed it, so I'd imagine 1.65+ is well over 375 since the TDP goes out the window above 1.5 volts

 Try and keep it under 1.5V for 24\7 operation

[Lays]

 Try and keep it under 1.5V for 24\7 operation

 

Yeah at first I just wanted to see if it was even stable, but in the process I degraded the living crap outta it.   So basically everything above 1.4v takes .025-.05 more volts now  Kinda sucks lol.  I can't really run 4.8 anymore 24/7 because it takes nearly 1.5, and I don't feel like murdering it anymore lol.  (I have intel tuning plan, but still)

[ProKoN]

Yeah at first I just wanted to see if it was even stable, but in the process I degraded the living crap outta it.   So basically everything above 1.4v takes .025-.05 more volts now  Kinda sucks lol.  I can't really run 4.8 anymore 24/7 because it takes nearly 1.5, and I don't feel like murdering it anymore lol.  (I have intel tuning plan, but still)

 

when was the last time you confirmed stability at lower voltages.

 

I find my last few cpus required slightly more voltage for stability shortly after "burn in"

 

i have been giving my chip more volts than i dare to admit 1.7V+ and it still runs like a champ.

 

degradation typically occurs when excessive voltages are applied for prolonged periods of time hours, days, weeks

[Lays]

when was the last time you confirmed stability at lower voltages.

 

I find my last few cpus required slightly more voltage for stability shortly after "burn in"

 

i have been giving my chip more volts than i dare to admit 1.7V+ and it still runs like a champ.

 

degradation typically occurs with excessive voltages for prolonged periods of time hours, days, weeks

 

 

When I tested for 5ghz at 1.545v, I did it for 12 hours of prime.  Think that'd do it?   And when I had my previous 24/7 1.452v 4.8 ghz settings, it was stable for 10 hours of prime before I did the 5 ghz, after the 5ghz it blue screens after a few hours of prime at 1.475v, (which it previously didn't) it now requires 1.492 to be stable.

 

I tested like 2 weeks ago btw.

 

 

Where do I check that "CPU total tdp?" I just installed XTU

[ProKoN]

far right hand side  close to the bottom ." blue wrench"

 

 

12 hours prime @ 1.545V... Im skeptical that ruined anything, sounds like something I would do.

 

so your saying 4.8GHz used to require 1.452V but now needs 1.492V?

[Lays]

far right hand side  close to the bottom ." blue wrench"

 

 

12 hours prime @ 1.545V... Im skeptical that ruined anything, sounds like something I would do.

 

so your saying 4.8GHz used to require 1.452V but now needs 1.492V?

 

 

Yeah, and the 5ghz that was stable for 12 hours of prime at 1.545v, I tried 5 the other day and it took 1.587 or something like that, otherwise it'd blue screen the second I opened prime.

 

Also, I used to run 1.452 24/7 without an offset, so the voltage would never go down when the cpu wasn't in use.   (it was like that for like 3 weeks of 24/7 run-time)  not sure if that could of damaged it either.

 

 

Thanks I found it!  I may do a test later to see the TDP at 1.65v, but I'll have to ice-box my room first

 

[BoredErica]

-

Hello Meltymoon.

I am Darkwizzie from OCN forums. My Haswell overclocking thread has 12,000+ replies now. The thread is at http://www.overclock.net/t/1411077/haswell-overclocking-guide-with-statistics

I too have been keep a spreadsheet/chart/graph of overclocking results, but as you can guess, I only take Haswell CPU overclocks. The spreadsheet is here:

https://docs.google.com/spreadsheet/ccc?key=0AjXDCk5eCp1gdEdENjlDYWl6ZnV4OVlNc0lMU1V3c1E&usp=sharing

 

The spreadsheet is a little old, because I have a backlog of entries I have not added yet which I will hopefully get around to adding later this week but still contains 100+ entries in its current state.

 

I just saw this thread. Originally I wanted to add your Haswell OC results onto my chart. Problem is, most of your entries are not Haswell-based and the stability test verification seems a little iffy to me (one guy simply took a shot of his HWinfo display, that's not proof at all). However I think there are a few I can glean from the chart.

Posting to let you know I exist and to ask for permission to copy any entries to my chart. I can add that the members are from LTT.

[MeltyMoon]

Hello Meltymoon.

I am Darkwizzie from OCN forums. My Haswell overclocking thread has 12,000+ replies now. The thread is at http://www.overclock.net/t/1411077/haswell-overclocking-guide-with-statistics

I too have been keep a spreadsheet/chart/graph of overclocking results, but as you can guess, I only take Haswell CPU overclocks. The spreadsheet is here:

https://docs.google.com/spreadsheet/ccc?key=0AjXDCk5eCp1gdEdENjlDYWl6ZnV4OVlNc0lMU1V3c1E&usp=sharing

 

The spreadsheet is a little old, because I have a backlog of entries I have not added yet which I will hopefully get around to adding later this week but still contains 100+ entries in its current state.

 

I just saw this thread. Originally I wanted to add your Haswell OC results onto my chart. Problem is, most of your entries are not Haswell-based and the stability test verification seems a little iffy to me (one guy simply took a shot of his HWinfo display, that's not proof at all). However I think there are a few I can glean from the chart.

Posting to let you know I exist and to ask for permission to copy any entries to my chart. I can add that the members are from LTT.

I didn't really want to exclude anyone when it came to submitting entries into this, that's why mine are ranked with colors. Red means iffy, anything higher provides a certain sense of certainty, especially the greens or purples . But you do have my permission to copy the information from this chart, maybe just post a link to this thread somewhere in your description?

[BoredErica]

I didn't really want to exclude anyone when it came to submitting entries into this, that's why mine are ranked with colors. Red means iffy, anything higher provides a certain sense of certainty, especially the greens or purples . But you do have my permission to copy the information from this chart, maybe just post a link to this thread somewhere in your description?

Sure, will do.

I have the same sentiment about exclusion. A few members attacked my chart's credibility, because they wanted verified 12 hr prime on every single entry. I've decided to do a normal entry, and an optional verification status that is harder to meet, and if a member meets it, I just tick the box. 

[ProKoN]

@Dark_wizzie

 

Do you mind if I put a link to your stats doc in my guide?

[BoredErica]

@Dark_wizzie

 

Do you mind if I put a link to your stats doc in my guide?

No problem. I would appreciate a reference about where the spreadsheet came from, but that is not required.

[ProKoN]

No problem. I would appreciate a reference about where the spreadsheet came from, but that is not required.

 

Intel Haswell Overclocking Database. This Database is Authored and maintained by @Dark_Wizzie. A highly regarded member of Overclock.net. Thank You

 

https://docs.google....c1E&usp=sharing

[Faa]

@Lays lock your BCLK to 100.00MHz they tend to have stability issues a lot when the bclk is overclocked, some people were claiming it degrades the chip faster. Think my 3930K degraded as well, I could do Cinebench at 4.9GHz with 1.42V now it won't even render a single block

[Morfy]

Has anyone got any idea why, for the love of god, when i go from 4.3ghz to 4.5 ghz on a 3570k, i get worse scores in cinebench? But in super Pi is getting better, sort of.

[ProKoN]

Has anyone got any idea why, for the love of god, when i go from 4.3ghz to 4.5 ghz on a 3570k, i get worse scores in cinebench? But in super Pi is getting better, sort of.

 

 

watch your temps close. monitor all cores

 

your cpu is most likely overheating and throttling.

[Ryan Leech]

I tested my new E8500 @4GHz for 7 hours and it proved to be fully stable and not to mention quite cool at a max of 49C

[skitz9417]

heres mine amd fx 6300 @4.8ghz on 1.488v 

 

 

[Sundin]

Quick question:

I'm quite inexperienced with overclocking, but I've trying to play with my new rig.

I got 4.4ghz with 1.200 core voltage.

Is it any good?

It's rock solid stable! Been running intel extreme tuning utility since last night and core temps never passed 82C and average of low 70.

For some reason I can't run aida64 stress test or prime95, cause these tests override the voltage I've set when using asus adaptive mode. And I don't wanna use a fix voltage when iddleing.

My rig is a

4770k

Asus z87i pro

Cooled by noctua nh-u12s

Memory at stock it's a pair of corsair dominator 1600.

Any advices for a beginner?

[whoohaaah1]

Quick question:

I'm quite inexperienced with overclocking, but I've trying to play with my new rig.

I got 4.4ghz with 1.200 core voltage.

Is it any good?

It's rock solid stable! Been running intel extreme tuning utility since last night and core temps never passed 82C and average of low 70.

For some reason I can't run aida64 stress test or prime95, cause these tests override the voltage I've set when using asus adaptive mode. And I don't wanna use a fix voltage when iddleing.

My rig is a

4770k

Asus z87i pro

Cooled by noctua nh-u12s

Memory at stock it's a pair of corsair dominator 1600.

Any advices for a beginner?

 

 

you just answered yourself. prime/aida are synthetic tests which overvolts any adaptive/auto voltages. for stress testing (esp synthetic stress tests) always do fixed/manual voltage. soon as you are ok with the stability, then go back and put adaptive for 24/7 (i usually do your stable voltage +0.005 to +0.010 adaptive). 

[Thylaine]

I made a typo in my form should be 4.5 Ghz not 45 lol