i7 Batch No# Overclock Database

[CryptiK]

This will be a self updating database of i7 overclock results.
Doing it manually is too arduous so this will make it much easier for me and everyone else. This will keep them all together instead of threads trailing off and no longer being updated by the OP.

Please post a screen shot of the settings/OC you upload in this thread - you must show a screen of at least a 1 - 2 hours Prime 95, 1 - 2 hour OCCT or Linx 5 - 10x Max Mem run. If you do not, I will remove your data from the spreadsheet. I'm Please respect the submission rules, this will help everyone. You can fill in the form first then post a screenie later though.

Ensure you state your FULL BATCH NUMBER, IE: 3xxxAxxx - just writing the first 4 digits and letter is useless.

If you use different cooling for your max OC, please state the cooling type in parentheses after the 24/7 OC cooling type. For example: Air (LN2 Max)

Fill out the following form with your information. Your information will be added to the spreadsheet.

i7 Batch No# Overclock Form

View or download the spreadsheet here:

i7 Batch No# Overclock Result Spreadsheet

This thing can just take care of itself now, so just download/view the spreadsheet to see everyone's OC's as they are added.

To download the file, Click FILE then select DOWNLOAD AS then choose the format you want.

EDIT 1 - Form updated to include fields for socket type (s1156 or s1366), Vtt (QPI), BCLK and Multiplier.

[dan7777]

heres mine... fully stable max setting 10 runs... sorry for the double pic

[FtW]

Added myself to the spreadsheet & a few screenies..

[Sam_oslo]

Xeon W3520 (#3848A352) @4.41GHZ(210x21) 1.32v On air.

VVT=1.23/1.26v (BIOS/load)
Uncore x18 (3780MHZ)
3x2GB @1680MHZ 7-8-7-21 1.65v


1 hour OCCT stable



5-run LinX stable


HyperPi32M stable

[Sniper]

Adding mine, W3520 3848A352.

Water cooling, Laing DDC, HK 3.0, PA120.3 + TFC Xchanger 360.

Ambient/Rad air intake temp 22.1 degrees.



Without HT:

[CryptiK]

Nice work, keep them coming.

Even if you have posted before somewhere else, post it up here so we can finally have everything in one place

*Rules updated and relaxed somewhat:

Tests now accepted:

OCCT 1 - 2 hours, Prime 1 - 2 hours, Linx 5 - 10x max mem.

[Vinas]

What about Orthos, Intel Burn Test, etc.... What about LGA 1156 cpus, are those ok? It would be nice to compare those as well.

[CryptiK]

Orthos as far as I know, only loads 2 threads - prime 95 v25.9 has 4 or 8 threads as required and is identical to Orthos in stress level - Orthos is just Prime95 with a sexy GUI.

Intel burn test is acceptable as long as 'max' stress is used, 'normal' is way too easy.

This spreadsheet will be s1366 only so as to keep things distinct.

[Vinas]

Orthos as far as I know, only loads 2 threads

Using a command line switch or batch file will load four, eight, (or whatever you want) threads. Several people use this and it works well so that's why I mentioned it...

Intel burn test is acceptable as long as 'max' stress is used, 'normal' is way too easy.

Isn't the standard much better than one to two hours of prime?

This spreadsheet will be s1366 only so as to keep things distinct.

Maybe this thread would be better served having both sockets together? This would help deduce which chips perform "better". Even if better is not higher clocks, but maybe lower heat, etc. I think having a socket type delimiter would be enough distinction for most folks. Unfortunately, with all the other 1336 i7 thread out there, without including the new chips this thread might fall by the wayside pretty quickly.

[CryptiK]

Using a command line switch or batch file will load four, eight, (or whatever you want) threads. Several people use this and it works well so that's why I mentioned it...

Isn't the standard much better than one to two hours of prime?

Maybe this thread would be better served having both sockets together? This would help deduce which chips perform "better". Even if better is not higher clocks, but maybe lower heat, etc. I think having a socket type delimiter would be enough distinction for most folks. Unfortunately, with all the other 1336 i7 thread out there, without including the new chips this thread might fall by the wayside pretty quickly.

I wasn't aware people do that with orthos. I guess its fine so long as there's as many threads as cores/HT cores (HT on = 8 threads, HT off 4 threads).

Linx vs prime stability is a contentious issue. In mine and other experience Linx can pass, then fail Prime, sometimes in a lot less than 2 hours. For others its the reverse. I really just want some proof the OC's are somewhat stable, so a decent length of time for any real stress testing program on max stress should be fine for that purpose.

After some consideration, I agree I'd like to see a database for s1156 CPUs too, and I think this self updating format is the best way of doing this. I have added a socket designation field to the form, so s1156 CPUs can be added also. Thanks for the suggestion.

[Vinas]

[Onoff312]

My brand new i7 setup watercooled by Dtek Fusion V2 and Pa120.3 temps are just as load finished load was 70-75, Getting new block soon hope to see improvments and sorry about the 50minutes i have a house open and had to leave.

[road-runner]

i7-960 batch 3938A611, foxconnflaming blade gti, water cooling heatkiller block.

[hetsaq]

Batch #3910A372. This is on air. Noctua NH-U12P With 2 120X38MM 97CFM Delta's

CPU 1.368V
QPI 1.35V
PLL 1.82
IOH 1.18

[doz]

Just got this. Could pass wPrime and SuperPi but linx fails after 5 passes or so. Under 80c load as well in LINX!

[rge]

5 ghz max OC on water, 1.58v, ambient 18C

4.5 ghz, stable prime 11+hrs, linx max mem 20+ on water, 1.39v cpuz load, ambient 23-24C

3847B005 i7950




http://valid.canardpc.com/show_oc.php?id=707206

[CryptiK]

Just a reminder to add your overclocks.

We are the forum with the least number of properly documented overclocks! We are supposed to be a leader in terms of overclocking but we are not looking very good at the moment.

Come on guys, get your OC entered!

[AlfaBoy]

I already submit my result and here is my screen:

[CryptiK]

Nice work there ^.

It's important that as many people as possible post no matter what you overclock is, especially since these newer batches of CPU's seem to be more hit and miss.

The more information we can get the better.

[trn]

W3520 Batch 3945A906:

I've had 3 i7-920's (all in the 39XXX range) and this Xeon. The w3520 is the best chip by far; I do believe the urban legend that Intel figured out how to bin the chips better and Xeon's are better than i7's. I think Intel figured out the better binning after 2008 because in '08 it didn't seem to mater if the chips were 920's or w3520's.

4.2GHz 24/7


4.5GHz run
Cooled by water and cold weather, over-volted all settings because I didn't have time to tweak the voltages to their min's.

[trn]

If possible I would suggest to make the spreadsheet sort by batch number instead of date the data was entered. It is a spreadsheet all about batch #'s.

[CryptiK]

If possible I would suggest to make the spreadsheet sort by batch number instead of date the data was entered. It is a spreadsheet all about batch #'s.

Since google docs is limited in it's abilities, I can't make it auto-sort by batch number. It also causes issues if I move the timestamp column, so it has to stay where it is. However I can manually sort it by batch number so they go in ascending order (oldest batch first). Done now.

Make sure you post your full batch number! ie: 3XXXAXXX

3945A is not a full batch number and is essentially meaningless.

Also can we get some image cropping & resizing happening so there's no need to scroll side to side? Surely we all know how to use photoshop

[trn]

Figured something like that might be the case. Sorry about my incomplete batch #. I realized my error after I reopened the spreadsheet and saw mine vs everyone else's.

3945A906 is my full batch number.

Since google docs is limited in it's abilities, I can't make it auto-sort by batch number. It also causes issues if I move the timestamp column, so it has to stay where it is. However I can manually sort it by batch number so they go in ascending order (oldest batch first). Done now.

Make sure you post your full batch number! ie: 3XXXAXXX

3945A is not a full batch number and is essentially meaningless.

Also can we get some image cropping & resizing happening so there's no need to scroll side to side? Surely we all know how to use photoshop

[CryptiK]

Not a problem I have edited the spreadsheet to reflect that. Thanks for your contribution, and good suggestion about sorting by batch number. It is definitely easier to read & find a batch now.

[sRHunt3r]

Intel i7 920 Batch # 3938B024
Speed : 4410Mhz Ht on
VCore : 1.3250
Vtt : 1.355
Cooler : Thermalright Venomous-X

Prime95 1 hour / Cpuz / Real Temp