My l039b470 also needs 1.52v to be linx stable at 5.0Ghz . So 1-35v~1.4 chips are the gems :DQuote:
Originally Posted by Johnny87au
Printable View
My l039b470 also needs 1.52v to be linx stable at 5.0Ghz . So 1-35v~1.4 chips are the gems :DQuote:
Originally Posted by Johnny87au
No...you're right!Quote:
Originally Posted by Haes
Mine is running (with Prime 26.4, which is much harder!!!) @4,5G@1,24 and 5G@1,408 - 1,416
Abit more progress with 1.35v 5.05Ghz LinX stable and Intelburn with improved timing on memory and abit higher Freq!
http://img534.imageshack.us/img534/2...7261tlinx1.png
Mine does 4.9 with 1.392 but needs a couple bumps for 5.0. Hoping a BIOS update will help out and I still am learning the platform so more stability might be possible higher up, haven't tried dropping PLL voltage yet. Just reinstalled Win7 with SP1 so now I can get back to clocking.
i have installed SP1 for Windows 7 but still i have low Gflops...
Disable HT and get latest linpack ;)
Why do i have to turn off ht?
Linpack gives less GFlops with HT enabled. The number of GFlops is irrelevant as long as the system is stressed properly, though...
Quote:
Originally Posted by zalbard
That is the problem though. With HT on its not stressing the cpu to the max and that is why the Gflops are low with HT on. Temps will jump up with HT off.
I don't have SB so I can't comment on that. Using SP1?
I know that all previous CPUs run much hotter with HT enabled during LinX stress test.
Linx won't stress Sandies hard enough!Quote:
Originally Posted by bro20000
Try Prime 26.4 and customize the run to 960K!
You will see it is much harder to get it to 5Ghz with your voltage!
You don't need to disable HT , just use IntelBurn test and select 4 threads when testing stability and max temps. Here i did a quick run for you guys , HT is enabled one run is 4 thread produces much higher temps and stresses cpu more , below run is with 8 threads temps don't even reach 80c.
SP1 + IntelBurn + merged with latest linpack module.
http://img842.imageshack.us/img842/9364/29660516.jpg
True,linx + AVX 4 threads=most stressfull test.better than prime ;)
Agreed!!!Quote:
Originally Posted by emertX
My new 2500k at 4.9 ghz is linx stable (20 runs, all memory used) and OCCT stable (4 hr run) on a vcore of 1.385. I haven't tried downloading SP1 yet.
However, I'm still not prime95 v26.4 stable yet with a vcore of 1.410.
Has anyone tried some real world apps like handbrake or some cpu intensive games on their linx stable (but not prime95 stable) voltages? I'm wondering if I need to keep increasing vcore to achieve prime stability in order to prevent errors in real applications (especially handbrake encoding).
Also, has anyone compared their prime95 v26.4 stable vcore to their linx stable vcore using SP1, AVX 4, and the new linpack modules? If so, could you post your values? I'm curious which needs more juice (and how much more) for stability.
I'll try handbrake tonight on my linx stable but not prime95 stable vcore and see what happens.
You are right guys, this method is more stressfull. Still my 4500mhz 1.25(in bios) is linx stable:) I have changed to 4threads in linx options and i had 4c more temps than with 8 threads.
http://img255.imageshack.us/img255/5...threads.th.jpg
Uploaded with ImageShack.us
Yes, I've been stalemated by the prime95 issue since Jan 9th. :-)Quote:
Originally Posted by drnick
I have a test scence that I setup for Lightwave 3D that on my I7-920 took over 20 minutes to render. On the 2600K it takes only 15 minutes but during that time all 8 threads are fully loaded and alot of heat is generated (not as much as linx + avx).
My stability test is now IBT + SP1 + AVX, Vantage, then 3 or 4 Lightwave renders. No BSODs and I've got to call it good. At some point it seems as tho the machine is stable and prime95 will never get enough voltage.
Just installed a new Windows 7 with SP1 and a new BIOS. Previous LinX stable voltages are NOT stable now, and heats up significantly more.
But 2500K @ 4.9 and garbage memory speed/timings (1600-CAS 9, ignoring RAM till CPU is 5.0 stable again) is pushing 130+GFlops.
I got a bad batch of i7 2600K: L041B202
4.9Ghz HT ON @ 1.476v load
http://i458.photobucket.com/albums/q...00K/49LinX.jpg
Currently testing 5Ghz 1.35v in Linx with 4Core 4 Threads and its getting there half way through now average 135GFlops!
5Ghz 1.36v in bios! Linx stable and intelburn Stable with High GFlops!
http://img510.imageshack.us/img510/6...61tlinx135.png
great result could u show some prime :D , i wish my cpu do like your , 5 ghz is not stable linux even with 1.5v , i still testing thoughQuote:
Originally Posted by bro20000
bro20000 u sure do have a great chip there mate!
Seems motherboards and BIOS versions do too! T_TQuote:
Originally Posted by bro20000
Updated to the latest 1253 BIOS and now 4.9 no longer turns off after an hour.
Hopefully I won't fall sleep before this test finishes so I can update it with a full 20 loop LinX screenshot :)
Has anyone had to tweak with the short/long power limit settings? Maybe on this new version the AUTO is more aggressive.
http://www.zhoulander.com/projects/sandy/4900-test.png
thanks mate! Have got some more results with even lower Vcore coming! Am getting weird results in temps thou! Core 1 70c Core 2 83c Core 3 80c core 4 78cQuote:
Originally Posted by Johnny87au
Am not sure if my waterblock is on perfect to get a 13c difference from core 2! Has anyone else used this thermal compound IC Diamond 24-Carat!
Anybody has tried to play with OCing a H67? I'm trying to find a small mini-ITX, and I came across a little cute monster, a ZOTAC H67ITX-A-E, but that's a H76.
I'm wondering if there is chance to fool it and OC the CPU. maybe some kind of software could push it in W7?
I know it's a tricky question, but anybody got any tips about OCing a 2006k on a H67?