Printable View
Well if it fails say 12 hrs or beyond ..... don't worry about it as i usually only do 12hrs, but i figure for fun i run longer this chip and i am 13hrs and 46mins right now on it.
Don't panic or throw :banana::banana::banana::banana: around if it fails in less then 12hrs.... me figure it out for you... usually rounding error = CPU volts or Ram timings and if you get a EXAMPLE: Error = AEB666 was expecting 455455HH error or anything like that ... usually means RAM voltage.
I cleared my CMOS because I made it madd by messing with the timings and I now notice that my memory has even gotten snappier? Also what I did different was left everything on auto cept the 5-5-5 part. I compared my screeny of my 52.3 where I timed all the settings myself and notice also that the board has loosened a couple of settings like the read to write delay is now 8 when I set it myself I had it set to 7 and the dimm fine clock delays now are tighter then what they were when I had set the settings my self.. and the row refresh cycle time I had set to 50 before but, with auto it is 52. I find this odd but it is a good thing for me I guess maybe some one else can see if this happens to them as well. I tested for stability and it is still stable as it was as well..,
http://img388.imageshack.us/img388/2302/whatnl7.th.jpg
My timings are the same as yours, same bios I think too, running 803M. Only difference is I am running 1:1 with PL10 and I have half the cache :( Great chip you have there :)Quote:
Originally Posted by truehighroller
Larry
I'm selling it for $280..Quote:
Originally Posted by Big Lar
I just thought it was rather wierd that I let it at auto and it loosened this and that and wuluah, it's snappier.
Yup, seems this board has a mind of its own :D I have set all ram parameters and gotten a lower score that way, now I only set the major four and leave it be. Pretty happy with the setup now, no need for change. Nice price on your chip, it should go fast. :)
Larry
I figure if aome one has a good watercooling setup they can push it further then me.
So what things in bios cause a drop in performance or a big drop in performance?
PL timings? CLK Skews (Delay? Advance?) That 55 timing you put on on one of the timing options?
I upped the voltage on my E8500 to 1.38V 4050Mhz and I've passed blend so far at 3 hours 23mins 54 seconds, wish I never had to leave it for 8-12 hours I want to try refine the settings and lower voltage..
I set the pl to 8 myself, didn't touch clock skews, I left everything on auto and wualah, everything is alot snappier now. My guess is because Ket put those new tables in the bios for me and my table must have been included. It loosened a few see above for the details and made the fine timing delays tighter by itself.
Tried this bios 1306 and workd great ;)
EVERYTHING on auto?Quote:
Originally Posted by truehighroller
Timings wise yes, besides the norm 5-5-5-15 everything else is set to auto and I get an increase in latency by 2.3ns yep.
Edit: Hell I bet I can hit the 49ns's if I pump 2.4v through the memory and crank it up to a 266 strap again.
Its not surprising. I discussed this some time ago with various people, tighter does not always mean better. Stressing components too much can produce lesser results than relaxing things a tad. Following screen is my results with my E4400 @ 3.3GHz, not too bad. Trying to get my CPUs FSB wall stable (425).Quote:
Originally Posted by truehighroller
http://img.photobucket.com/albums/v187/bizket/33GHz.jpg
So using FSB strap and pumping voltage gives a lower latency? Or I guess it basically just overclocks the ram?Quote:
Originally Posted by truehighroller
If I let Orthos blend test go on for 5 hours is that stable enough to start going tinkering settings possibly trying to lower voltage and all that? 12 hours repeated is a nightmare.
What settings are you guys running to get 4.1ghz or better? Of a E8500 and what version of the bios
I have a P5Q with 4 gigs of PC8500 G skills and a E8500 with a custom water cooling kits on the way, want to know what bios and settings to get it stable over 4ghz for 24/7
I can't wait to be in the 4ghz + club :)
+1Quote:
Originally Posted by Quad-Damage
I have same moboish, cpu and ram as you.
You been trying to clock to 4Ghz yet? Even underclocked I am getting memory errors, don't know if its the memory or another thing causing it though.
^^ waiting to get it from the egg
Anyone any idea how I can stabilise Core 1? Is the only way more voltage?
4.2Ghz and 1.424V the first core is going strong in prime blend test for almost 8 hours yet first core failed in 1 hour 11 mins :\
Can you run 4GHz stable @ 1.42v?
Try a custom run of 10Kfft's in Orthos or Prime for 15 minutes. If she does not pass try adjusting the gtl for that core, usually higher but you never know.Quote:
Originally Posted by Willhub
If it passes 10kfft's, look at ram timings,ram voltage and NB/vtt voltage.
Larry
I know that wasn't directed towards me but thanks for the tip.Quote:
Originally Posted by Big Lar
I'm having a not-so-fun time with this pos E. I'm at 4ghz 1.440v (after vdrop/droop LLC disabled) PL of 12:confused:, all pullins disabled and insane PLL of 1.58v, nbV of 1.3v.
Anything less on any of these values and it's crash city.
Hi, yes 4Ghz is stable at 1.42VQuote:
Originally Posted by Ket
http://smg.photobucket.com/albums/v3...ursfinally.jpg
1.34V actually, but I want to push the FSB and 500fsb causes blend to fail all the time.
My ram cant be the problem, ram is at 2.04V, was under 1066Mhz and 5-5-5-15, stock is 1066Mhz 5-5-5-15 2.0-2.1V.Quote:
Originally Posted by Big Lar
Its G.Skill PC8500.
I was running prime and adjusted the second core GTL to like 0.7+ and it still failed, dont want to go too high dont want to damage anything.
What will 10K FFT's do? if it fails with standard surely that means instability?
Quote:
Originally Posted by ReverendMaynard
In another thread related to this mobo I posted my voltages and other ram settings, have a try with them.
http://www.xtremesystems.org/Forums/...3&postcount=22
Do you know where your CPU tops out for frequency? If not I would run a high multi to establish your max frequency first and to rule out any FSB wall. If you can't run 4.2GHz with a lower FSB and higher multi, you know your CPU can't do 4.2GHz. That is, unless your willing to push on toward 1.5v.
Well I am at 9 multi atm as I'd rather have 470fsb rather than 445 fsb, but it seems equally as stable/unstable at 9.5 and 9.
I upped the GTL ref for the CPU to 0.700, I'll try prime again soon. I've ran memtet for ages now and it got one error at the start but thats it, does that mean lots of instability? Have I done enough memtest yet?
http://img.photobucket.com/albums/v3...ub/memtest.jpg
Any specific reason for 3 different values in HCI memtest ?
What? You mean the amount of memory?Quote:
Originally Posted by Leeghoofd
Well it would not let me test enough memory in one to just have 2 and split it, I supposed I could have just split the rest in 2.
Should not effect the test its still doing ram whatever.