Overclock.net › Forums › Benchmarks › Benchmarking Software and Discussion › Super PI 32M 4ghz efficiency challenge
New Posts  All Forums:Forum Nav:

Super PI 32M 4ghz efficiency challenge - Page 47

post #461 of 2130
Nice work Loud, impressive stuff for air. Posted this on XS, but figured I would post it here too in case you don't check one or the other and can offer some insight.

Is there anything I'm missing for getting PSC to boot at 2600, with bclk or with the 26x multi? Either just gives me 55 postcode error. Tested back on my MVF and everything is fine, same standard PSC profile timings, 2600 and boot @ 1.82v. No IMC voltage changes. But it's just not working on the formula.

I can boot in on the 2400 multi and bclk up to 2600 once in windows and it will pass 8m no problems (didn't bother running 32m) so it should be able to post with those settings I would have thought. 1.60 bios. Any help appreciated, really need PSC working on the 2600 multi for some 32m runs I have planned.

Thanks.

P.S: Samsung works just fine on Z77-OC, 2666+ no problems.
post #462 of 2130
Do you have any other PSC kits to try?

vccio and vtt? I need less on Asrock than Asus.
post #463 of 2130
Hey CL3P2O,I just tried a few sets of pis mems and your kit the 2000 6-9-6 kit was the best of the 3 different kits.This kit was as good as my Patriots,may even be better
post #464 of 2130
This is the kit here,no stability testing,havnt had any time yet.This was on air,
post #465 of 2130

yup.. where you at voltage wise? .. prolly around 1.74v or so? smile.gif For SB I was rocking them at 7-10-7-21 up to ~2200mhz with a bit more voltage for 2D.. and 7-11-7-21 for 3D action. I popped a stick just before my Ivy arrived .. So I havent had the glory of testing them on this platform yet.

 

 Nice results for certain! I cant wait to get mine back for some 32mil.. this CL10 @ +2600mhz just isnt cutting it for the times you guys are posting wink.gif

post #466 of 2130

dbl post*

 

Looking at the posted times of Mike and 636.. just wanted to point out a few things.

 

Mike - looks like your running maxmem @ ~600MB ..but your copy size is off.. try making your copy file larger. It should help free more available mem.. and bring your total to ~557xxxxxxx or 567xxxxxxx if your lucky.. higher even. You should have an initial loop time lower than 5.375s

 

636 - looks like your running maxmem @ ~630MB ..and although your initial loop time is drastically better, your available is still low [or seems to me].. Just curious if you are playing only with maxmem settings or if you have started increasing CW file size as well..

 

..compare to mine with MM @ 600MB, pf @ 384MB on C:\ ... using 632MB wazza file :  C->E only

 

 

700

 

**when i get available up to 567xxxxxxxx in Pi.. my initial loop goes down to 5.370s

 

 

 I may be asking you to share a bit much.. but would love to discuss if your up for it.

post #467 of 2130
Speed of initial loop depends on whether the best 16K result recorded in pi_rec.dat is slower or faster than 0.093s. If 16K "record" is lower than 0.093, your initial in 32M will be slower, but loops overall faster ... or so it is on s775/s1366 at least, didn't test this on IB.
Here's a pi_rec file which you might want to put in your Pi folder to see how it affects the initial - http://www.mediafire.com/?13h9jhfl678d51u

Speaking of maxmem and CW size for IB - I tested various maxmem sizes between 600 and 900 and didn't find much difference.
On the CW, however, you might want to go at least twice the maxmem. Small CWs (especially ones done via OCX tweaker) gave me higher available/system cache in task manager, but overall result was not better than using manual CW or CDT.
post #468 of 2130
Quote:
Originally Posted by Sam OCX View Post

Speed of initial loop depends on whether the best 16K result recorded in pi_rec.dat is slower or faster than 0.093s. If 16K "record" is lower than 0.093, your initial in 32M will be slower, but loops overall faster ... or so it is on s775/s1366 at least, didn't test this on IB.
Here's a pi_rec file which you might want to put in your Pi folder to see how it affects the initial - http://www.mediafire.com/?13h9jhfl678d51u
Speaking of maxmem and CW size for IB - I tested various maxmem sizes between 600 and 900 and didn't find much difference.
On the CW, however, you might want to go at least twice the maxmem. Small CWs (especially ones done via OCX tweaker) gave me higher available/system cache in task manager, but overall result was not better than using manual CW or CDT.

Thanks Sam, will try that file out and I agree manual waza is faster for me.
post #469 of 2130

Sam just ->  biggrin.gif

http://www.youtube.com/watch?v=2PFlij7kP7E&feature=related


Edited by CL3P20 - 11/21/12 at 6:32pm
post #470 of 2130
Quote:
Originally Posted by CL3P20 View Post

dbl post*

Looking at the posted times of Mike and 636.. just wanted to point out a few things.

Mike - looks like your running maxmem @ ~600MB ..but your copy size is off.. try making your copy file larger. It should help free more available mem.. and bring your total to ~557xxxxxxx or 567xxxxxxx if your lucky.. higher even. You should have an initial loop time lower than 5.375s

636 - looks like your running maxmem @ ~630MB ..and although your initial loop time is drastically better, your available is still low [or seems to me].. Just curious if you are playing only with maxmem settings or if you have started increasing CW file size as well..

..compare to mine with MM @ 600MB, pf @ 384MB on C:\ ... using 632MB wazza file :  C->E only




**when i get available up to 567xxxxxxxx in Pi.. my initial loop goes down to 5.370s


 I may be asking you to share a bit much.. but would love to discuss if your up for it.
Not sure how you figured this, but I probably was running 600mb MM. I try to do the cw manually but not sure exactly what i'm doing. I need work on it and the last time that you helped me was when sandy first came out. I did try a few different file sizes ranging from 2gb to 3.5gb. I must be doing it all wrong if you're using a 632mb wazza file.

I did try to run 32m using my MVG instead of the asrocker and times are terrible with it. I must be missing some setting in the bios or something because it's pretty much the same OS as I used with the asrocker and it's right around 8min.
Daily
(15 items)
 
  
CPUMotherboardGraphicsRAM
5960x Asus X99A II Asus GTX 1060 Dual Team Vulcan (4 x 8GB) DDR4 3000 
Hard DriveOptical DriveCoolingOS
Samsung 960 evo LG Bluray  Thermaltake Water 3.0 Ultimate Windows 7 64 
MonitorKeyboardPowerCase
Dell U3415W OCN Ducky  Seasonic X750 Corsair Carbide 600Q 
MouseMouse PadAudio
Razer DeathAdder Black Edition Glorious  Creative Sound Blaster Z 
  hide details  
Reply
Daily
(15 items)
 
  
CPUMotherboardGraphicsRAM
5960x Asus X99A II Asus GTX 1060 Dual Team Vulcan (4 x 8GB) DDR4 3000 
Hard DriveOptical DriveCoolingOS
Samsung 960 evo LG Bluray  Thermaltake Water 3.0 Ultimate Windows 7 64 
MonitorKeyboardPowerCase
Dell U3415W OCN Ducky  Seasonic X750 Corsair Carbide 600Q 
MouseMouse PadAudio
Razer DeathAdder Black Edition Glorious  Creative Sound Blaster Z 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Overclock.net › Forums › Benchmarks › Benchmarking Software and Discussion › Super PI 32M 4ghz efficiency challenge