I did mine with both e-w and n-s config with 2 120x38mm without any problem at all before I switched to the Mega Shadow.Quote:
Originally Posted by Havox
http://img197.imageshack.us/img197/5158/img0311g.jpg
Printable View
I did mine with both e-w and n-s config with 2 120x38mm without any problem at all before I switched to the Mega Shadow.Quote:
Originally Posted by Havox
http://img197.imageshack.us/img197/5158/img0311g.jpg
I am having some problems loading into windows with an uncore of 4200mhz or so, I want to run the ram at 1000mhz+, although I can not get into windows with such speeds, so I am forced to run 1600mhz and 3600mhz Uncore. I tried upping the Vtt to something like 1.5V and nothing and the Vcore was upped to 1.5V, just to see if it worked. Now the Ram I left at 1.65V, it is rated for 2000Mhz at 1.65V. Stock timings I had cas 9. Now any settings I am missing when trying to run such a high Uncore? Please enlighten me.
71proste
Try the bolded changes:
Frequency Control
CPU Clock Ratio ( 21X )
CPU Host Frequency (Mhz) ( 215 )
MCH Strap ( Auto )
CPU Uncore Frequency ( Auto )
CPU Clock Skew ( 0 ps )
Spread Spectrum ( Disabled )
PCIE Frequency (Mhz) ( 107 )
Memory Feature
Memory Speed ( )
Memory Control Setting ( Enabled )
Memory Frequency ( 1722Mhz / 2:8 )
Channel Interleave Setting ( 6 Way )
Rank Interleave Setting ( 4 Way )
Memory Low Gap ( Auto )
tCL Setting ( 9 )
tRCD Setting ( 9 )
tRP Setting ( 9 )
tRAS Setting ( 24 )
tRFC Setting ( 88 )
Command Rate ( 1t )
Voltage Control
EVGA VDroop Control ( Without VDroop )
CPU VCore ( 1.3875 )
CPU VTT Voltage ( +75 )
CPU PLL VCore ( 1.65 )
IOH PLL VCore ( 1.65 )
DIMM Voltage ( 1.65 )
DIMM DQ Vref ( +0mV )
QPI PLL VCore ( 1.275 )
IOH VCore ( 1.1 )
IOH/ICH I/O Voltage ( 1.500 )
VTT PWM Frequency ( 610KHz )
CPU PWM Frequency ( 1210 )
CPU Impedance ( Less )
QPI Signal Compensation ( Auto )
ICH VCore ( Auto )
CPU Feature
Intel SpeedStep ( Disabled )
Turbo Mode Function ( Enabled )
CxE Function ( Disabled )
Execute Disable Bit ( Disabled )
Virtualization Technology ( Disabled )
Intel HT Technology ( Disabled )
Active Processor Cores ( All )
QPI Control Settings ( Enabled )
QPI Link Fast Mode ( Enabled )
QPI Frequency Selection ( 4.8GB/s )
OC Recorvery ( Disabled )
PnP/PCI PCI Express
Maximum Payload Size ( 4096 )
Michowski
Try these changes:
CPU VCore ( 1.35v )
CPU VTT Voltage ( +200 )
CPU PLL VCore ( 1.65 )
IOH PLL VCore ( 1.65 )
QPI PLL VCore ( Auto )
QPI Frequency Selection ( 4.8GB/s )
1.375 actual should take you pretty much anywhere you want to go on air or water. I generally tell people to keep it under 1.4V for 24/7 usage, as long as their temps are manageable.Quote:
Originally Posted by Michowski
Thanks for the reply capper. Im currently testing for 4.2ghz. I got up to 4.0ghz stable. I ended up before just going back to defaults and starting over. I think before I just got lost in what I was changing and didnt stay on top of it. I got prime running now so we'll see how it goes. One article that I read and started using for a guide helped me a lot. Its the voltage guide on evga forums. Tells ya what you might want to change for whatever happens when you test for stability. http://www.evga.com/forums/tm.asp?m=...09&mpage=1&key. A bit lengthy but definatly a must read if you want to OC and are new.
Highest temp ive gotten so far is 73c on e-leet testing for 4.2ghz. Right now there all just sitting at 70c
hmm thanks for the pic theres no way mine can go that far there. NB heatsink cover seems to be crooked anyway to fix?Quote:
Originally Posted by iTravis
I just got a blue screen after about 20 minutes or prime. The error was irql not less then or equal to. My current settings are as follow...
Frequency Control
CPU Clock Ratio ( 21X )
CPU Host Frequency (Mhz) ( 200 )
MCH Strap ( Auto )
CPU Uncore Frequency ( Auto )
CPU Clock Skew ( 0 ps )
Spread Spectrum ( Disabled )
PCIE Frequency (Mhz) ( 100 )
Memory Feature
Memory Speed ( standard)
Memory Control Setting ( Enabled )
Memory Frequency ( 1067Mhz / 2:8 )
Channel Interleave Setting ( 6 Way )
Rank Interleave Setting ( 4 Way )
Memory Low Gap ( Auto )
tCL Setting ( 7 )
tRCD Setting ( 8 )
tRP Setting ( 7 )
tRAS Setting ( 28 )
tRFC Setting ( auto )
Command Rate ( 1t )
Voltage Control
EVGA VDroop Control ( Without VDroop )
CPU VCore ( 1.36875 )
CPU VTT Voltage ( +175 )
CPU PLL VCore ( 1.65 )
IOH PLL VCore ( auto )
DIMM Voltage ( 1.65 )
DIMM DQ Vref ( +0mV )
QPI PLL VCore ( 1.125 )
IOH VCore ( 1.1 )
IOH/ICH I/O Voltage ( 1.500 )
VTT PWM Frequency ( 370KHz )
CPU PWM Frequency ( 940 )
CPU Impedance ( Less )
QPI Signal Compensation ( less )
ICH VCore ( 1.05 )
CPU Feature
Intel SpeedStep ( Disabled )
Turbo Mode Function ( Enabled )
CxE Function ( Disabled )
Execute Disable Bit ( Disabled )
Virtualization Technology ( Disabled )
Intel HT Technology ( Disabled )
Active Processor Cores ( All )
QPI Control Settings ( Enabled )
QPI Link Fast Mode ( Enabled )
QPI Frequency Selection ( 4.8GB/s )
OC Recorvery ( Disabled )
no dice I just tryed again to get it in that config and its by like 2 mm that it cant fit.
TryQuote:
Originally Posted by Michowski
QPI PLL VCore ( 1.125 ) to 1.2V
IOH VCore ( 1.1 ) to 1.2V
and PCIe 101-102MHz.
hint hintQuote:
Originally Posted by Michowski
vcore to vtt ratio.
dunno y u didnt check that.
based on this is the increament of ioh voltage
and based on cpu /qpi impedence is qpi pll voltage.
and how to know what to adjust
use ure rtl's as indicator..
Thanks but no changes :( same story,I can run prime95 215x21 ht on/off no problemsQuote:
Originally Posted by Capper5016
http://img38.imageshack.us/img38/204...5stable.th.jpg
but bclk 215+ totally unstable even for spi 1M
on re2 bclk 218-220 spi32m with no problems at all
it seems like bclk 217 is the max bclk for this board:shrug: spend 5 hours yesterday night and still didn't find solution
Ahh ok, I finally figured il just up the vcore and then just keep bumping it up in speed there after. And whats rtl's and how do I use them as indicator's?
try IOH pll to 1.5 undervolted, helped me the other day benching wprime @ 4.2ghz on water. Wouldnt finish 5% then made the change and it was stable.
I have no problems booting at 21x220 4 cores + HT, and 12x240 2 cores enabled. Its not a board problem. Did u try QPI SIgnal compensation -> less ?Quote:
Originally Posted by 71proste
use cpu tweakerQuote:
Originally Posted by Michowski
give u example
on bclk 4.2 200 dram 2000mhz Cl7 rtl is 53/55/57 for A/B/C
sometimes theres a change like 53/56/58 which is not stable
so i had to change my dqvref for channel A and it then become very consistent 53/55/57 which is how i passed 12 hours prime.
for 2050bclk 4.3ghz etc etc etc.. was the same
this is for dram multi 10
check for others
u will notice a trend
where inconsistencies will happen one day pass easily 20 mins
others will fail in 1 min
then u will also see the scaling of ure proc and vtt voltage to dram voltage etc etc to bclk increament.
this is how to clock stably and consistently.
increasing voltages without a reasoning especially huge jumps from 1.1qpi to 1.2qpi is pure stupid for mere increments of 2 bclk.
test is always min fft 128 max fft 4096 in windows diagnostic mode 5.4gb.
then u will see ure vcore to vtt ratio's and dram increaments.
always noticed that ultimately every voltage affects rtl's.
test it.
example
dram multi 8 for bclk 200 normally stock vtt around 1.21-1.23 CL7 rtl is 50+
but if u increase the vtt to 1.4v.. rtl is 60+
test it.
also this is why many ppl are not entering the 8 hours prime table thread.
cause increasing voltage randomly.. without reasoning is pure silly.
pcie clock is related to qpi clock.
for bclk 200.. theres no need to increase pcie clocks.
and hence no need to raise qpi pll or ioh voltage. since theres no bottleneck at all.
of all ure hardware ure dram chipset is the most limiting factor for bclk.
u need to test and rule this out
sometimes bad dram chipset strengths just pure bad luck.
elpida hypers however .. NP's with BBSE's
yeah...some people had no problems with 240+ on air but that's not the pointQuote:
Originally Posted by The_Game
i also never said this board sucks :)
i know it might be the cpu but why RE2 has no problems with spi 32m bclk 218-220?220-222 is the max for 98% RE2
when oc cpu and push over bclk 216 the ratio goes down from 21 to 20
this fact makes me feel it's more board(my board) issue than cpu
i found more people with the same issue i hope i'm wrong :(
I turned my vcore down to 1.364v, ran another test and after 10 minutes i got a blue screen that said something system exception.
Never tried R2E. I was using T3eH8 before and with the same CPU i was able to boot at the same bclk ~227-8 this was also the max with E760 in hot summer days. Also i`ve never had problems with turbo throttling on classified the multi drops from 21 only when reaching arround 95 degrees, and it drops to 17 which is a normal high temperature throttling. You can try lowering the multi to see if u can reach higher bclk, also try diffrent settings for B2B for me 3 works best.Quote:
Originally Posted by 71proste
re2 is much worst than evga if we talk about bclkQuote:
Originally Posted by The_Game
b2b?sorry what do u mean?
i tried ratio 16-21 and it's same
B2B Cas Delay its an option added in S61A/B BIOSes, before that it was only available on asus MBs now GB included this too. This option is similar to PL on 775. I dont think that this will help you for higher bclk but who knows you can try :).Quote:
Originally Posted by 71proste
Quote:
Originally Posted by 71proste
U sure? not anymore with newer bios i think.
http://valid.canardpc.com/show_oc.php?id=667814
got it too 245 at the class 760
what do u plan to avhieve with pure unfounded guess work.Quote:
Originally Posted by Michowski
overclocking nowadays seems to be just pumping voltages
lol.
know thy hardware and its limitations.
this requires empirical reasoning on increasing settings
gave a holy grail hint already
same as that unfounded uncore higher than 2x issue.answer so simple rtl.
My 759 died today im gutted as its going to be a very long rma wait.:/
what was your max bclk on re2 before bios changing?Quote:
Originally Posted by Rol-Co
we still talk about 2-3% of users :D
if I knew I wouldn't change my re2 to evga :( serious
temp is not a problem as you can see i run prime95 215X21@4.5Ghz and the max temp was around 80'CQuote:
Originally Posted by The_Game
it's so weird no problems with 215x21 but 216x21 is so unstable,the ratio drops to x20 :(
b2b works same,no difference
i can try lower ratio like 12-16 but this not what i am looking for
i tried 17-21x already but same story,I used same or similar templetates and spend 3 days working on it-still no progress...
Max bclock is cpu related not the board, so a god chip is needed too. old bios have the 222 wall like some board still have because of the high qpi.Quote:
Originally Posted by 71proste
New bios solved that issue and using 1504.