Bloodrage P09 bios

[Lord Trident]

G25 was when the table change

sorry PCI-e was always 102 worked and 103 was knowen to BSOD so you luckly to get 103
so you need to do the hardware mod to get higher PCI-E

[To(V)bo Co(V)bo]

Hey lord any info as to if the Xeon memory mutipliers will get unlocked like the classified and rampage?
future developement I hope

[ritchwell]

well i just flashed to p09 from p06, some good some bad. The good is that i am finally able to get my ram to the speed it should be running instead of 1083 when it posts, with default setting i can overclock to 155blck instead of 150 blck so increace of 5. Now for the bad it seems i have to increase voltage (was at 1.36 core voltage now at 1.392) to not get BSOD when running prime to get back to my overclock in p06, since its still summer i really don't want to down clock from the setting i had which was stable in p06.

[Mescalamba]

well i just flashed to p09 from p06, some good some bad. The good is that i am finally able to get my ram to the speed it should be running instead of 1083 when it posts, with default setting i can overclock to 155blck instead of 150 blck so increace of 5. Now for the bad it seems i have to increase voltage (was at 1.36 core voltage now at 1.392) to not get BSOD when running prime to get back to my overclock in p06, since its still summer i really don't want to down clock from the setting i had which was stable in p06.

Um.. only 155 BCLK with or without changing VTT/uncore voltage? Cause.. I´m not sure how for you, but for me, BCLK is linked with uncore voltage.. I can have stable 200 BCLK with default vCore.. but not without changed VTT/uncore volts..

Usually I´m running 160 BCLK, default vCore and +220mV VTT (mainly cause of my ram.. needs lots of VTT). My vCore is max 1.28V.. vDroop enabled.

[ritchwell]

Um.. only 155 BCLK with or without changing VTT/uncore voltage? Cause.. I´m not sure how for you, but for me, BCLK is linked with uncore voltage.. I can have stable 200 BCLK with default vCore.. but not without changed VTT/uncore volts..

Usually I´m running 160 BCLK, default vCore and +220mV VTT (mainly cause of my ram.. needs lots of VTT). My vCore is max 1.28V.. vDroop enabled.

Yeah after flashing i just tried to see how far i can increase my blck until it can not go to windows. Right now im at 3.8(180blck x21 rounded off) with +40 VTT and uncore +220. Core voltage from cpu-z says 1.392v. I also have a C0 also, thinking about getting a D0.

[Lord Trident]

Xeon has someone tested and have a problems as i not hear of any problem
so if there not a problem then why fix

and please people who complain about increasing the voltage like richwell this is fine
and the vid table from intell in the pass were wrong and the voltage were jumping a lot higher then what you have to set so 1.392 is fine and it may have even been going higher in the past.

[saaya]

well i just flashed to p09 from p06, some good some bad. The good is that i am finally able to get my ram to the speed it should be running instead of 1083 when it posts, with default setting i can overclock to 155blck instead of 150 blck so increace of 5. Now for the bad it seems i have to increase voltage (was at 1.36 core voltage now at 1.392) to not get BSOD when running prime to get back to my overclock in p06, since its still summer i really don't want to down clock from the setting i had which was stable in p06.

what do you mean with the mem at only 1083 thing?
only 150 bclock?
i can get close to 200 without increasing any voltages...
what multipliers are you using?

and about vcore, check actual vcore with a dmm if you can, might be the same actual vcore

[PiLsY]

Cheers for the heads up on pcie 103 Lord Trident - realised myself this morning when I couldnt post from being powered off at the mains until i dropped it to 102.

Guess thats all yesterdays testing written off . Still not a lost cause yet tho!

[ritchwell]

what do you mean with the mem at only 1083 thing?
only 150 bclock?
i can get close to 200 without increasing any voltages...
what multipliers are you using?

and about vcore, check actual vcore with a dmm if you can, might be the same actual vcore

I would set my memory manually also changing the multiplier of memory to get at the speed of my XMP ram but during post it would state 1083, now with the p09 at post it would state the right frequency.

I have a C0 i7 withought changing anything in bios i wanted to see how high i can get my blck.

Sorry Trident did not mean to come out like that, its just my initial reaction since going to p09.

[Mescalamba]

Would be nice to know which memory exactly you have.. cause I runned in lots of problems just due wrong memory SPD setting or simply cause my ram sux at higher BCLK.. (no issue for Micron based ram users.. but rest, well, not that good..).

Its possible that wrong SPD can hold you from reaching higher OC.. just post your memory pls.. I´ll think what can be done.

[saaya]

Cheers for the heads up on pcie 103 Lord Trident - realised myself this morning when I couldnt post from being powered off at the mains until i dropped it to 102.

Guess thats all yesterdays testing written off . Still not a lost cause yet tho!

same here with me on G25... 103 pciE doesnt work reliably for me either, had to back down to 102... i dont think theres a change in pciE clocks with P09 compared to G25... only some vcore and vtt things seem to have changed, or maybe they updated the intel code block...

I would set my memory manually also changing the multiplier of memory to get at the speed of my XMP ram but during post it would state 1083, now with the p09 at post it would state the right frequency.

I have a C0 i7 withought changing anything in bios i wanted to see how high i can get my blck.

Sorry Trident did not mean to come out like that, its just my initial reaction since going to p09.

huh really? it works perfectly fine for me... it always shows the correct mem clocks at post for me...

Would be nice to know which memory exactly you have.. cause I runned in lots of problems just due wrong memory SPD setting or simply cause my ram sux at higher BCLK.. (no issue for Micron based ram users.. but rest, well, not that good..).

Its possible that wrong SPD can hold you from reaching higher OC.. just post your memory pls.. I´ll think what can be done.

yes, i wouldnt be surprised about that at all!
the spd detection and reading and memory timings beeing set clearly doesnt work 100% like it should even in the latest bios, so a badly written spd or badly interpreted spd can def limit the mem and bclock overclocking as some mem timings have a direct effect on max bclocks like rtl and tras and trfc...

[ritchwell]

Would be nice to know which memory exactly you have.. cause I runned in lots of problems just due wrong memory SPD setting or simply cause my ram sux at higher BCLK.. (no issue for Micron based ram users.. but rest, well, not that good..).

Its possible that wrong SPD can hold you from reaching higher OC.. just post your memory pls.. I´ll think what can be done.

I have these OCZ XMP RAM http://www.newegg.com/Product/Produc...82E16820227388

[Mescalamba]

Hmps.. I would try this..

For example 160 BCLK ..

1.65V vDIMM, +200mV VTT, 8-8-8-24, command rate 1T, tRFC 88 or 92 or more.. simply try how much you need.. it doesn´t have much impact on performace..

Memory multi x10, uncore multi x20. CPU multi 20x .. Turbo disabled, PPM/EIST/CXE disabled, HyperThreading disabled, Virtualization and Execute Bit disabled as well. vDroop enabled. vCore stock..

Yea and XMP disabled, use manual SPD ofc..

[ritchwell]

Will try setting, Thanks.

[To(V)bo Co(V)bo]

Xeon has someone tested and have a problems as i not hear of any problem
so if there not a problem then why fix

and please people who complain about increasing the voltage like richwell this is fine
and the vid table from intell in the pass were wrong and the voltage were jumping a lot higher then what you have to set so 1.392 is fine and it may have even been going higher in the past.

The processor works yes, but the memory multiplier is locked. It would be nice if it was unlocked as the classified and rampage does.

[PiLsY]

Cant seem to get anything 100% stable out of this board.

Must be something im missing but im not sure what. Been using By SPD mem timings so shouldnt be that. Can't even get 205bclk through prime for over 6 hours. Worked backwards from 228fsb, tried millions of combinations of vtt, vddr, vcore, vioh - enough to say that its not voltage stopping me. I wonder if the PCP+C silencer 750 isn't as compatible as previously thought. I figure Im pulling about 600watts total, maybe 650. If I run furmark, superpi 32m on core 0 and prime 95 3 workers on core 1,2,3 the psu fan kicks into hairdryer mode. Rebooting from a stable prime95 run will often cause a hang and oc recovery needing a gradual work back up to the bclk it was at before (keep voltage the same, 166mhz then 175mhz then 185mhz then 195mhz and so on or itll just hang and oc recover again). Tried various cpu and ioh delays - not played with drives yet though.

Think ill have to try memory timings next, but I cant see how it can be that? Im running 8-8-8-24-1T-58 @ 1600mhz+ whereas 7-7-6-16-1T-48 is stable at 1600mhz same voltage on the EX58-UD3R.

ARGH!!!!! Days of this now!!

[oggie]

Noob question here, what difference does pcie value make ?

[Mescalamba]

PCIe let you use higher BCLK.. It's cause Intel, when they designed this platform, forget, that maybe some crazy guys will like to have BCLK over 220 (225). For example, us. So, they made dividers for BCLK:PCIe only up to those 220 (maybe less.. cause usually boards end at 200).
And no dividers over this area.. so you can go around this, by increasing PCIe freq, which let you use already existing dividers, simply because you increased both values, between which these dividers are.. eg. BCLK:PCIe. Even 102 helps a bit.. but not much.
For example with "Classified" .. at 120 PCIe, you can reach around 240 BCLK.. depends on cooling ofc. And no, its not much stable.. plus its dangerous.. and you HDD and other stuff will virtually disappear from system (cause its little bit too high BCLK for disc controlers and such stuff). But, for higher CPU OC, its usefull, especially on DICE/L2N.

Yea and one thing worth noting, continuous use of higher PCIe results usually in various dead things inside your PC (no I don't mean mosquitos) usually HDD or graphic/audio cards (graphic cards usually survive a lot.. audio much less..). For short time, it should be quite safe.. what wouldn't we do for those extra MHz on CPU core.. right?

Plus quite low values (like 105), can make higher BCLK little bit more stable..

[saaya]

Cant seem to get anything 100% stable out of this board.

Must be something im missing but im not sure what. Been using By SPD mem timings so shouldnt be that. Can't even get 205bclk through prime for over 6 hours. Worked backwards from 228fsb, tried millions of combinations of vtt, vddr, vcore, vioh - enough to say that its not voltage stopping me. I wonder if the PCP+C silencer 750 isn't as compatible as previously thought. I figure Im pulling about 600watts total, maybe 650. If I run furmark, superpi 32m on core 0 and prime 95 3 workers on core 1,2,3 the psu fan kicks into hairdryer mode. Rebooting from a stable prime95 run will often cause a hang and oc recovery needing a gradual work back up to the bclk it was at before (keep voltage the same, 166mhz then 175mhz then 185mhz then 195mhz and so on or itll just hang and oc recover again). Tried various cpu and ioh delays - not played with drives yet though.

Think ill have to try memory timings next, but I cant see how it can be that? Im running 8-8-8-24-1T-58 @ 1600mhz+ whereas 7-7-6-16-1T-48 is stable at 1600mhz same voltage on the EX58-UD3R.

ARGH!!!!! Days of this now!!

what speed IS stable?
stock speed?
after how much time does it fail/hang/freeze?

[oggie]

PCIe let you use higher BCLK.. It's cause Intel, when they designed this platform, forget, that maybe some crazy guys will like to have BCLK over 220 (225). For example, us. So, they made dividers for BCLK:PCIe only up to those 220 (maybe less.. cause usually boards end at 200).
And no dividers over this area.. so you can go around this, by increasing PCIe freq, which let you use already existing dividers, simply because you increased both values, between which these dividers are.. eg. BCLK:PCIe. Even 102 helps a bit.. but not much.
For example with "Classified" .. at 120 PCIe, you can reach around 240 BCLK.. depends on cooling ofc. And no, its not much stable.. plus its dangerous.. and you HDD and other stuff will virtually disappear from system (cause its little bit too high BCLK for disc controlers and such stuff). But, for higher CPU OC, its usefull, especially on DICE/L2N.

Yea and one thing worth noting, continuous use of higher PCIe results usually in various dead things inside your PC (no I don't mean mosquitos) usually HDD or graphic/audio cards (graphic cards usually survive a lot.. audio much less..). For short time, it should be quite safe.. what wouldn't we do for those extra MHz on CPU core.. right?

Plus quite low values (like 105), can make higher BCLK little bit more stable..

Lots of usefull info there as usual ....... learn something new every day So whats safe for 24/7 ??? 102 ??

[PiLsY]

Yeah stock speed will be stable im certain.

I've not had a single fully stable setting yet when clocking over 185bclk. Ive used no lower than 18x cpu multi and my vcore has been set to 1.5375 bios (1.51 idle 1.57 load).

I know this cpu does 3.8ghz @ default vcore and 4.2ghz @ 1.44 load. As youve seen in links before on this thread ive managed up to 4515 into windows @ 1.5375vcore so im happy to eliminate the cpu clock as causing instability up to say 4.3ghz (above that itll need testing).

Uncore wise I have been into windows at 4.3ghz+ using 1.7vtt / 1.63vcore / 1.95vdimm @ 2151mhz 9-9-9-24-1t and 4.515ghz cpu.

Ill try and find exactly what is stable and what is not later today. Should probably try single and dual channel too.

Anyway i'm going to leave the cpu @ 1.5375, set vtt to 1.54, vddr to 1.78 and see where I get.

Ive been recommended to put nb and sb volts up to 1.25v so ill try that too.

I havent managed to get anything stable enough to even optimise 6x or 10x multis. 8x is useless without cas 7 unless I really ram the ddr volts for cas 6.


I've noticed that as I increase gpu clocks my psu starts to buzz. Not a good sign i dont think. Combine it with a huge cpu clock and it pretty much vibrates! However voltages are rock solid stable and there's no particularly hot air coming out. My Corsair TX650 certainly didnt, but then I need 4xPCIE 6 pin for the 260gtx sli.

[ritchwell]

Would be nice to know which memory exactly you have.. cause I runned in lots of problems just due wrong memory SPD setting or simply cause my ram sux at higher BCLK.. (no issue for Micron based ram users.. but rest, well, not that good..).

Its possible that wrong SPD can hold you from reaching higher OC.. just post your memory pls.. I´ll think what can be done.

By the way can you be able to list the micron base RAM if i decide to upgrade. TY

[Mescalamba]

By the way can you be able to list the micron base RAM if i decide to upgrade. TY

Heh, I would ask Saaya probably if I were you, cause my knowledge about ram is limited.. almost every manufacturer have some..

Currently I know that CSX value 1333MHz CL9 are D9KPT based, as well as OCZ Blade 2000MHz CL9. And usually every Crucial ram is Micron based.. but its sometimes pretty hard to find good one.. (not all Micron chips are good).

Best are D9GTR/GTS .. but pretty expensive.

[Mescalamba]

Lots of usefull info there as usual ....... learn something new every day So whats safe for 24/7 ??? 102 ??

Heh.. good question, I've seen HDD dead after month on 115, which means thats probably not safe..

My guess is that 105-110 should be ok.. but really, I can't tell, cause I like my HW and my PCIe is on default.

Its bad that QF engineers ignored Saaya and haven't made BR with separated PCIe from SATA/SAS and such.. If that happened, we woulnd't be limited that much.. On other hand, there isn't mobo that have this modification..

[saaya]

well if you use the sas pciE then the clock should be 100 even with high pciE, at least i think so...
about what mem uses what chips, im way out of the loop, sorry...

pilsy, hmmmm def try stock speed to make sure your os isnt hosed, and def try low uncore and low mem clocks as well...
and check the error log in windows... does the system reboot or freeze? power issues are more likely to cause a reboot than freeze in my experience...

freezes are usually "soft" errors, caused by corruption... althought that does happen as a result of voltages going too high or low as well... but only if it fluctuates a bit... if theres a serious prob with power of either the board or psu, itll most likely cause the system to shut off or reboot... but im sure you know that already