Printable View
OK this is the P09 bios for the bloodrage
http://www.yourfilehost.com/media.ph...1_8A1F1P09.zip
dont know the fix list but it may just be g25 bios made into public.
but this is the latest and i will sort out the list of problem for the next beta
thanks
thx for the post. i got the board in today. cpu comes tomorrow.
Thanks Carl! :)
Awesome! Thanks!Quote:
Originally Posted by Lord Trident
thanks :surf:
I'm not complaining, but this is like a doctor giving you medicine, and when you ask him what it's for he says.......I don't know, try it and see.Quote:
Originally Posted by Lord Trident
funny my doctor never said stuff like that, you must have funny doctor's
as i said its mostly G25 with a few more fix so there is not a list but why dont you try it and see if it helps or not, and if you find a new bug please report
i know i found one already but i want to see if anyone else see it or not
Thanks lord trident, Im sure if it wasnt for you and the feedback from us, that we wouldnt have gotten any support at all.
Glad to see progress, at least there is light at the end of the channel.
THANKS AND KEEP'EM COMING :toast:
no changelog? too bad the past bios releases had one...
has anybody tried it and noticed any diferences to G25?
any new settings?
I ripped the p09 and g25 bios apart in my hex editor, and i can confirm they are different. Not just a rename to shut us up, some parts are similar but totally different module locations.
Build date is 07/07/09 so its not that new. Good timing tho, just in time to flash to after my leak test. thanks:up:.
Nothing odd to report except Im struggling to get over 1450mhz cas 6 without using huge vtt. Worked fine on P07 up to 1667mhz @ 1.25vtt / 1.60vdimm (then wouldnt post beyond that no matter what). Tried 10x and 8x multi. Using 6gb Crucial ballistix 1333mhz kit w/ D9GTS, 260GTX sli, Xonar D2 PCI + i7 920 c0.
Cas 5 and 7 both scale the same as with P07.
the higher you set vtt the bigger the discrepancy between actual vtt and what you see in hwmonitor.
at 1.4v its actually more like 1.3v real under load iirc, use the pads to meassure with a dmm :)
I got some issues when OCing with it (seems like BIOS doesn´t always remeber what I´ve set) but after CMOS clear, its fine.. Maybe its just my "feel" but it seems to me, that it has somewhat improved VID table, cause now my voltage at CPU never goes over 1.25V without me, increasing it.. I needed to increase to 1.3V to have 3.2GHz stable.. before I just leaved it on default and it worked. Voltage reading is same, so I guess they´ve fixed VID table.
No other problems noticed so far.. seems stable.
fixed or improved? :D
fixed would be great!
for me it always seems to display idle vcore for some reason...
in the voltage section the vcore setting is 0.935v iirc, but actual vcore in hwmonitor and meassured is 1.25v
So at this point would u guys suggest G25 or P09 or P07 or 2 of 3? I got to get mine up and running today. I have high hopes. :pimp: :party:
I remember my 955BE showed 0.975 cpu v in the bios when I first updated it for the 955BE. I set it to 1.35 before booting. I got 1.70 inside of windows....:shock2: I was on air!!! :wierd: lol
Just wondering - whats the difference between CPU turbo mode Always on and CPU turbo mode enabled when using current feedback override?
When set to always on it sometimes boots at 20x anyway, whereas "enabled" with CFO on sticks it at 21x permanently.
Cheers for the heads up on the VTT saaya - just noticed that myself :p. Im finding though that despite the vtt drop my bios value for any given speed is equal to the Gigabyte bios value even tho the BR is undervolting by about 0.08v.
It rebooted itself after an hour of Linx @ 4.1ghz HT on - giving vdimm a notch or two and trying again. Already way ahead of the same setup with EX58-UD3R rev 1.0 on cpu, memory and uncore clocks and at a lower voltage too!
My cpu vcore reports as exactly what its running - I can see power saving turn on and off and vcore rise and fall accordingly. Not tried with load calibration on as my idle is exactly what is set in bios, load is reading off dmm pad @ 0.03 to 0.08v below idle. Most excellent :up:.
P09 is the most recent one according to Lord Trident from Foxconn. Should be like G25 or better, so probably the best! :up:Quote:
Originally Posted by CobraXP
LOL! :rofl:Quote:
Originally Posted by CobraXP
onQuote:
Originally Posted by PiLsY
multi jumps dynamically based on load, lowest is 12 and highest is 22 for single threads (since vista shuffles threads from core to core quickly it doesnt work on vista, and on xp it doesnt work cause award messed up their code and never fixed it, the only way to get 22x is to disable 3 cores or 2cores and ht)
this depends on how the os handles eist though, if you set the cpu to be running at full speed all the time, idle or load (power management) then it will essentially be the same as the bios setting "turbo always on"
always on
multi is locked at 21 at all times, idle, partial load, full load... only drops to 20x if cpu current exceeds the programmed maximum (if you disable current override it should always stay at 21 or 22, even if vcore is 2.4v and the current is so high it damages or kills the cpu)
it always boots with 20.Quote:
Originally Posted by PiLsY
turbo multipliers are basically reverse thermal throttling, implemented through eist.
thats the best way to look at it, thats what it really is...
the real clockspeed of any nehalem cpu is its max turbo multi x133, for the 920 thats 2933mhz.
but since that would blow the 125W tdp envelope easily with 4 cores and 8 threads, the cpus are basically constantly throttling down.
traditional thermal throttling is:
run full speed, if temp exceeds tmax, throttle
for nehalem its the other way around:
throttle, if temp is ok, current is ok, thread load is ok, run full speed
once the eist driver loads with the os, itll say temp ok, current ok, thread load ok, and then enables 21x or 22x multipliers.
sounds like the BR vtt pwm is pretty good then :)Quote:
Originally Posted by PiLsY
your idle vcore is higher than load vcore? hmmm really weird, some people report higher load vcore, some higher idle vcore... im really starting to wonder if maybe the cpu has to do with this, and some request a certain vid for load and another for idle, and some cpus are programmed with a higher vid in idle than load and some the other way around... or the board reads the load and idle vid table incorrectly, so some vids are read too high, as a result sometimes idle vcore is read too high and set too high, and sometimes load vcore...Quote:
Originally Posted by PiLsY
seems to be the same on other boards as well though... intel really did a sloppy job making sure award and ami get their basic bios kernels working properly... theres so much stuff that still doesnt work 100% like it should when it comes to power management and nehalem cpus... even on 1156... :rolleyes:
I'm too scared to try P09. I'm currently running on G25 and I finally got it rock stable @ 4.3GHz with a Vcore-load of around 1.42v. Going from P08 to G25 screwed all my overclocks and I had to redo them from square one. So.. I think I'm going to stick with G25 unless something listed in the change log persuades me to go through all the retuning again ;) Hint hint :p
P09 is fine.. tough, can´t say that its different.. will play with OC a little..
You've got two BIOS chips on your board for this reason. ;)Quote:
Originally Posted by MagicBox
Switch the jumper on the board to the other BIOS chip and keep you OC profile saved for G25. :)
Hmps.. few notes.. its not that good as P06/P07 for C0 users. I´ve managed my usuall 4,2GHz without much problems, only one note.. this BIOS version is picky about uncore voltage:vDimm ratio. Which is quite ok for health of PC, but not good for poor suckers like me, who has ram that doesn´t work good over 1.6V. :) Basicly you need to increase vDimm with increasing uncore voltage. Yea and it boots at 103 PCI-E. :D (before it was only 102 for me).
i can boot at 103 with G25... you HAVE to increase vdimm when increasing vtt??? what??? LOL
are you sure? what happens if you set vtt to the same as vdimm? vdimm gets increased automatically?
and if you reduce it, vtt gets reduced as well? 0_o
Got up to 222 bclk in windows without too much hassle. Only been using PCIE 102 - not tried 103 yet. I have to use HUGE vtt to get over 205bclk. Tested stable @ 205 1.28vtt reported. To get 215 bclk prime stable I need 1.63vtt reported. Extra +360mv needed from bios.
Can't get more than 4.06 with HT on stable no matter what vcore. Previous best was 3.98 on the UD3R. Currently testing in Linx @ 4.33 HT off compared to previous stable best of 4.22 on the UD3R.
Moving on to uncore clock and hopefully sme mem clocking tonight.
Having great fun here :D.