thanks, just flashed F7 and noticed the PLL override setting was gone, gonna try F7E
Printable View
thanks, just flashed F7 and noticed the PLL override setting was gone, gonna try F7E
Np and welcome. :) Nice looking rig, I use the white tubing myself.
Quote:
Originally Posted by dinos22
Sorry I meant the SATA2 ports.. I need to transfer the data off of my RAID0 which is 3x640 GB WD blacks to a 2TB that I just ordered.. I just want to make sure that having the 640's hooked up to the SATA2 wont cause any immediate damage.
Naa man I'm still using 1,2,3 & 4 noproblems...Quote:
Originally Posted by aver
Nice looking rig there, Nice job on the sleeving looks like you have experience and not a nooby :)
man sleeving those power cables is something else. really like that
oops think i killed my 2600k , mine needed a lot of v core 1.65 for 5.373mhz and ran all baenchies but tonight i had bsod a few times with something like failure to something signal on secondary processor , now all it tries to do is a permanent boot loop , so everything comes on as though its starting the boot process then clicks off and tries again and again, could be board but i think its the cpu myself
dont take long with that kind of voltage huh glen?
that sucks.
you need a better chip anyways
:D yeh mate got to look on the bright side :up:,can't see it being the motherboard . it tries to post but just loops and nothing on the led panel with the boot codes ,i'll get a cpu to try tomorrow
what cooling subzero? if subzero that really isn't promising.
LOL BTW nice cable sleeving i see them in microcenter for like 5-10 bucks per extension they are sleeved like that, I would do it if I kept this thing in a case. Or if my case had a window. I would do black, when get a gold plated waterblock black tubing, gold plated blocks for my GPUs and so on.
I will be interested to see if it was the chip. I just sent my board back as it was doing the exact same thing. I thought it was the board for me as it started after flashing the backup bios.Quote:
Originally Posted by GLENBOY
Aver,
If i recall, you *backflashed* the backup BIOS with an older version, is that right?
Maybe the main bios can be backflashed, but perhaps it was backflashing the backup that killed it?
I remember WAY back, with an Abit BH6 board (rev 1.1), they released some BIOS (around the time they broke CAS latency changes temporarily), I think it was version NK, and if you backflashed to an older version, it would brick the BIOS (the dos flasher would reach around 30%, then suddenly "insta-speed" 3x faster up to 100%, then the board wouldn't POST anymore. You could backflash successfully by hot flashing *twice* (the first hotflash would brick, the second one would succeed).
There was a 1.0 board BIOS version which did the exact same brick...I think it was two bios versions after the initial release of that board. I recall posting (maybe on the old ocforums?) about warning people NOT to backflash.
LOL I still have an old spare bios chip abit mailed me from back then...
Has ANYONE here successfully backflashed the backup BIOS with alt-f12 and NOT bricked their boards? (I don't suggest you try it until you get your RMA'd replacement from the recall...)
I've pestered Dinos22 in PM, but i thought i'd post these images here anyway. Is anyone else finding the Marvel SATA 3 ports are giving slow write speeds?
Intel SATA 3:
http://i24.photobucket.com/albums/c1.../Attointel.jpg
Marvell SATA 3:
http://i24.photobucket.com/albums/c1...ttoMarvell.jpg
You be one of the first to kill one if its the chip..... My 2500K has had the max the board can set 1.7V + the LL1 overvolt & 1.9v+ through the memory for days & she loves it :DQuote:
Originally Posted by GLENBOY
This is my max overclock for 24/7.
Using L1 4.0v set in bios with everything Auto.
If you notice under load = 1.356v max cpu clock 4.7
System is 100% stable, 30+ passes on LinX a few hours ago.
This time I left everthing enabled turbo ect. Running bios F7
No problem here flashing the backup, on UD4 or UD7?Quote:
Originally Posted by Falkentyne
Also have seen many users not have issues as well, here and other forums.
Marvell is always terrible like that :DQuote:
Originally Posted by jackbrennan2008
i'll report back this evening getting a new cpu to try i had flashed the backup bios to f6x but that was weeks ago and its been 100% sinceQuote:
Originally Posted by aver
i'll try new cpu tonight so will know whether its cpu or mboardQuote:
Originally Posted by C-N
I wonder why they limited Vcore at 1.7? Seems like 1.9-2.0 would be more of an ideal Vcore limit!
Not very extreme is it.... makes me wonder how much further these chips will go as I've just seen one run x58. Maybe its something I have done but my UD5 has no x58 multi x57 is the max you can set.Quote:
Originally Posted by Lsdmeasap
Edit: Think x57 is the limit of 2500K's ... 2600's go higher just waiting to see one go quicker than x58
There is no need for more volts. If you do need more I can give you a bios for 2.1vcore heh
I dont but someone might, Its good to know its not a hardware limitation :up:Quote:
Originally Posted by dinos22
lsd,Quote:
Originally Posted by Lsdmeasap
Remember, I was talking about BACKflashing the backup, meaning to an older BIOS, not forward to a newer one. I haven't seen anyone talk about flashing the backup to an older bios....
Done it both ways on my UD5... no problem, just pump F12 3 times to switch bios's & flash it.Quote:
Originally Posted by Falkentyne
Sorry I didn't catch that, but ya I have back flashed the backup as well when going back to F3 to test some things (From both being on newer F6x or above)
uf my friend thats very bad news gigabyte has been standing the flag "use our ports" i talked to them so if intels become degraded we are f*kd :down:Quote:
Originally Posted by Lsdmeasap
I did the following :Quote:
Originally Posted by Falkentyne
Flashed to F7 and set it as my backup bios, then realized no LCC option so I went back to F7e. When I tried to flash F7e over F7 in the backup something happened and it went into infinite boot loop. I have a new board coming Monday so I will be able to tell you if my CPU or Mobo was the culprit.
They won't degrade like the Intel ones, but they are not going to be very fast is what I meant.Quote:
Originally Posted by andressergio
Also terrible for SSD and don't even think about RAID SSD on them.
P67 Intel 6Gb/s ports are what you should use!
well did a chip help or what? cus....Quote:
Originally Posted by GLENBOY
thats what mine is doing now and no more then 1.61v real with great load temps :confused:
i hope its the board i did not even push the chip yet:(
i looked around local to just get a cheap one but no 1 has them.
i dont think i can get a replacement from newegg bu i'll try.
sending both back hope they still have a few around to send me.
what a sickener cowie ehQuote:
Originally Posted by cowie
***********WARNING********
well good news bad news got another 2600k , installed and straight away the led boot codes started going c1,29 4d etc so i though thats good so it looks like cpu has expired so put original cpu back and just the boot loop again and the led code panel was blank just a red flickers so swapped cpu out and put the new one in and hey presto all the boot codes coming up again c1,29 etc, so plugged everything back in and all booted up fine so definitely the cpu had expired mine had good load temps but needed a lot of v core but temps were in line and i have run probably 100 3dmarks ok i needed about 1.685 volt on the old to be fully stable at 5373, my symptoms before it expired was i had a few bsod's with failiure to communicate with secondary processor and also just before it would not start again when you ran super pi it wouldn't run at all just came up not responding, as said i have been running 4200 all auto and 5373 for benchies but mine needed a lot of vcore right from the start to bench 5373 but as said i have done 100's of benchies at that vcore and speed all ok, also i had upped systam agent voltage the same day to 1.125 i think so not sure what the absolute cause was but v core rather than heat is my guess
oh so mb ok then?
sounds like the same thing i have glen poor chip
just a cycle of on off and flicker in code leds wtf...
i did not even give it alot of vvt .1.20 dram 1.66 sa was 1.035 vcoree real was 1.58 till last night i pushed 1.60 it did not boot so cmos and i was back to bios.
today ok for 1 hour then whiling fighting the flash to back up bios it went bad.
less then 20 hours benching oc'ed in a weeks time???? and i thuoght the 980's were weak.
well now i can buy another to have a back up.
Ed Marvell ports aren't THAT bad lol. I have been using them for so long for my C300.
Any idea why after a hour up & running the 2600K :confused:L3 Cache Read & Write lose speed?
Just keeps getting slower & slower...
I think i got the max i can out of my chip now.
http://i24.photobucket.com/albums/c1.../overclock.jpg
http://valid.canardpc.com/show_oc.php?id=1632365
And the SuperPI run,
http://i24.photobucket.com/albums/c1...757/1234-1.jpg
what are you using to test cache speed???Quote:
Originally Posted by TheHardCase
I definitely second that question.
L3 Read will drop to 40000MB/s & Write 30000MB/sQuote:
Originally Posted by sin0822
maybe is better to buy a sata port card or raid one, in that case what card can perform same as the intel ? to raid ?Quote:
Originally Posted by Lsdmeasap
AIDA64 eh, i wouldn't trust every program, but it would very well tests cache speed, although i highly doubt it does so. See if you run superpi if you see the same degradation.
Why there is only 2 LLC levels? Other boards has 5. Is it a hardware limitation for this board or something?
We shall see.... I'll join the Gigabyte P67A-UD7 party soon.Quote:
Originally Posted by sin0822
Bit of colour coordination :D
http://img.i4memory.info/reviewimage...7A-UD7_008.jpg
Yes the other boards with multiple levels have digital VRM systems, this has an analogue system. You have three levels, standard, level1, and level2, you have what 2 less levels than the ASUS boards?Quote:
Originally Posted by nijel2@comcast.
Either way Standard is meant to be for normal people or those who want a energy efficient system, level1 is supposed be for mild overclockers who want really low vdroop and vdrop, level2 is supposed to be for extreme ocers who want the most stability in voltage delivery. That is why level2 will never drop voltage under 10mv of what you set it at idle and will increase voltage during load. So what you set is what you get plus more, a lot more actually.
In the past there was no LLC, then we got smarter PWMs(analogue and digital) and now most of them use separate ICs for loadline equation and phases switching. Now we have LLC in multiple levels on the good boards(Gigabyte and ASUS mainly) not every digital vrm board has multiple levels of LLC.
For instance digital PWMs can do all this loadline stuff from inside the PWM is one of the benefits of a digital PWM, the VRM is under software control. With the Rapamge3 extreme and other x58 board with digital PWMs the max we saw was 3 levels of LLC 0%, 50%, and 100%. For X58 GB had almost all their X58A with standard, level1, and level2 in an analogue vrm, this is was pretty damn good because it countered all mainstream ASUS boards with a more complex PWM system. So they were on par in the top Rampage/Maximum VS. UD9/UD7(formula=UD5).
Now that Analogue and digital PWMs are just as precise this kind of ruins of one digital PWMs strong points of being more precise(precision comes from the PWM logic circuit in the case of PWMs software is much more complex than hardware thus is more able to be more precise in theory) because they usually always are(a few mv). Anyways with VRD12 Intel is trying to force everyone to use digital PWMs by the end of 2012, so that is why you see the shift toward a requirement that all PWMs allow a minimum of 5mv voltage increase as well as an extra SVID digital channel. This is pretty damn huge, not many analogue PWMs OR Digital PWMs have this ability, but companies that spearhead this industry(Intersil=analogue) (Chil=Digital) really hit the ball with it, Intersil got its VRD12 certification before Chil did which is impressive as well.
Intel also added in stuff like 4 phases server boards that boast 180amps which is pretty damn hard to do. This new spec demands that the VRM industry basically improve heavily on the analogue(or make it semi digital which it already is) or adopt even better digital. It is pretty impressive what Intel has done. The thing is that analogue and Digital PWMs aren't all that different, right now we have a large mix of analogue and digital signals and both types of PWMs have to use ADC(analogue to digital converter) or DAC(the opposite converter) to change all the signals to what they need. in the past they were almost all analogue in the future they will almost all be digital, its on Intel's time. Digital allows for much smaller VRMs though.
This is VRD12 PWM spec its a damn good read:
http://wenku.baidu.com/view/7e1bbcd8...6f533221c.html
thanks for the quick tutorial sin. that was a good read.
sooo good eva :up: im wainting for mine to arrive alsoQuote:
Originally Posted by eva2000
cheers
Sergio
nice george :D
click on link :shoot:
http://img703.imageshack.us/img703/4...7hoodsmall.png
lol Dinos! U like kahlifa too or just taking the piss ??
i run it before every 32M SuperPi wazza it goes faster :D
keeping it gangsta ay dinos lol
just linked a better video :D:D:D:
Hi folks!
I just assembled my new rig (shown in sig) and I wanted to flash my BIOS from F4 to something newer. Tried F7 in Qflash, invalid BIOS. Tried F7e in Qflash, invalid BIOS. Tried F7e with FlashSPI, invalid BIOS. Tried @BIOS, it crashed after reading the F7e file. Am I overlooking something obvious? How can I flash my BIOS?
Thanks!
use this app and do it in windows
http://download.gigabyte.us/FileList...gbt_atbios.exe
once you flash it make sure you also copy this new bios to your backup bios by hitting ALT+F12 during PC startup
Thanks for that Dinos.
Dunno wth just happened, but I connected a 2TB hard drive from my old PC to the new rig so I could copy over all my old data, rebooted, and...DISK BOOT FAILURE, INSERT SYSTEM DISK AND PRESS ENTER!
Both the C300s and my optical drives don't get detected by the "IDE search" before POST, but they are detected by the RAID utility and the array is fine. The pre-POST (!) search only detects my other two SATA drives connected to the SATA 2 port.
I've checked in the BIOS that XHD is on, RAID 0 is on, boot device priority is correct, just looks as though my OS has got fried somehow. Any ideas? :confused:
When you add/remove drives you always need to go into Advanced BIOS page and then into the first section "Hard Disk Boot Priority" and be sure you move your OS drive to the top of that list.
Same applies with clearing CMOS, loading Optimized, or flashing the BIOS.
Yeah I already did all that. Just reconnected my 3TB drive that I disconnected to connect my old PC's drive and windows booted just fine. Seems somehow that Windows has made my 3TB drive the active partition! How did that happen?!
Did you install windows again? If so whatever disk is set first will be the disk with windows boot files placed on it.
If you are just moving disks around windows wont set anything active without you manually doing it.
No, I didn't. But it has definitely set a 100MB NTFS partition as active on my (as yet) unformatted 3TB drive. Attaching a screen shot.
I suppose I just need to make my RAID 0 array the active partition now, do a start-up repair and things should be good?
That 100MB partition is only created during windows 7 install, if you do not create/manage your own partitions during the install.
I've never seen it move, or be moved, are you sure you did not just install windows 7? Or install windows 7 with the disk in question set as the first disk in the BIOS boot Priority, that's the only way it would be there.
I think you are confusing "Active" partition with "System" partition. What you need to do is first be sure your intended OS / C drive is set to the top of the BIOS Hard Disk Boot Priority, then you need to delete the hidden partition from your array. Yes that will cause windows to not boot, then run startup repair and it will put the boot files onto the C Drive not in a partition, you can manually rebuild the BCDStore as well (That's what startup repair will do)
Thanks LSD. Yeah, I am pretty sure I had my SSD array as first boot device when I installed Win 7. It was 3 a.m. though, maybe I changed it later! :D
Right, that might also explain my less than stellar Windows start-up times. I was wondering why it was taking as long as it was. Gonna do a startup repair and brb.
You guys are awesome! :up:
Edit: Urg...I'm now stuck in some sort of endless 57 Windows updates loop every time I start up and shut down the system. :-/ I think it's running outta space on that 100MB partition or something and not applying the updates and then retrying every time I start up / shut down! *roll eyes*
Haha, ya pretty sure isn't too convincing :)
I think you probably changed it later and just didn't realize it until now.
Here is how to manually rebuild the BCDStore in case you want to do it that way, second method
http://support.microsoft.com/kb/927391
Right, so system is back on it's feet and now, back to where I began here...
Is F7e the best BIOS to update to, for the moment? I am planning on overclocking my system quite substantially, but I've heard stories of new Sandy chips dying with more than 1.4V, so I'll probably be limiting myself to 1.35 - 1.40V vCore. This new machine is gonna have to last me a while, sure can't afford upgrades like this too often! :D
Ok, VERY dumb question here:
Ok, what is the benefit of increasing the TDP and TDC (watts and currents) in the Turbo Boost section of the BIOS, even if you have Turbo boost disabled (and thus the independent core multipliers and watts and amps are all greened out?)
Like, what benefit would setting them to something like 150 amps and 250 watts do?
Default is 93 and 95 or around there....
I tried setting it to 40 amps and 50 watts (I had to temporarily enable turbo boost to do this) and then re-disabled it, so the fields were greened out again, but the 40 and 50 still showed. I was wondering if it would even take effect if turbo boost were completely disabled.
Result: HELLO CLEAR CMOS :rofl: :ROTF: (computer kept freezing at the CPU string display in POST, either freezing completely, or turning itself off after 15 seconds while on that string and turning on again....not even the overclock failed feature would kick in...it kept showing 100x34 so it tried to reset the settings, but I guess my amp/tdp limits stopped the cpu from working?)
Pressed the clear cmos button, loaded my saved profile, and I was back to normal. But what happened anyway?
Does raising the Amps to 150 and watts to 250 cause any boost in performance when overclocking (example: 5 ghz and 1.4vcore?), if all poewr saving options are set to disabled?
BTW I was doing this on a UD5. UD7 should be the same, anyway...(so I posted here)
I answered your PM. 40 amps and 50 watts lol that is just not going to work man, even at 1600mhz. You could never load the CPU past 50% lol.
I have adjusted turbo mode many times and increase those values to 300 each, only to find it benefits those with UD4s and UD3s. I believe that GB automatically set the reported tdp and tdc ont eh UD7 and maybe UD54 to somthing very low like on P55 boards, so that you do not need to set these values yourself.
Makes no difference on my UD5 even if you set 1000/500. I need 150/150 give or take a few to run 5G @ 1.4v otherwise the CPU throttles the multi back through lack of juice.Quote:
Originally Posted by Falkentyne
Anybody else have USB3 problems with this board? I got a Sharkoon Quickport Duo USB3 dockingstation for sata HD's. And most of the time it does not get a connection with the computer. And when it does work, it's not stable. It disconnects from time to time.
If I plug it into USB2 I do not have the problem, but that's of course much slower.
It also works fine with via the USB3 ports on my GA-X58A-UD7 ver1 mobo.
Dinos is it possible to get skew parameters opened up in a bios rev?
The skew parameters that the M4E has for the BLCK? They are obviously optimized but It would be nice to have that though, maybe to see if I can push higher than 107.5. Either way blck is CPU dependent.
Basically yes. I want to try and tweak skews to try and stabilize certain bclk and multi combos more than for extra bclk. For example my 54x multi is really sketchy with most bclk and I need it for max mem speed.
well I went to mc to get another chip but it dont work still?
i knew it,i a week i kill the mb unless its a classy or dfi lol
on ne live chat now to make sure get replacement mb
*well looks like i'm crap out of luck ne said no replacements only refund
refund ain't bad spend the money else where you have 3 months to buy again. Personally id save the money. BTW i have never killed a piece of comptuer hardware just for no reason. I have physically damaged, voltage damaged, current damaged probably, corrupted, and just plain old DOA, but never killed a piece of HW with shock. I always use anti-static mat, gloves, and wrist band. I killed an X1150 and it really pissed me off. Its like people think static shock doesn't exist, it sure as hell does, everytime i sit down i touch a computer case and zap, every day every time i go to sit on my computer chair.
i never shocked one to death i have gloves i use when i build.
i may have put one or two componets in shock thuo......
was me all the way on this no dought. :D
i'll join you guys/gals again soon.;)
Guys, would you do a little easy test?
If your are using DVID for your 24/7 4C/8T stable clock, please run 1 thread LinX AVX for 10-15 min and see if you have stability problems (BSOD).
Can't stabilize low loads.
dvid? use svid.
How? Auto?
Works for high loads, but for low 10-20% loads Vcore for 4.7GHz = Vcore for 3.4GHz and that's where a got BSODs. Doesn't scale properly for x47 for all cores. It'll work probably for something like 41-43-45-47 for the cores.
default 3.4GHz > 10-20% load=1.262V > 100% load=1.296V
clocked 4.7GHz> 10-20% load=1.262V (BSOD) > 100% load=1.368V
Tried all LLC levels. Fixed Vcore with LLC L2 is the best option for now, but the idle Vcore is high.
clocked 4.7GHz> idle 1.344 > 10-20% load=1.356V (BSOD) > 100% load=1.368V
turn EIST and C1E on with auto voltage.
All is ON. I never run them OFF. It just doesn't scale. That's why I asked for more LLC levels earlier. Might help for better tuning Vcore for the entire load range. I can't believe no one else has this problem. I'm on my third chip now and the problem is worst with chips needing more Vcore for the same clock which is perfectly logical - wider Vcore range scales worst. I guess people don't spend much time at low loads yet.:)
idk mine scales just fine, maybe you are using it too much? you can't really expect 4.8ghz on auto vcore lol. It kinda hits a limit at 4.5ghz-4.8ghz for me. One of my chips does it fine at 4.8ghz the other wont at 4.5ghz. Of course you can just change the vcore but then the voltage doesn't scale.
I have posted this about 10 times now on different threads without a reply... I might be going in the wrong direction and just looking for some input.
IMO: P67/H67 was a stopgap with the intention of getting SB to retail faster.
With Z68 coming out right on the heels of the new production schedule for P67,
where is the market for these boards now?
With this kind of timetable, would you still want a P67 over a Z68? why?
For starters Z68 would be a costlier platform overall.Quote:
Originally Posted by Kingleo
Quote: The chipset, which is said to include integrated Intel HD graphics with support for both digital and analogue outputs, is designed to offer performance overclocking features for Sandy Bridge processors.Quote:
Originally Posted by Kingleo
The chipset will allow for 'performance over clocking of processor, memory, and processor integrated graphics,' with Intel clearly hoping to capture the performance enthusiast market with this revision.
Hope this helps...
I notice
my 2600K can bench @ 5.4GHz with 1866Mhz 8-8-8-24 @ 1.60V setting
<Kingston 2000C8 (elpida hyper)>
but It got BSOD @ 5.4GHz with 2133Mhz 9-11-9-28 @ 1.62V setting
<G.Skill 2133C9-11-9-28 @ 1.60V (PSC40nm)>
same setting
thanks..
just trying to figure out what to do when its time to replace this board. Thought z68 would make a good HTPC when 2011 is released.
TryQuote:
Originally Posted by PatRaceTin
CAS Latency Time 9
tRCD 11
tRP 9
tRAS 28
tWL from 7 to 8
All rest auto & std performance level @ 1.64v
If it works you can reduce vdimm & tighten things up.
I am still unable to update the BIOS on my P67A-UD7.
Qflash and FlashSPI say the BIOS file is invalid and @BIOS reads the BIOS file and then crashes. What do I do?
You tried this version of @BIOS?Quote:
Originally Posted by kbk_75
http://download.gigabyte.us/FileList...gbt_atbios.exe
Yeah, you are right. I'm just trying to 24/7 maximize the chip for my cooling setup. Wish we had more options to precisely control the voltage. I guess I'll settle down or try another board when the time comes. Unfortunately only UD7 has the features I need. They better tweak the LLC.Quote:
Originally Posted by sin0822
As said early, use @bios ver 2.1 and don't forget to remove the old and install the new one before flash ( or it will bring you back an error ) ( i have got the same crash with @bios before try to uninstall it and reinstall the good one ) ....Quote:
Originally Posted by kbk_75
Anyway. someone tried the last F7f beta bios ?
Ran into same problem http://i4memory.com/f90/gigabyte-p67...49/#post140444 - just a question guys @bios works fine in winxp right ?Quote:
Originally Posted by dinos22
http://img.i4memory.info/reviewimage...qflash_005.png
http://img.i4memory.info/reviewimage...qflash_007.png
Have yet to install windows xp heh
I use @bios in W7 and it is fine so i assume it will be ok for XP. The new one from GB I mean, the one on the disk doesnt work and q flash doesnt work. Bit of an oversight on that one I think.
yes eva it works on xp :up:Quote:
Originally Posted by eva2000
Should work, then you will be able to flash normally with Qflash ( it's really a F3 bios bug )Quote:
Originally Posted by eva2000
Just tried F7f bios, first funny thing is the Bclock who was set automatically at 99.8mhz with previous bios is now at 100.3mhz... the second thing is voltage drop seems lower ( without LLC ) and more stable ..
my setup is stable with my 24/7 settings for 5ghz, so i hope it overclock better or same of the F7f.
Thanks guys. I installed the newer version of @bios over the older one, which was a mistake. Uninstalled the previous version, installed the new version and managed to flash my bios to f7e, no problems at all. :)
No worry i have do the same the first time, and i was going crazy to search what happend.Quote:
Originally Posted by kbk_75
Guys, i have a doubt on how i set the voltage if it's the soft CPUZ or Hwinfo who are buggy or me who set something wrong, i have some strange reading from thoses 2 softs.
Vcore: 1.5V in bios ( just for test purpose ) CPU PLL overvoltage enabled ( reported as 1.488 on windows )
LLC: auto
Multi: 50x
Turbo: Disabled
All other setting in AUtO ( EIST; C1 states etc )
HWinfo when idle :
http://img15.imageshack.us/img15/8338/hwinfo.png
CPUz when idle before and after test it( 1.488Vcore, normal with 1.5V set in bios )
http://img714.imageshack.us/img714/2093/cpuzstop.png
HW info max load
http://img827.imageshack.us/img827/5520/lynxhwinfo.png
CPU-z Max load
http://img815.imageshack.us/img815/1213/cpu3tun.png
I believe this is just the software who are confused by the different voltage / different states and don't update it enough fast, so don't show the 1.488Vcore and they read not well the voltage. but i was ask me if the voltage set on max load is not using the "turbo one " or something like that.
@ Lanek
LLC Auto is giving you the same as standard, set it to level 1 to eliminate the droop although by the looks of it you have a nice chip that may only need 1.38v for 5GHz. The reason the voltage is not reduced at rest using the C states is because you have gone above vid. if you want it to reduce leave vid normal & + - voltage with the dynamic vid.
Lol nice point: i understand now why i had so much trouble for stabilise at 54x multi ... my vcore was still too low even when i was up it. You just have save me from some future headache.
I have test it again, but this time with real time Vcore read...
LynX idle = 1.48V
bench start and vdrop occurs when clock raise
LynX Load = 1.38V
then the first run is complete, clock decrease and vcore reincrease to 1.48V
Little question anyway... as i don't remember the last time i have play with DVID, it's standard Vcore +0.1 +0.2 etc ? sorry i absolutely don't remember how work this setting.
.
( in addition, yes this chip is 2h Lynx stable with thoses settings, and many hour on different games ( im never enough patient for make run Prime or Lynx for 4-5hours lol )
Got new chip today, fourth in a row, second c-batch, and looks like this one scales Vcore perfectly:
Don't think this one gonna BSOD at low loads, but we'll see. So far so good. Next step - optimizing Vcore without braking the good scaling.Code:Vcore = Auto, LLC = Level1
Clock 3.4GHz 4.7GHz
SVID 1.232V 1.356V
idle 0.984V 0.984V
13% load 1.200V 1.248V
26% 1.224V 1.284V
39% 1.236V 1.320V
51% 1.236V 1.344V
63-100% 1.236V 1.356V
Glad it's not the board giving me troubles.:up:
3dm05@ 5415, GTX460sli, lots of vcore :shocked:
http://img98.imageshack.us/img98/9426/18800407.png
Uploaded with ImageShack.usmhz, GTX460 sli, lots of vcore :shocked:
jeeeze Tom thats so good :shocked:Quote:
Originally Posted by OC Nub
congraaaaaaaaaaaaatsssss my friend :up:
Thank you.
had the board not wanting to change bclock at all this morning. I flashed the bios again and it works fine now. If you see the board acting up a bit try flashing the bios just in case its corrupted (especially if you push the board crazy like some of us do hehehe) :up:
yup i found this helped with memory clocks as well, for some reason my settings wouldn't stick so i flashed and it worked perfectly.Quote:
Originally Posted by dinos22
I believe my board do it by itself, each time i push to far, freeze in 3Dmark11, reboot i got a niiice " Checksum error, can't load HPA, ..." reflash with backup bios ( 3third times tonight in 1 hour, )... ( sometimes even without OC )
What i have find is flash using @bios, then reflash using Qflash with the same new bios ( F7e or f ) ... look like it stabilise a bit the thing about this " auto reflash " .
If someone have a fix for this will be good ..
Anyway someone can indicate to me how use Dynamic Vid ? set to normal and then ? how work this value, it's SVID +0.20 voltage or Stock +0.xx or need use - 0.xx ) Don't want to come with 2Vcore or try boot with 0.4V.