With F8x,I donn't have cold boot(F7e yes) and is much more stable(Main and Backup),but need more Vcore then F7e.
Printable View
With F8x,I donn't have cold boot(F7e yes) and is much more stable(Main and Backup),but need more Vcore then F7e.
It happened again:confused:...Bclk stuck @ 99.8. Probly killed ME in eeprom
ive had that happen too but bios flash will fix that up mate, just flash the bios with qflash to something different you used, i like this F8x seems alright
Yes I know,SP1 running here from November.
Will try this combo F7f(Main),F8x(Backup).
BIOS F7f:
http://img121.imageshack.us/img121/5...le1385vf7e.png
BIOS F8x:
http://img33.imageshack.us/img33/842...1stablef8x.png
heheh it is a bit at start but once you know what you're doing in terms of settings it should be doing the same thing every time, if not time to clear cmos or flash bios if it wont do certain settings like no bclock change, thats a giveaway the bios may be corrupt
I am using official F7 as backup bios and F8x as main bios and things are very good now even OC fail recovery is much better than it used to be. I used Qflash within bios to flash the F8x
So you guys using SP1, have you noticed any performance increases in benchmarks?
yes yes! thats what i get. But i first need to go in to the bios to set my overclock again to get past the 4E error. sometimes it works on the first go and sometimes i need to do that 3 times for it to work. Thats the strange thing.
AAA oke! i thought it was the same thing and that PLL voltage setting wuld overide the overvotage setting. So i wil disable it, because i think its on auto now.
Yes ive read it a couple of times, but thats not the problem.
And thats the strange thing about it, because when i get the 4E error it won't even boot @ 1.5 where as before i needed 1.35.
Its just if as that 4E bug occurs it holds everything back. Until it is releast through a bios flash or setting it first to default and than reentering the OC again.
i'm using F8x
im testing since first releases and worked as a charm i feel system faster or maybe its placebo effect dunno but i decided to go serious on testings and put 1 SSD with NO SP1 and one with SP1 RTM and benches no difference, in some cases scores more NON SP1
about the GFlops dif on LINX i really dont know if its a flaw or its the new memory management that SP1 brings
But 4 nothches more of vcore dont justify going to it inmo
cheers :up:
Sergio
F7e for me seems like more stable, faster in high multies bench...and with less Vcore too.
Take off all usb peripherals when oc from bios...it will help to avoid endless post code 50 hangs:)
4E is an overclock failure on CPU ratio most likely since CPU outputs maximum loading at this stage. I am speaking to bios team about a few different things and we are trying to get some concrete and replicable issues but at the moment nothing consistently shows a fail. There should be a timer for 4E error to reset the bios and let you load up defaults and then go back into bios and change settings but if it isnt working for you let me you may have to clear cmos until it's addressed a bit better.
if anyone seriously wants something sorted you need to provide system spec in a post and procedure how to get the error to show up every time. The only other suggestion i have if you do encounter certain errors is to flash the main bios again just to be certain the bios isnt suffering from some corruption or whatnot
I think f8X recovers from failed oc alot better and may be just a tad more effec, but what it gains in effec it loses in about 20mhz stable for benching over f7e.
These are just my thoughts though, when you find a better chip give it a shot, may only pertain to higher frequency clocks.
@ Dinos22
mostly it is reproducible by resetting system wile in windows, but youre right i will look further into it an see if i can reproduce it on camera.
the idea is to have specific set of steps to send to bios guys so that they can replicate it and then try address that specific issue, we're on the same page :)
The 4E will recover on its own, it just has a little longer recovery period than on previous boards. But if I get 4E I just hit reset, the board powers off, hit power button, it will 4E again and power off, I hit the power button again, then it goes past 4e code to 50 then 70 reboot and then will go to 75 and allow you back to bios, not a big deal. Im more than mildy ocd so I cant help but to watch the post codes everytime it boots.
Only a couple times has it failed to recover and do the endless reboot at C1, trying something stupid with memory caused this. With f7e and f8X I havent had it happen at all.
Setting the multi higher than the chip will boot at causes 4E, also going above your max bclk.
dinos22, please, could you talk to the bios guys to unhide [Any Kay] option in Power Management Setup > Power On By Keyboard. I really like it for easy PC_ON, no need to reach the power button. I have to mess with _EN_CODE.bin for every bios coming out to show it and make it selectable. I do this for the last 3-4 years with every gigabyte board I had. I can't get it why they need to hide it. It always worked for me, it's not like giving problems or something. Doesn't make any sense to hide so simple but useful feature from us.
yeah thats true
C1 is memory related you are right, the IMC just freaks out at that point probably going homie what the fluck are you thinking, OC fail :D
Thanks for that.
Yes, mouse click works, but every small unintentional mouse movement wakes it up also and that's why I use the keyboard and not the mouse.
Weird.
4E for me is multi chip can not post at, etc you have a 55x multi chip or to much bclk and tried to post above 55X
C1 for me = Andre Yangs FF that or instant reboot loops with no code at all.
It means i can safely return the chip and be sure it won't post for the store and test another :D
I havent had the instant reboots with no post code but had C1 endless bootloop a couple times, had to clear cmos.
I have tried a few SB chips,this one makes it past 4E at 55x multi but best Ive been able to do is make it to wondows load screen and hangs. Tried lowering bclk but same thing happens, voltage doesnt seem to help.
Ok, you guys seem to get 4E as well. One of my procs stops at 4E on 52x100. The other at 55x100. :(
Thanks, I will have to upgrade my window unit AC :D. Best I can manage is around 25c at 5400mhz. I am getting hooked on this HWBot ranking and would love to try some better cooling but Ive never tried to insulate a board, this is all pretty new to me. May try a water chiller and see how it does.
Once you found a >56X chip, just make sure decided what you want to do with it.
Either you going to bench it or put it on 24/7 safe rig, because time's ticking when it started to get >1.54Vcore...degradation will start...no matter what cooling you slap on it
Well this chip has been running well over 1.65vcore for a couple weeks now every night for a few hours benching/playing, up to 1.32vtt, and up to 1.80vdimm. Im waiting for it to die but it just keeps going. I got these 2500k's to play with, not to worried about the longevity. It has been fun to just say F-it, give it some juice and have a blast. Im sure it will come to an abrupt end one night benching but for now Ill keep having fun.
hmmm i am a bit confused ... same proc, same ram , same vc, same psu, same hdd .... Ud5 spi 32m run @ 6m 35.xxsec while on my ud7 its 6m 37.xx sec ...same bios settings and same ram sub timings ... how come ud5 is giving me faster runs? I wanted the ud7 to be my main board ...do i need to set anything else ? or this board is really slower than ud5 ?
If an ant walked on top of my case, it could raise my superPi times by that amount. The same goes for the gravitational pull of the moon. :rofl:
Two seconds on a 32m run is within margin of error. I've had 1M runs finish at 9.385, then other runs at 9.406.
Even two EXACT same boards at exact same settings might not give the exact same spi time.
BTW voltages (error correction) can play a part as well, if you're borderline stable.
well yeah you are right :)) I also noticed that i need to put 1.44V LLC level 2 to stabilize @ 5ghz ( latest bios with PLL enabled ) while 1.42V LLC level 1 in ud5 ( f5x ) .... hehehe we are actually running a local game in a local forum here and i am hoping that ud7 can do the trick as one of my friend is using that and getting better run :D
damn that is a nice tiny boost.
great board mates will run my Redlines on 3x2GB like lsdmeasap told me will do
thanks mate :up:
http://img130.imageshack.us/img130/9...hyperx8x32.png
http://img34.imageshack.us/img34/1415/p1080478t.jpg
Uploaded with ImageShack.us
Wow alessan, nice clean build you have there!
I wish I had cable management that clean...
New AIDA64 1.60.1300 "memory copy" bechmark is less arround 2000 MB/s,as Fiery explain:
http://img27.imageshack.us/img27/6006/1236ph.pngQuote:
That's normal. With the previous AIDA64 release Sandy Bridge processors could somehow get around our Memory Copy benchmark, and over-optimize it. It resulted in an unrealistically high score. We've fixed it by increasing the used memory block size to avoid any sorts of CPU optimization impacts. So the realistic Memory Copy scores are produced by AIDA64 v1.60, and please ignore the scores obtained by the previous versions.
http://img402.imageshack.us/img402/9411/1300.png
Sergio,
I saw 3 memory stick on your UD7.:confused:
Edit:
Ohh,just saw your notepad.
How 3 sticks doing,any boost in performance?
as if any of the AIDA scores are that realistic, what am I supposed to compared memory bandwidth mb/s to anyway?. SB is optimized so what. They shouldn't be un-optimizing it.
thanks stasio amazing lsdmeasap had the courtesy of testing 3x2 on giga P67A so i bought those, will need a bump on vdimm to do 2133 but not worth it im more than fine with 6GB on a dual channel mobo and see its very fast
i will lower to 6-9-6...to see if they take it
cheers :up:
Sergio
very happy passed linx and very nice bandwith :up:
i could lower 2 vcore notches but not vtt
http://img155.imageshack.us/img155/7...4linxx5ok1.jpg
http://img109.imageshack.us/img109/5...4linxx5ok2.png
http://img37.imageshack.us/img37/3903/capturake.png
im using CE1/EIST states
room temp 25C
Nice stable run.:up:
Will be helpfull to include ET6(Voltage) in your pic and adjust your pic to forum post.:rofl:
VTT.....V?
vtt=qpi/vtt
Random bsod...often
Any clue?
Thanks
http://img406.imageshack.us/img406/9...nshot001cw.png
More than likely, your Vcore is too low at idle or light load.
Are you using LLC?
If yes, try raising vcore by 0.020v, or raising VTT by 0.05v.
BTW I still don't understand how DVID works. How are you supposed to use DVID if you are NOT using speedstep/EIST/C1/C3, etc....?
And with rant mode on,
What happened to the days when you would simply get Roundoff errors or illegal sumouts and stuff in Prime95, instead of it either running forever, or blue screening? Is this some sort of error trap in the CPU that triggers an exception, to avoid possible corrupt data, or something? I never saw blue screens on my QX9650 system, unless it was a driver problem or something, only Prime or LinX just crashing.
Now on this SB, the only way I can make prime throw out errors is by running my RAM out of specs..it seems like CPU errors just usually cause a BSOD...(though I did have Rybka 4 (Chessbase) just "stop" calculating once, due to too little vcore...)
What OS are you using.. I remember an ASUS post about removing a driver to prevent randon BSOD, maybe its something to do with this.
My rig will LinX 5G @ 1.356v CPU-Z but needs 1.392v to be proper 24/7 stable & prevent BSOD. I find surfing the net & playing games for hours is a better stability test than LinX!!!
dumo here are the codes for BSOD that one is VCORE
BSOD Codes
0x124 = add/remove vcore or QPI/VTT voltage (usually Vcore, once it was QPI/VTT)
0x101 = add more vcore
0x50 = RAM timings/Frequency add DDR3 voltage or add QPI/VTT
0x1E = add more vcore
0x3B = add more vcore
0xD1 = add QPI/VTT voltage
“0x9C = QPI/VTT most likely, but increasing vcore has helped in some instances”
0X109 = add DDR3 voltage
0x0A = add QPI/VTT voltage
cheers :up:
Sergio
stasio i tried F8x and reboots the same...also i think GFlops are lower, how is F7f and where can i dld it ?
Thanks !
Sergio
Heloo,
i will ask right here
i tried on my ud4 3x 2gb of ram and performance was 2x slower then dual chanel 2x 2gb same ram
So i wanted to try out an ethernet card PCIE, card installed, did another reboot to disable both onboard nics,disable in bios then reboot and hangs at 4E every time. tried 5 or 6 reboots in a row, all hang 4E, then power down, clear cmos. power on, load my settings again
disable onboard controllers again, and then it boots into windows fine, this board is so maddening.
Sergio,I have both(Main and Backup) on F8x,btw
F7f
Guys just got myself this board and a 2600K. Which is teh best bios for clocking a 2600K and G Skill 2 X2gb 1600mhz 6-8-6-24 1.5v RAM? I have tried to read the thread to find this out but after reading afew pages thought I would just ask!!!
yes
man reboot loops are terrible when are they going to fix that ? i tried all bios and same, if i lock or fail in Win for some reason takes like 6 times to enter bios and sometimes more...i took out the 3x2GB as tough it was stable makes cold boot, on the other day i turn on PC and starts to loop
Picked up a GTX470 to play with. The card isnt a great clocker but scores ok in 05 and 06.
http://img833.imageshack.us/img833/4791/95647239.jpg
Uploaded with ImageShack.us
I am experiencing some issue with f8x ... it want s more voltage :( my vcore on f7f only @ 1.42 for 5ghz while f8x wanted 1.44 for me to stablize it :( also observed that f8x want LLC= level 2 vs level 1 on f7f
Gskill 2200 died:(...Testing with an old stand bye 2GB GTR with bios F7f
@Dumo ...how long do you have that gskill? I am actually thinking if which one to get, the 2200 Gskill or Adata .... if the gskill will just die in a few weeks/months ...then probably i will just gor for the Adata ...and BTW ...what dram voltage did you apply ? Am asking since my gskill 2133 perfect storm has been running for a few days already with 1.78 dram voltage ( for me to go 6-9-6-18 @ 2133 Mhz ) and would like to know if I will expect the PS to die soon :D
bart, mostly @ 1.66V 7-9-7-25. I'm waiting for RipjawsX 2133 8-9-8-24 from newegg
http://www.newegg.com/Product/Produc...82E16820231450
@Dumo ...wow died at 1.66v only ... too bad ... i just hope that my PS will not die that soon considering I am putting 1.78v to make it run 6-9-6-18 VS the 9-9-9-24 stock timings hehe
@Falkentyne ...yep ...am now at 1.42 again with LLC=level 1 @ 5ghz speed ( 50 x 100 )
I have a new kit of gskill 2133's that cause lots of random bsod's. I put the mushkin redline 1600 6 8 6 kit in and run them at 2133 7 10 7 timings and get no random bsod. The last 2 kits I have tried from gskill have been like this, my 1600 pi's are flaky on x58 platform. Last good kit from gskill I tried were the 9600 pi's, I will stick with mushkin for now.
yea i hope they don't pull an OCZ, i mean how bad can memory QC get that two kits i buy back to back different ratings are both bad?
hmmm seems like i am lucky ..well maybe so far ... I have the PI black 1600 and been with my x58 for quite some time and no issue,I actually bought 2 its and the other kit is actually I am using for sandy ( alternate for my PS )and the black can go 18XX speed and can finish spi 32M for several times as well.
I had thought the chip was starting to degrade, but once I switched out the memory the bsod stopped. The gskill's will still pass stress tests but normally on a cold boot or even sometimes a restart will give a bsod as soon as I get to windows. Tested with 2 kits of mushkin, one kit of redlines and a kit of radioactive using the same voltages and settings and not one random bsod. That was a wasted 159.00 for the 2x2gb kit and another 170.00 for the 1600 3x2gb kit, this wont get them anymore business from me.
i´am using this RipshawX kit currently with the asus pro and until now no further problems no matter which speed i choose. hope they will have a long life :D
http://www.memoryc.com/products/desc...nel/index.html
@OC Nub .. good for you you that its the ram and not the proc ...I think have pretty good choices of HW's available in your area ...here in PH we usually see gskill as the main ram available in the market. there are few dominators but they are very pricey here. Kingston are just the ordinary ( value ram with 1333 speed LOL ) and it will be a bit hard just to get the HyperX. On my side, I need to either wait for my relatives from the US to be me one and bring it here or get the available gskill rams. Luckily the gskill thats with me right now are ( actually came from new egg and not locally since price is cheaper in new egg and a relative came last christmas season) still alive with no issue.
So who tried F7f bios for me are most the same for some reason F7e seems stable, but reboots looops i hate man
u guys use S1 or S3 state ?
I find F7e to be better than F7f and F8x. At least for memory clocking.
@OC_Nub: great run there.... very nice indeed.
Some 3D testss with F7f/F7e. For me both are better clocking on cpu and ram compatibility (I can clocked my D9 GTR) compared to F8x.
http://img163.imageshack.us/img163/8...enshot022k.png
F7f
http://img87.imageshack.us/img87/8186/a1signature.png
dumo that is a nice chip, me jealous.
Awesome chip, and great score with the GTX560!
Nice work :up:
I was trying to order the Gigabyte GTX560Ti super oc'ed card but they have been out of stock. Just ordered a gigabyte GTX470 super oc'ed hoping it will clock beter than this EVGA. Also hope I can sli them lol, I wasnt thinking and pulled the trigger to fast.