Printable View
Quote:
Originally Posted by lowdog
In case anyone wants to turn on ACHI after installing under IDE mode. Enable the RAID/ACHI with a reg switch. Swap to the white sockets (different controller), boot, it'll ask for the Intel drivers (below). Swap back to the IC10R sockets and that was it. Pita but it works.
http://img227.imageshack.us/img227/3...92008qixl3.png
OMG it failed installing windows and rebooted the came up with a post screen saying Intel ACHI bios not installed...WTF!!!!!! and my raid array was corrupted....it just hung there so I rebooted and now it just hangs on the Gigabyte splash/boot screen :rofl: with error code 75 which means Detect and install all IDE devices: HDD, LS120, ZIP, CRROM.
To be honest I'm totally flabergasted, can't even enter bios now.
reset to defaults
use basic setup
no drives plugged in
check if you can get into bios
try again with non OCd setup first
that is really odd
what bios are you on
i am going to test Raid controller with some iRAM drives soon and see if mine works
both lan ports on the motherboard just stopped working :( yes it's the motherboard and not the cable. I tested the cable with my raid controller and works fine.
Quote:
Originally Posted by dinos22
Shut the board down and unpluged the power cord, hit the clear cmos button and left it for 20 minutes, it powered back up but had reverted back to bios F2 which is what it must have shipped with, wierd because this is now the third time I am having to flash back to the F4m bios.
F4m is what I originally flashed to because most said it was just as good as F4j but just had the coloured overvolt voltages in bios.
Man this board and it's bios are flakey as.....and I haven't even had a chance to try any overclock, all the problems I have been having have been all stock - 0ptimised defaults!!!!
Mine as well as others are rock solid. You seem to have every issue known to man happen with that board. So much that I'm calling it user error. Pull your battery for at least 30 seconds while depressing the power button for at least 15 seconds, pressing the cmos reset button is useless.Quote:
Originally Posted by lowdog
You have P6T, why not just use that?
Quote:
Originally Posted by ReverendMaynard
Excuse me but I think the fact the board can't even boot straight back up after powering off entails alittle more than just the blanket statement of "user error"....I'll throw that one back at you next time you have a faulty piece of hardware :p:
Don't worry I know how to reset cmos and I have been removing the battery and draining the board of power on various occasions. Like I have been mentioning I can get the board to reboot but only after jumping through the necessary hoops that are required to do so with this particular board....not just the simple "push the power button" that should be the case.
At least my false ego isn't so huge that I can't admit where I have made mistakes as in the "not enabeling the usb keyboard function" in bios, but in other areas I know fairly well what I'm doing and enough to ascertain if a component has some issues.
regarding the P6T....I will use it but atm I'm not giving up on the giga until I know without doubt it's faulty and that is becoming more of the case the more I play with it.
You have no issues with your board/system and I'm truly glad for you regarding that but please don't just assume based on your problem free encounter with the board that others may not have such a smooth encounter and that that may be due to a faulty board, which does happen
hey man try to relax your mount on the board
if there is too much flex it could go crazy
Np but atm all I am using is the stock Intel heatsink (thats why I won't try clocking it yet), I have the Swiftech backplate for the GTZ but am still waiting for the block to arrive in stock early Jan.Quote:
Originally Posted by dinos22
For some reason when I was installing windows before on the raid array and it crashed it changed back to the F2 bios automatically, obviously installing on the raid array had somehow corrupted the bios?????/ beyond me.
And don't worry my mem is fine, already run multyple passes of memtest86 so thats not the issue. PSU is fine as is cpu, just mobo/bios thats seems dodgy.
I'm trying to like this board but so far no good, I'll judge Giga after I get a replacement board and see how that one pans out. The big retailer where I purchased from in Aus said they have sold heaps of the UD5 with no returns and only a few of the Extremes and already had 3 returns of the Extreme....mine will be the 4th so far.
Horay horah finally sorted it out :rolleyes:
Incompatabillity with Thermalright Toughpower 1200W psu.
Solution - hold down the power button for 7 seconds (the time it takes for the power to drain out of the mobo which is indicated by the DDR PHASE LED remaining lit or the onboard power switch remaining lit) then press the power switch imediately after those leds/power switch lights go out and it powers on.
My speculated conclusion is something to do with the 2 phase power for DDR ram implementation has an issue with this particular PSU.
So yeah, user error after all :rofl: or perhaps more like user persistance to figure out the problem. ;)
glad ya got it sorted out.. time for a new psu i'd say hehQuote:
Originally Posted by lowdog
I did have hard time with AHCI, as I did install Vista in IDE mode.
I did clear that with turning some swithches in registry but RAID mode was still no go.
So I decided to reinstall Vista X64 in RAID mode on single Intel 80GB SSD.
I noticed that F4 final is back at the GB website for download and just downloaded it. Is it still a no-no to install this BIOS? I am currently on F4r......
Many thanks.
I got the lan ports working again.... rather than pull the system apart I hit the cmos clear button like 50 (because hitting it once was not enough) time with the power supply unplugged.
Hmm, my LAN1 port died few days ago also.
Perhaps is time to rise ICH volts? It's good to start meeting all the possible side effects of our OC.Quote:
I got the lan ports working again.... rather than pull the system apart I hit the cmos clear button like 50 (because hitting it once was not enough) time with the power supply unplugged.
lan chip is not on the ICH.... i miss the broadcom lanchip that my old 965P had... i dont like the realcrap stuff.Quote:
Originally Posted by vega22
:ROTF::ROTF::ROTF:Quote:
Originally Posted by dinos22
:D
BTT, what bios should I use right now? Still on F4J for my builds because that one worked well.. but there are some issues with VMware fixed in F4M right? So F4M it is? Some votes please ;)
I wait for F5 final, im good with F4j for now.
i am still on F4f (edit-sry this is the right name heh)
i tried F4R but saw no improvements
actually RAM kinda clocked worse which is odd as its supposed to improve RAM compatibility with corsair QVL loaded in it and all
Quote:
Originally Posted by vega22
my ICH volts were at 1.26
I miss the marvell network chips. I agree about the realcrap stuff.Quote:
Originally Posted by Hornet331
I'm using bios F4r and I have reason to believe the 2:12 memory multi is not working right at stock. more on that when linpack finishes.
F4J doesn't post for me with some memory dividers btw.. 6x is working, 8x usually isn't, 10x is working again etc - weird.
Edit: Also what is the deal with the white SATA ports? I connected a few SSDs to them and they weren't even recognized.. Is that some kind of jmicron dedicated (crappy) controller? Works fine on the blue ports (ICH10R ftw)
has anybody tried the new F4 on GB's web site? how does F5a work?
Of course, but the ICH is the master of I/O peripherals. Anyway 1.26v is enough.Quote:
lan chip is not on the ICH
yeah I know it worked fine on my evga board.Quote:
Originally Posted by vega22
maybe a dumb question, but did you install the drivers for the gigabyte controller/enable it in the bios?Quote:
Originally Posted by jcool
Im in F4j. I can achieve without problems 19x211 @ 1.40v Qpi - 1.38v Vcc Bios/linpack x64 stable. However for 20x200 i need to raise the Vcore over 1.40v. Perhaps 20x200 be easier in F3, but I dont dare to flashing back for testingQuote:
F4J doesn't post for me with some memory dividers btw
the 2:8 ratio works fine on bios F4r I was trying to using 2:12 at stock to run 1600 on the ram but it randomly errors out within 30 minutes. the 2:8 ratio is solid though just check the pic.
I'll test F5a and F4 final later. I'm just happy my system is running ok at stock atm so I can get back to OC. I'm glad I use true image to keep a backup os install running.
I am on F4m, but using mem channel timings and round trip latency values from F4j, as my mem likes them better.
Only other differences on F4m, pink and red volt warnings, F4m bios loads smooth and fast instead of slow and choppy like f4j and earlier bioses, and F4m allows direct enter on CAS instead of scrolling. But I am prime stable at 4.2, 21x200 1.376vcore load in either F4j or F4m.
Using F4 final or F5a, I can not even boot past 178 bclk. I spent hours, tried manually locking every visible setting to F4j/F4m values, tried pink and then redlining every voltage including channels, altering clock/pcie drive, if it could be changed in bios, I tried it. Bclk above 178 on F4 or F5a is no go on mine. Computer stays on, does not give bios screen, does not load windows, blank monitor, mouse/keyboard no green light, only way out is clear cmos....which I used upwards of 100x.
Very good idea. The best of F4j and F4m together...Quote:
I am on F4m, but using mem channel timings and round trip latency values from F4j, as my mem likes them better.
Well i tried F5a, and will be flashing F4m as soon as i can find a link for it.
I can boot no problem, and my 4.2ghz settings (200 blck) are at least 10 mins prime stable, however i stopped it because on this bios i cannot do any form of restart - whether it be restarting from windows, or even doing a ctrl+alt+del during post, it just gives me a black screen with no post codes. I have to press the reset button on my board. My NB temps were also higher than usual (59c idle, 61c load - noticed because of the flashing green LED), and i figured out why. I had my IOH Core voltage on Auto, which was fine before but according to Easytune with this bios it's setting 1.22v when it's 1.1v stock.
What's the difference between IOH Core, ICH I/O, and ICH core? I assume ICH core is the SB, so which is NB out of ICH I/O and IOH Core? If IOH Core is the NB which it seems to be, then it's a very different chip indeed because at 1.1v idle it has the same temps as my old x48 NB did at 100% load with 1.7v!
P.s anyone got a link to F4m? Searching back through this thread but can't find one. Station drivers don't have it :(
NVM found one :up:
http://www.fileden.com/files/2008/7/...EX58EX-F4m.zip
Yep, actually I should have mentioned, sometimes, not always when i first boot with bclk >180 on F4 or F5a, I will get into windows first time. But when restart 100% time or sometimes just first time from bios, I get blank screen, computer running, no load, no bios screen...then I had clear cmos to get out.Quote:
Originally Posted by Xello
Yeah tell me about it.
F4m seems better so far, in fact i just ran everest and i'm approaching the 20k mb/s read barrier, that's with 1600mhz and 9-9-9-24, completely stock (3.8ghz uclk). Faster than 1600mhz 9-8-8-18 was on F4j :)
Haven't tried lowering cas or overclocking yet, will verify 4.2ghz stability first.
P.s superpi also got a major speed boost, .4s faster than identicle settings on F4j and went under 10s with 4.2ghz which i never managed before :)
http://img.photobucket.com/albums/v4...star/12121.jpg
Now to see if it's stable..
best bios for ex58 extreme is 4j or 4m?Quote:
Originally Posted by Xello
I'm still calling F4J the golden one.
F4j is good, but I'm not having any issues with F5a thusfar.Quote:
Originally Posted by ReverendMaynard
I'll probably sell the board before I flash it from F4J. I refuse to beta test for Gigabyte if they aren't going to pay me.Quote:
Originally Posted by Lukee
Can't say so far, only been testing F4m for an hour (stable so far). If stability is the same, which i suspect it will be, then F4m is faster than F4j for me on everest and superpi - and therefore probably benchmarks in general.
Hmm, this experience has been far from bad for me, though i know 1 or 2 have had fatal problems. They're pumping out bios pretty fast, i have faith that in the next month or two they will get one out that will resolve everything and have the best bits from F4j and F4m included :) For the people with fatal problems, well it's just like me and the trouble i had with my old 4870x2's - sell and move on (change manufacturer)Quote:
Originally Posted by ReverendMaynard
P.s had no idea flashing bios would wipe my overclock profiles though! How infuriating... lucky i have a good memory. Miss my blackops more ever day :p
I sort of of miss my DFI X48 TR2SB+....sort of. CMOS Reloaded FTMFW.Quote:
Originally Posted by Xello
you think its better on other boards?Quote:
Originally Posted by ReverendMaynard
look in the asus pt6/rampage II thread. :p:
has anybody had any problems with the board not detecting 6gb of memory?
i've just picked up a UD5, but i'm having problems with the DDR3_1 slot. the board refuses to post with just a single stick in there (post code 6B).
with 3 modules installed, it sees 4gb in the bios (or 2gb with 3x1).
CPU-Z sees all 6gb, but says it's running in dual channel mode and windows sees 4gb
I've tried resetting the bios a few times and tried a couple of different versions but nothing has worked so far :(
i found a post on tweaktown with a similar prob (here), but the solution doesn't really work for me as i cannot boot the system with only 1 stick in the DDR3_1 slot :shrug:
damnit, i think it's gonna have to go back, but now the shop have no more in stock :(
Yeah that is a board fault, maybe a bent pin somewhere or something, bios has always reported correct memory for me.
Sorry to hear that :down:
Seems you got a dead ramslot, its bad luck that its the first one.
gutted :(
now i dunno what board to buy :shrug:
People come from all over to report problems and ask for help with a bad board. Its not really reflective of the general reliability of the boards you're looking at buying. Better to check several forums and see if there's any particular recurring problem that might indicate avoiding the board. Its not like anyone makes a point of coming in to a thread and saying my board's great it work perfectly. So I'll say hey my boards great and it works perfectly....so far.
-----------------------------------------
Tried a few games on it today, just to compair 4870x2 same driver gpu speeds etc.
QX9650 @ 4Ghz
http://img248.imageshack.us/img248/8315/40782783nn4.jpg
940 @ 4GHz. Gained a bit.
http://img201.imageshack.us/img201/314/dx10vu1.jpg
a bit is good. :p:
13 fps avarage is nothing to sneeze at, also min fps is 15fps higher :yepp:
If that F4 final really available somewhere?
If I flash to F5a (or F4) and hit load optimal settings, reboot, put in my OC settings 21x200, it will load windows 1X. Any further reboots and I get blank screen, no bios entry, and had been clearing cmos. Xellos suggestion works for me as well, I can shutdown, hit reset button, than boot up windows fine back into my OC settings. But each time I restart, I have to power down and hit reset button.Quote:
Originally Posted by Lukee
Are you not having to do this? If so are u using vista or xp, just wondering if something obvious, or one of those things that is not worth tracking down and just wait til GB releases better bios....I am just curious at this point.
Here ExtremeQuote:
Originally Posted by BulldogPO
http://www.gigabyte.com.tw/Support/M...ProductID=2957
And Here UD5
http://www.gigabyte.com.tw/Support/M...ProductID=2958
I seem to have the cold boot issue between 150 and 200 as well.. had a rock stable setting with 170 BCLK and 12gb ram (53h prime lol as well as 8h linpack), however on a coldboot it won't start. I have to start with 150 BCLK first, then go to 170 and it works. Strange..
well "good to know" (although its bad :p: ).Quote:
Originally Posted by jcool
So if there are more people with that problem and that board and with otherboards the problem dont occure at least i know where the problem lies.
Right now im not certain if it is the board or the cpu.
Well, let me help there. So far, I've had 2 EX58 Extreme setups, first with a 940, 2nd with a 920. Both show the same behaviour when running 12GB...
However, the 920 can run at 200BCLK even with 12GB of ram, while the 940 couldn't. Now running 200*19 on the 920, mem at 1600.
Vcore 1,25V load, Vdimm 1,6V, VTT 1,45V (will try to lower that one later on). Passed linpack for 3 hours now so I'm guessing its stable.
The other mobo + 940 wouldn't even post with the exact same settings (same mem, same bios). So I am guessing the CPUs aren't created equal (yet).
it still puzzels me why there is a hole of 50mhz... i would understand if the cpu wont cold boot at 200mhz bclck... reminds me to the FSB holes in the first 965p chipsets/C2D. :D
You must have not tried clocking a 680i lol. Swiss cheese lol.Quote:
Originally Posted by Hornet331
i knew why i never owned a board with corruptoforce chipset. ;)Quote:
Originally Posted by ReverendMaynard
corruptoforce...now that is a new one. Fitting indeed lol.Quote:
Originally Posted by Hornet331
I'm testing bios f4k at 3.7 and the new linpack based linX 5.4.1 is putting the screws to it big time.
edit: yes that is 918MB out of 12GB free.
What is key to get 4GHz speed for Uncore?
Now I can run tests at 20x200Mhz fine but Uncore is only 3600MHz.
I just figured this out earlier and it took me allot of work to get 4ghz uncoreQuote:
Originally Posted by BulldogPO
I just want to verify that was it, cause I messed with so many settings I am really not sure what I did lol I will play with it tomorrow some more. Got my 06 score over 30k at 4.5 and thanks to vince's advice I think I can I get to 31k hopefully :D
Trying to get my g.skill 1600mhz pi black run 2000mhz but it wont post. Running with 200x18 qpi x36 uncore 16-17. Multi 10
Have someone manage to boot with 10 multi ? What uncore should i choose ? And volt ?
what bios are you using? I know f4r anything above 2:8 leads to corruption.Quote:
Originally Posted by ante_ante
Bench2die, can you post your bios settings?
I´m using f4r .Quote:
Originally Posted by safan80
dram voltage is key for uncore :yepp:Quote:
Originally Posted by BulldogPO
You do realize that the new i7 "capable" linpack routines from intel cheat and use only half the CPU when HT is enabled right? Actually, temps with prime (blend mode) are currently 5-10C higher. Improve your cooling m8.Quote:
Originally Posted by safan80
It's fine to test memspeed/BCLK tho (IBT freezes after 20 seconds if the IMC can't handle it hrhr)
69C on the nb? Insane.Quote:
Originally Posted by safan80
The new version of Linx 5.4.1, with one gui runs two versions of linpack on core i7 with hyperthreading, so temps are a few C higher than prime and load is 99.9% on all 8 threads.
http://www.xtremesystems.org/forums/...&postcount=192
A bit interesting, but don't know the mobo type....havn't tried on my ex58-ud5 yet.....
http://www.itocp.com/attachments/mon...oVDal2iTvV.jpg
http://www.itocp.com/attachments/mon...wkmuOAzoIP.jpg
It says that:
Intel C-STATE Tech→Enabled
C State package Limit setting→C7
It seems for asus only :(
its no cheat, intels current linpack version (10.1) was only desigened for up to 4 cores. ;)Quote:
Originally Posted by jcool
Give them time to update the code to support more threads. In meanwhile just run two instances. :yepp:
I have killed my board. Well, is not dead, but is seriously sick.
I was 1000+ linpack rock solid in 19x211/x36/x16/x8 with 1.38750 Vcc, 1.40 Qpi and 1.28 IOH. So i wanted to tighten the Uncore to x17, leaving the Vdimm in 1.66v. Then i ran Linpack, the system was frozen in 14 mins, and after reboot i could see one of my 750Gb HD completely dead, and the board no longer accepts more than just a little OC. Clear CMOS, battery off, flashing to another Bios..nothing helps.
Too much stressing, i think...
sounds to me like the QPI and IOH voltage too high.....Quote:
Originally Posted by vega22
so it's better not enter the pink area....for long-term usage
You're positive that the HDD is dead? Or does it show up as failed in the Intel Matrix boot screen? That has happened to me twice pushing the uncore to x17 but a cmos reset brought it back and has been fine ever since. Your problem looks like a trashed cpu, not motherboard.Quote:
Originally Posted by vega22
1000+ linpack really necessary or were you trying to burn the cpu?
How many linx runs do you guys do? I've switched to using it today after getting a freeze in prime after 19~ hours, the first time i've ever had instability after so long :( So if this is going to start happening regularly i'd rather use a quicker test.
I hear linx will tell instability quicker so i'll try it, right now i'm on 21/50 runs. Temps are about 6c higher than blend. Is 'residual (norm)' the value to watch? It has to be the same all the time or else you're unstable right?
I used it after a restart so the 'max available memory' was fluctuating around the 5.1gb mark, i got it on 5020mb with 'all' memory to use.
Seems simple so far, if this can do a good test in even 4-5 hours i will use it from now on :)
Xello - I would think shifting from prime to linx (linpack?) would be a major change for everyone.
Post a link if you have one - I need to do some homework ...
If it can tell us stable or no - sooner than p95, I'm all for it!
I dont get all this 24 hours linpack and prime stable BS, If I can run blend 3-4 hours and not freeze while gaming or anything else its stable enough for me. If I wanted to stress my cpu that hard I would take it outside and bash it with a sledgehammer for a while.
Is this for e-peen? (I do understand if you're folding but come on)
I got from here (link @ bottom of first post)
http://www.xtremesystems.org/forums/...d.php?t=201670
I think this latest version supports HT, but i'm not even OC'ing with HT enabled yet so i'll worry about that later.
Currently on 35/50 runs, 2 hours in. If this is a 100% way to prove stability then it's much quicker than prime, but i'm not sure if it is 100% with 50 runs. You can choose up to 1000 or more i think :shrug:
I always done 24hr with prime, but with these set numbered runs i'm not sure what to choose.
Lol yeah i'm gonna stick "100 passes linpack" in my sig when finished :rolleyes:Quote:
Originally Posted by Bobbylite
It's for peace of mind so that when i'm in the middle of a heated frag session with my crosshair on someone's head and about to get a sweet kill my computer doesn't go oh, sorry mate, i'd rather show you this blue screen instead.
If you can get that peace of mind with less testing, go ahead what does anyone else care :up:
40 hours of COD4, CODW@W, 2 TWL 5v5 matches and 2 pickup scrims and I run at the settings in my sig...not even a remote little stutter. :)Quote:
Originally Posted by Xello
RvrndMynrd - CODW@W.
EDIT: I just noticed that you use a Lycosa keyboard....bought one last night and I hate it...but my G15 was completely worn out.
I love linx because it tells me in half day, what prime used to tell me in days. I plotted out all my OC points, 4, 4.1, 4.2, 4.3 with and without LLC enabled in half a day using linx 10 runs near max mem (and either second linx running with small mem same time or prime running to get 100% load). Using linx all my OC points are usually within 0, 1, or sometimes 2 notches of prime stable 12+hrs. Though new version runs 100% 2 instances by itself, using half mem.
I will always run prime 12 hrs runs, as I never crash when prime stable, so just works for me. But linx is speed king to getting preliminaries done quickly. Cant imagine having to use prime for quick runs anymore.
That's cool, did you ever try F4m? I got very measurable speed boosts, going from F4j.Quote:
Originally Posted by ReverendMaynard
I just dropped from 1.4v (what i needed for 4.2ghz stable on f4j with prime, 1 notch less was unstable) to 1.37v with linpack, and i'm currently on 15 passes. I read on another forum that 5 passes on linpack is enough to verify a small tweak, and 30-40 passes should be done for final settings. So either linpack is useless and i'm actually unstable, or this new bios has done the trick :cool:
Yeah i think i will use it this way too, have you ever been linx stable and failed prime blend?Quote:
Originally Posted by rge
to each his own :up: :yepp:Quote:
Originally Posted by Xello
Depends, running 10 runs usually nets me within 0, 1, or 2 notches. Running 100 runs+, I havent yet, but problem is I use them both as I overclock, and almost always run linx beside prime for first hour, trying to crash prime early and reduce number of 8 hr prime reboots (and wasting time). So not the same as only using 1 at a time getting that stable then trying the other, that I cant completely attest to.Quote:
Originally Posted by Xello
Yes, the hdd is dead. I tested this on another machine and neither works. No, I have not tested for pleasure 1000+ linpacks. Is the number that results from adding three-four hours to find the least stable Vcc, plus three-four hours to find the least stable IOH, etc. Linpack is very quick to find errors in Prime95 stability, but not as fast adjusting stability based on himself.Quote:
You're positive that the HDD is dead? Or does it show up as failed in the Intel Matrix boot screen? That has happened to me twice pushing the uncore to x17 but a cmos reset brought it back and has been fine ever since. Your problem looks like a trashed cpu, not motherboard.
1000+ linpack really necessary or were you trying to burn the cpu?
Yes, this worries me ... But a hard disk broken i wants to think that means problems at the board. Perhaps everything is broken ...Quote:
Your problem looks like a trashed cpu, not motherboard.
The HDD was most likely a fluke. I've never seen a board kill an HDD unless there was a ground issue aka lightening strike.Quote:
Originally Posted by vega22
Tightening up your uncore didn't kill your board/cpu/drive. Linpack put the nails in the coffin. I had one of my techs tell me that he wants to use linpack as a procedural test for replacement cpus. I told him to save the $ for every cpu that he runs it on so when it comes back failed, he can pay for it.
We are betatesters of a very new technology. We do not know what can do every thing to another.Quote:
Originally Posted by ReverendMaynard
For now, and if anything, i try not to put the uncore @ x17/3.600 in my new board when full load.
No we aren't.Quote:
Originally Posted by vega22
How to properly flash the bios for Gigabyte???
Is it just copying the ex58ud5.f4r file on a USB drive and go into Bios and ask it to flash the BIOS with that file ?
Is there anything else to do, like loading the optimized default values to finish of loading the BIOS?
Thanks.
That is all you have to do :)Quote:
Originally Posted by icecold316
Sorry but you are entirely wrong...Quote:
Originally Posted by Hornet331
a) the latest Linpack release officially supports Core i7 architecture
b) It only won't run 8 threads on an i7 rig depite the CPU being recognized and reported as capable of running 8 threads, guess why ;)
As I am one of those WCG maniacs who own multiple Dual Harpertown systems (real Octacore systems) I can tell you linpack is very well capable of running 8 threads / on 8 cores - Intel just didn't want it to with i7 because all the "normal" people using stock coolers would probably degrade or kill their CPUs :rolleyes:
If you use linpack on an i7, disable HT beforehand. It gives higher Gflop rating, faster runtimes and stresses the CPU more. Still, prime blend with HT on is currently the best option to stress i7.
Hey guys, just so you don't think I'm crazy or a noobie or whatever:
my replacement Extreme board's SATA ports are going bad just like the ones on my first Extreme board. I accepted that S3 resume problem. I figured how to work around the slow boot problem. But having my SATA drives not recognized on boot just doesn't fly. I called Giga Tech Support and they, again, were of no help. They even suggested that I get another board from another manufacturer.
I called MWave and, luckily, they are willing to take back BOTH defective boards without any restocking fees. They also told me that they've stopped selling the Extreme model - and if you check their website, you'll see that it has been pulled from their inventory. The service rep said that they've "discontinued" this model - though I means that they (MWave) stopped selling it because I still see it in stock and on sale at other places. Take that for what it's worth.
As for me, I'm hunting down an EVGA or Asus R2E board now.
Showing out of stock at NewEgg. Interesting.
never had a problem here :yepp:Quote:
Originally Posted by yenniedn
Out of stock at NewEgg but still on their website - which means they still stock it or at least plan to.Quote:
Originally Posted by MicroFly
MWave has removed the model from their website altogether. That's a whole different matter than just temporarily out of stock.
FYI:
http://www.mwave.com/mwave/DeepSearc...+x58&ALL=y&TP=
Problem resolved.
New PSU TT 1500W Toughpower and the board is powering up without issue. Finally resolved and I can use the thing and move on.
Thanks fornowagain for suggestion to try a new PSU (something I overlooked because the TT 1200W Toughpower had been working fine on other systems), you were the only one to suggest a solution of relevance. :)
maybe, i just noticed that they updated linpack for Ci7.Quote:
Originally Posted by jcool
But linpack still stresses the cpu much more then prime. :yepp:
30min prime blend:
http://www.abload.de/thumb/30minprime4j7t.jpg
30min linpack 2 instances:
http://www.abload.de/thumb/30minlinpackvo74.jpg
system was preheated by runnning 9 houres of boinc. :)
Boinc (WCG/Seti) loads the cpu even less then prime blend. :yepp:
http://www.abload.de/thumb/boincywiv.jpg
Glad you got there in the end mate.Quote:
Originally Posted by lowdog
Ok, didn't know you could run 2 instances - your point ;)Quote:
Originally Posted by Hornet331
Still it kind of sucks the stupid routine won't just run 8 threads in 1 app :shakes:
Uhm.. have a look at my boinc/WCG stats and you'll know why that info was superfluous :wasntme:Quote:
Boinc (WCG/Seti) loads the cpu even less then prime blend. :yepp:
http://www.abload.de/thumb/boincywiv.jpg
yeah i had been getting sick of prime it would only test 32% ( when ETQW would use 40-60% depending on if I used a ramdrive) of my memory and linpack via LinX tests 93% of my memory.you should adjust the memory setting in LinX you can make it use 90+% of your memory.Quote:
Originally Posted by rge
What settings you prefer?:
CPU clock drive -800mV
PCI-E clock drive -900mV
or
CPU clock drive -700mV
PCI-E clock drive -700mV
The default in all the F4s is 700/700, but load optimal or safe defaults after flashing turns this into 800/900. I always flash->load optimal->clear cmos, so I always use 700/700, but i have not compared between the two settings stability and performance.
800/900 is default, so i would leave it at 800/900.
Dualist added a HT option in linx (gui for linpack), it start 2 instances and you dont need to open linx 2 times. :)Quote:
Originally Posted by jcool