wohoo, got my Q6600 G0 today! batch #L739A466. But first need to take the last tests on this dualcore e2160 ;)
Tonight Iīll see if I was right with the vdroop, I think itīs going to be huge with this quad
Printable View
wohoo, got my Q6600 G0 today! batch #L739A466. But first need to take the last tests on this dualcore e2160 ;)
Tonight Iīll see if I was right with the vdroop, I think itīs going to be huge with this quad
I've seen all these DS3 mods around but none is catered for DS3L. Shall I presume it is the same for my DS3L too? After all, the PCB and design is the same except no extra features like the rest.
Bump :DQuote:
Originally Posted by maxwolfie
There ara new bioses.
But the F10 also throws errors under memtest86+ like F9 does :(
mobo: p35 ds4 rev2.0
mem: 2 x 1Gb A-data Extreme Edition DDR2 1066+
F8 works fine.
Latest BIOS on Gigabyte.us website for DS3P 2.0 lists F8 as latest BIOS revision adding two items.
Fixing 1TB hard drives and adding loadline CPU control (I'm guessing vDroop control)
Does F8 BIOS for DS3P 2.0 cause memory errors? Are these real errors or only phantom errors in memtest86
Thank you! Seems like ASUS built in BIOS flasher! Easy!Quote:
Originally Posted by DeNs
From my testing it seems that certain ones are just false positives. The offending addresses have been outlined in the last 10 pages or so and they're easy to spot if they come up.Quote:
Originally Posted by kenton
I've had the errors come up then tried Orthos with exact same settings and wasn't an issue - ran for a couple of hours no issues before I shut it down.
---dens
The solid caps being the only dif between the S3 and DS3 (I think), the PCB's should be identical so I'd give the vDroop mod a go and see how it fares. Good luck :)Quote:
Originally Posted by maxwolfie
---dens
On Gigabyte.US website updated page for DS3P including Revision 2.1 preliminary information
New revision? I JUST order 2.0 yesterday.
Anyone know if this 2.1 is DS3* wide or just P revision?
And here is why you shouldn't mess with vdroop "fixes".
http://www.xtremesystems.org/forums/...0&postcount=32
"Fixing" it with a mod is going to put more stress on the CPU and the voltage regulation circuitry. Likely to cook your mobo or CPU in the end.
Super complex explanation:
http://www.thetechrepository.com/showthread.php?t=126
Quote:
For example, the addition of extra capacitors, the favorite VR circuit modification of the over-zealous albeit misguided basement-hobbyest, often does more harm than good. Understanding why there are those that attempt to make these changes isn't hard to do considering all the recent (actually, rather historic) negative publicity surrounding the dreaded "Vdroop problem" that seems to be prevalent in today's motherboards. We're here to tell you that you're much better of learning why Vdroop exists (and why it's needed), rather than attempting to hack up your expensive, well-engineered motherboard in a horrific effort to "correct" the problem.
My PCB revision is 2.1 - is that what it's in reference to?Quote:
Originally Posted by kenton
swaaye: Thanks for that info. Pretty much finished tinkering with my chip under phase so i'll likely take off the vDroop mod now that it's becoming more obvious that it'll do more harm than good in the long run. Fun while it lasted :) Would be ideal if it actually managed to keep the same clocks even without the vDroop mod. We'll see.
---dens
EP35-DS4, Revision 2.1....newer P35-DS4.
It's dated for Dec 14th...a bit after we notified them of the memtest bug....hopefully an F11 is in the works to fix the bug....thanks for the update!Quote:
Originally Posted by tlac
I got one of these last time I ordered from NewEgg, the board has leds on it that flash different colors when rebooting. I have not seen where it is any better than the DS4 2.0 rev. myself....Quote:
Originally Posted by Richard Dower
Well, on Gigabyte.US (and I assume .tw and worldwide!) website it lists GA-P35-DS3P revisions 1.0, 1.1 and 2.0.Quote:
Originally Posted by DeNs
This was the way it was until yesterday.
Yesterday while downloading driver for preparation of system building I notice website had add revision 2.1.
2.1 was not listed when I made my order with Newegg two days ago! It probably doesn't mean much, but darn I just order revision 2.0 with Newegg.
And yes, I assume 1.x and 2.x refer to PCB revision number.
Thanks you for your input!!
P35-DQ6Quote:
Originally Posted by Richard Dower
Receive from Gigabyte bios F7C and F7dd.
F7C-added in M.I.T-Loadline Calibration(Enable;Disable)
F7dd-
1.Fix 400-600mA CPU Fan regulation
2.Fix 1TG HDDs
3.Special Bios for Intel@QST
Anybody interested to try PM to me.
Btw,I using F7C
Edit link for download:
HERE
http://www.gigabyte.us/Products/Moth...2506,2579,2633
Per that webpage
Difference between DS3P 2.0 and 2.1 is in Gigabit Ethernet Controller
Realtek 8111 for 2.0
Realtek 8111B for 2.1
Anyone know what feature B adds? Perhaps cost issue for Gigabyte??
I am confused .. lets examine the different DS3P boards for example...
Untill now, we had the rev 1.0 , rev 1.1 (different heatsink) and rev2.0 (different heatsink and IO connectors).
But WHAT are the differences between these new boards: EP35-DS3P Rev2.1 (newest bios = F2) and P35-DS3P Rev 2.1 (newest bios F8, with different changelogs then F8 for rev1.0/1.1/2.0 !)
They both have the "Dynamic Energy Saver" feature, both support "1600FSB" and "ATI CrossFireX™ with dual PCI Express x16 interfaces"
But these boards have a different bios version, so they are different :confused:
Notice any difference with the loadline calibration?Quote:
Originally Posted by stasio
About -"LOADLINE CALIBRATION(enable/disable):
The Loadline Calibration helps users enjoy the motherboard's ultimate OC capabilities while ensuring stable and opimal CPU voltage during heavy or low loadings.
Myself keep enable.To early to consider some improvement.Need time with testing
Where do you get F7c and F7dd?
F7C (release 15/11/07)from Taiwan,F7dd (release 23/11/07)from GermanyQuote:
Originally Posted by MuffinFlavored
I find the vcore enough stable on the P35DQ6 without this loadline calibration :Quote:
Originally Posted by stasio
http://img503.imageshack.us/img503/8...coredk9.th.png
It's nice.Soon I will post mine.Quote:
Originally Posted by Repliquant67
As I receive a lot of PM according Bios F7C & F7dd for P35-DQ6 and my internet provider working today with 25%,I decided to upload.
Link is>>>>>HERE
Edit my Vcore(Bios F7C):
http://img118.imageshack.us/img118/9...coreen6.th.png
3.4GHz,Bios 1.40000v.
OCCT 30min 1.39000v(idle 1.39000v)
Thanks Stasio, the F7C Bios got my 3.8Ghz overclock OCCT stable. The Vdroop was the only thing letting it down and the loadline calibration option cured it (or at least reduced it).
Guys, I going to get an E8400 from my local shop in 12 hours from now. But I still not sure whether will my GA-P35-DS3L (Rev 2.0) can support this 45nm processor. The change log of the F6 bios only stated "Support Intel 45nm Yorkfield CPU". Need some opinion on this.. Can support or not? Thanks
I'm willing to bet it will work okay. In the worst case the mobo will probably report it as an "unknown" processor while posting.Quote:
Originally Posted by irangan
well under load you usually should have no further vdroop. :up: (or else your psu is about to die. :D )Quote:
Originally Posted by Repliquant67
vdroop only happens when going form idel to load.
Hmmmmm... Ok... I will give a try then.. Will update when I got the processor.Quote:
Originally Posted by GazC
Thanks
what is the BEST bios to use for the DS3p v.2.0 ?
this is my first gigabyte ever, so i dont know much about them !!
Also whats that bios hack or something to show more options??
At least you've got someone to blame if it doesn't work! I'd use your original chip flash the bios to the latest revision before you put the new chip in though, just to be safe.Quote:
Originally Posted by irangan
LOL... I already dropped an email to local gigabyte distro. According to them, its will works. We'll see.. Haha..Quote:
Originally Posted by GazC
Already did that.. Using F7 now..
Thanks ya.
The key is CTRL+F1 in the BIOS root menu. This opens more options.Quote:
Originally Posted by lastdon00
I, too, would like to know the best BIOS for the DS3P 2.0 as there have been some discrepancies in other boards about RAM killing BIOSes.
I have one maximus now, because the F7b was not really good on my P35DQ6 with my QX9650 (can't boot over 4Ghz and trouble with vcore in manual).Quote:
Originally Posted by stasio
If someone test one orki with this new bios, please give the results ;)
My second run on my newly bought Retail E8400.
4ghz at 1.37vcore (I believe can go lower than this). Using Gigabyte P35-DS3L. Stable at least while posting this. Haha.. Temp: 33'C Full load on WC.
http://img235.imageshack.us/img235/5...oto0410op3.jpg
http://img235.imageshack.us/img235/2...ghzds3lpi5.jpg
Nice!Quote:
Originally Posted by irangan
Thanks... At least, its run on my board... But its took several on / off before its managed to POST. I think my board is re setting itself for this proc.. Oh ya.. Using F7 bios and avoid pairing DS3L with E8400. I cant get my proc above 4.2ghz even with 1.6vcore when I can prime 4.0ghz at 1.37vcore (still on going)Quote:
Originally Posted by GazC
Not higer than my qx9650 for the same vcore.
Temp are nicer and could alow higger oc.
For the 24/7 use, I don't believe that I could go over 4.2-4.3Ghz in water cooling.
mm got my q6600 running this weekend....really nice boost almost everywhere :up:
I see vdroop, but it didnīt have effect (yet) to overclock, because I oc'ed just a bit, 333*9=3ghz, that boosted a bit too, ie world in conflict benchmark lowest fps rose up from 13 to 22 :p: (avg 34->41)
When I have time, I tweak this more, then weīll see if the vroodp mod is essential. Lapping coming ahead too. btw, with this mobo and this cpu, is the coretemp 0.96 reliable? @ extremely cold house (my friendīs place) coretemp showed startup temps like 15/12/12/14 (celcius) and in load the temps went up to 28. Possible with cooooold air & tt bigtyp? :eek:
so there are some issues with ds3l and these chips then?Quote:
Originally Posted by irangan
what about other p35 boards?
Hey all im preety new to overclocking as such i have noticed that all the q6600 around here are running 3.6g and up i cant get mine past 3.4 can anyone help me with a some volts to set before i oc my specs are this:
Q6600 @ 3.4 1.47vc swifttech watercooling kit up graded the rad to a swiftech360 and a alphacool res i have 2gig of ocz 8500 reaper 2x320 seagates and a x1950xtx vid card got 2 3870 arriving next week on and thats all running a p35-dq6 board in a cosmos case the crrent volts im using on everything is ..
Vcore: 1.47 Dram:2.3 fsbv:.1 mchv:.05 can anyone please help me with what volts i should try please... i need help:confused:
I'm running with the F10 Bios since yesterday. I always update to the latest bios.
My settings are Core CPU = 1.35V, memory voltage + 0.2 (for 2.0V total).
No luck in overclocking..it's a project in progress!
I did too, and too bad for me :( , i canīt go now over 299Mhz fsb, before my E6400 was working at 425x8 (ddr2@850Mhz CL4-4-4-10) without any problem (crysis, Boinc, pes2008...)..Quote:
Originally Posted by chrispycrunch
But since i have updated the bios its impossible for me to get even close to my "old" clocks, i donīt know whats going on and i've started to feel desperate with this *** motherboard.
Am i the only one with this problem? :shrug:
The "enhance performance" memory option actually starts to do something with newer BIOSes. So make sure you set it to standard if you are having problems.
It changes performance level. With standard it's 9, turbo 8, and extreme 7. I always leave it at extreme.
swaaye i always check that, i will try this afternoon...
Finally figured it out. After I added the .5v I set the main one back to Auto, this must somehow override the manual adjustment's I made. So I know have 2.32v with F7 and no false positives after memtesting overnight like I did with F8 on the DS3P.Quote:
Originally Posted by necron66
Finally i managed to do it and its again running at 3.4Ghz... i'am happy again and waiting for the E8400 to come out XD
You trust auto-performance tuning? extreme? What does 9,8,7 mean?Quote:
Originally Posted by [XC] 4X4N
Quote:
Originally Posted by HitMaker
It's out - what nice numbers, and sweet overclocks. For me saving 5 seconds encoding a song, or having nice benchmarks would not justify the $200. Quad core would though :)
And when will flash HDD's come out? That bottlenecks the system no matter how fast the CPU!
It really affects RAM access latency big time. Everest's latency test will show you. It is separate from the other RAM timings.Quote:
Originally Posted by chrispycrunch
About to install F9a for DS3P.
It doesn't auto-tune as you say. All it is, is gigabytes way of setting performance level. As far as I've seen, it does not make any other changes. That is why I always have it at extreme. That's performance level 7. I have tried to lower it to level 6 with memset, but it hard locks every time.Quote:
Originally Posted by chrispycrunch
Do you guys have to hit CTRL+F1 every time to access the settings in the BIOS or only once?
Also, I haven't had an Intel system in some time, and the NB cooler is feeling pretty toasty to me. I am at 370 right now, but am a bit scared to make the push to 400 and hit 3.6 on this cooler. Having a fan helps but I don't have a convenient method since I removed my side fan hole to install a window in my CM690.
I have a Whisper XE somewhere that I need to dig up, I'm sure that would do the trick.
How about cooling the mosfets around the socket? I'm using ramsinks right now... sufficient?
Ramsinks will work fine. You have to hit CRTL+F1 everytime in BIOS, although a new BIOS might skip the need for CRTL+F1. The F7b BIOS I'm running on my DS3P shows timings in BIOS without hitting CRTL+F1.
I had a Thermaltake Extreme Spirit II on the NB which worked fine with a Q6600 running 450fsb with +2v to MCH.
damn i need a 1:1 divider for my Dq6 seems it doesnt like running 4x1GB over 500ish mhz really hurts my cpu oc = /
f11 works without this error :yepp:Quote:
Originally Posted by miahallen
new rev 2.1 out for p35/c-ds3r that adds a power saver featrure
p35-ds3r rev 2.1
http://www.gigabyte.com.tw/Support/M...ProductID=2746
p35c-ds3r rev 2.1
http://www.gigabyte.com.tw/Products/...roductID=2749#
f10 bios for p35c-ds3r all revisions out
f11 bios for p35-ds3r all revisions out
I recommend that DQ6 owners remove their heat-pipe system and apply some Arctic Ceramique. I did this about two weeks ago, and after removing the heat-pipe system I saw that the thermal pads for the mosfet sinks were making no contact with the thermal tape at all (the thermal tape was smooth). So I added some ceramique, but I also saw how someone in the Abit P35 thread bolted their heatsinks down, and I did the same for the mosfet sinks on the DQ6. So I hope what I did has helped... I think it did... Over the summer the lowest load Vcore load I could run 3.6GHz was 1.35-1.36V, and now 1.32-1.33V... I just wish Gigabyte had some NB and PWM temp sensors.
As well as doing this, I also recommend you remove the useless crazy cool back plate which simply doesn't contact the motherboard. In addition, put some plastic washers between the retaining nuts and the back of the motherboard to slightly increase the clamping force of the NB sink, that lowered my NB temps quite a bit.Quote:
Originally Posted by NJDevilsFan21
Yeah, P35-DS3R's F9 BIOS shows everything without Ctrl-F1. Also shows real voltages instead of "Ok" or whatever it was before.Quote:
Originally Posted by [XC] DragonOrta
New BIOS's-
P35-DS4 - F11
P35-DS3P - F9
P35-DS3R - F11
P35C-DS3R - F10
P35-DS3 - F12
Wanted to report my experiences with the DS3P
Setup was very easy. Booted first try. System is in signature.
No problem with 4x1GB either, absolutely no problems.
Currently Q6600 is at 400x8 Quad Prime for a little over 22 hours. I will stop it at 24 hours and call it a day. I have a g0 chip with a pack date of 12/18/2007. My VID is a 1.325v which is the highest you can get from the factory.
400x7 = 2800 was stable for 12 hours Quad Prime @ 1.30v before I stopped it.
Went immediately to 400x8 and I decided to try the new OCCT program and it failed so I bumped voltage to 1.325v and OCCT passed at 30 minutes but I'm a 24 hour kinda guy because this is my HTPC.
Anyhow Quad Prime failed (rebooted system) after 6 hours. I was at the computer when it happened. I was scrolling through the Prime logs and it just rebooted.
Added voltage (forget BIOS setting, vDroop load is 1.34v, Idle vCore is 1.36v) and it's been stable for 22 hours now.
Increaded 3DMark06 score from ~10500 (E6400@3.4GHz/8800GTS 640MB @ 700/2000) to ~14500 with no overclock on new G92 8800GTS 512MB yet.
I am impressed in upgrade over ASUS P5B Deluxe. MIT tweaker menu is stupid and poorly designed but the correct options are there and overclocking is as they say great success
I wish I could go higher but I ordered the wrong Zalman cooler (universal instead of intel only) and the "AMD Style" clip was blocked by Gigabyte heatpipe on one side and heatsink on other. I didn't wish to mount Zalman pointing up anyways.
I am using ACF7Pro w/ MX2 paste (one dot in center of CPU size of rice)
Temps are (load/100% fans) 67/65/65/67 and idle (75% CPU fan/60% Zalman 120MM exhaust and 7v modded Zalman intake and 7v modded 92mm Scythe side intake) are in mid 40s. I forget but mid 40s. Computer is silent with the exception of the Raptor and it may need to go.
I have this problem that has occured yesterday, basically Windows locked up due to a Media Player Classic bug.
Pressing the reset button did nothing, and 1 press of the power button shut off my PC, it then went through a power cycle phase where it powered up with fans spinning for a few seconds then shuts off before POST. I let it power cycle a few times until I was sure it couldn't POST without some help.
I turned off the PSU for about 20 seconds, then powered up again.
However the BIOS automatically reset to defaults, so I put everything back to normal including my OC (Q6600 B3, 9x360, 1.39V, RAM was 1:1 yes underclocked), saved my settings, computer powers off and automatically back on again like usual when saving.
But POST just says 266x9
I get into Windows, everything I set was right like 1:1 RAM and CPU volts, but FSB was default 266.
I tried F7/F8/F9/F10 BIOS still same issue, sometimes it would power cycle about 5 times and boot up with 166MHz FSB instead, every setting will stick just not FSB.
Tried the new F11 BIOS, I put 340x9, saved, but still read as default 266MHz (at least it didn't power cycle so many times before it POST)
I had this P35-DS4 Rev 2.0 for 2 weeks now shipped with F7 BIOS no problems at 360x9, Prime95 stable and Folding ever since.
This Q6600 B3 is several months old only and was running happily in a P965-DS4 360x9 before too.
I will test more tomorrow with F11, but anyone have any ideas wtf is the problem? Cheers
F11 for the DS3R fixed the bug for the memory in Memtest. Everything is smooth as silk so far. :up:Quote:
Originally Posted by Cob
I have problems with it at extreme, usually. Turbo works until I'm around 400 FSB. It's a strange thing to test because I've had Prime95 run for hours fine and other times error out in a few mins. So, I definitely suggest setting it to standard until you find that the rest of the settings are stable.Quote:
Originally Posted by [XC] 4X4N
One thing I'll say is that on the P35-DS3R this setting makes much more of a difference to latency than on 965P-DS3. I believe the P35 boards go lower on the performance level when set to Turbo and Extreme. Compared to Turbo on 965P-DS3.
It's pretty cool when you can beat out Athlon 64s and their integrated memory controllers with this performance level. But as I said it's tricky to figure out if you're stable sometimes.
http://kenton.org/3200q6600.JPG
How is that?
It's a g0 revision with the worst VID 1.325v.
I can't go any higher, I fear, due to temps. My personal limit for these new 100C tJunct procs is 70C (for 85C procs I did 65C) and I'm at 68-70C on Core 0 and 3, and I'm at 65-66 on Core 1 and 2.
So
I bought a
P35-DS4 REV. 2.1
anyway it came preinstalled with BIOS F8.
THOUGH weird thing when i go tot he giga website for revisino 2.1, it says newest and first bios is f10
go figure that one out
Ok, no idea what I did, but looks like F11 may have fixed my OC troubles..., I set my OC to 6x400 (Q6600 2400MHz), but it posted as 6x266, went back into BIOS immediately and saw MIT listed it as 6x400.
Didn't change anything, just exited BIOS without saving and then it posted as 6x400 :D
booted into Windows and it indeed is 6x400, amazing, first time i've been able to increase FSB in the past 2 days with any available BIOS version.
went back into BIOS and put 8x400, saved, didn't power cycle and instantly POSTed.
Finished Prime95, so looks like F11 fixed whatever problem I had :D :D :clap:
my rev 2.0 shipped with F7, only F8 and above are listed, but I flashed with an F7 i found on the internet :DQuote:
Originally Posted by lastdon00
Q6600 SLACR G0 2.4 @ 3.5 * Thermalright Ultra Extreme 1x 12" Scythe flex
* GA-P35-DS3R * 2x1gb Cellshock DDR2 pc6400 @ 1180/5-5-5-12 *
Thermaltake Toughpower 750-W * Inno 3d 8800 GtX @ 660/2000 * SyncMaster 226W
I,m back to bios v f5.
Anyone overcloking 4 gb on this mobo?
Bios version?
2 x 1024 can overclock very well but 4 x 1024 wont.
I don't know what to do! I get random crashes.
Should I send 2 sticks back?
I just wanted to give back my experiences to this forum to confirm that F11 is stable, with 30 minutes worth of memtest and orthos testing (I'll do a complete test later).
F10 resulted in memtest errors.
My current settings are 365x8 = 2.92 @ stock. My next goal, pending stability tests, is to either go 5-5-5-15, 400x8, or 4-4-4-15 @ 365x8. I think a good temperature goal @ stock would be 35C idle and 60C load.
RAM brand: ocz
RAM rated speed: 6400
RAM actual speed: 870MHz
Latency: 5-5-5-15 @ 1.9V
CPU Clock Ratio (Note)____________ [8] <<<----CPU Multiplier
CPU Host Clock Control_ [Enabled]
CPU Host Frequency (MHz)__________ [365] <<<----FSB Speed (Front Side Buss)
PCI Express Frequency (Mhz)_______ [100]
C.I.A. 2__________________________ [Disabled]
System Memory Multiplier (SPD)____ [AUTO]
Memory Frequency (Mhz) 876
DRAM Timing Selectable_______ AUTO
CAS Latency Time_____________ [5] []
Dram RAS# to CAS# Delay______ [5 []
Dram RAS# Precharge Delay_____[5] []
Precharge Delay (tRAS)________[15] []
ACT to ACT Delay (tRRD)_______4 _____[auto]
Rank Write to READ Delay______3 _____[auto]
Write to Precharge Delay______6 _____[auto]
Refresh to ACT Delay________42 ______[0]<-
Read to Precharge Delay_______4 _____[auto]
Memory Performance Enhance__________ [TURBO]
High Speed DRAM DLL Settings________ [Option 1]
System Voltage Control____ [Manual]
DDR2 OverVoltage Control__ [+0.1V]
PCI-E OverVoltage Control_ [+0.1V]
FSB OverVoltage Control___ [+0.1V]
(G)MCH OverVoltage Control [+0.1V]
CPU Voltage Control_______ [Standard] - 1.35
I did not memtest F8 bios for DS3P 2.0 but I just loaded F9 and ran memtest with it. It completed 1 full pass without error
D9GMH @ 800MHz, 4-4-4-12 @ 2.06v, 4x1GB 4096MB
Quote:
Originally Posted by miahallen
Quote:
Originally Posted by miahallen
Quote:
Originally Posted by miahallen
I would also like to confirm no memtest86+ v1.7 errors with F11 for me :up: Thank you Gigabyte! :clap:Quote:
Originally Posted by tlac
Quote:
Originally Posted by Cob
Quote:
Originally Posted by blurp
Yes F9 for DS3P 2 passes on memtest no errors. :up:Quote:
Originally Posted by miahallen
I also got no memory errors with memtest+86 using the F11 bios setting for the P35-DS4 ver 2.0 motherboard using 4 x 1GB of corsair TWIN2X2048-6400C4. :)
sorry, fot the ds3p the bios F9 wich improvement give?Quote:
Originally Posted by necron66
thanks
It looked the same as F8, but I didn't use F8 for long due to the ram errors, so I'm not really sure. My g.skill PC2 8500 HK's run good now tho.Quote:
Originally Posted by diablos
Now i have the F6, F8 and F9 wich improvement give?Quote:
Originally Posted by necron66
Same here......Now they need to get the fan headers to work properly,This board has the worst fan control Ive ever seen:mad: Dont know about you guys but I think ET5 needs a major overhaul also It flat out Sucks!:DQuote:
Originally Posted by miahallen
I have followed these posts for the past couple of weeks; thanks for all the tips.
My own results are as follows:
First: I have a GA-P35-DS3R v. 2.0 MB, with Q6600 G0 stepping cpu (OEM version from newegg). I've installed 8 GB of Corsair TWIN2X4096-6400C5DHX ram. I am using Suse Linux 10.3. cpu cooling is an Arctic Cooling Freezer 7 Pro. I have an Antec Sonata III case, and I've added a 120 mm internal fan to keep the various chipsets and graphics card cool (nvidia 7300 GT, I believe; no fan on it).
I had the usual memory errors with BIOS F9, F10. I was using F8 until I realized that it was overheating my CPU; these bioses are clearly changing around the various voltages (or the way they are reported). F8 was hot. I thought F11 was also giving the same memtest errors, until people reported no problems...so I went back to the tweeking drawing board...
Using F11 (or any of the other bios versions), I've only been able to OC to 3.0 GHz (9X333), using standard 5-5-5-18 RAM timings and keeping RAM speed to the rated 800 MHz (though I think it will work slightly OC'ed; 832, I think). To achieve error free memory, according to memtest, I had to +0.2 V memory, FSB, MCH and bump the CPU voltage up to 1.40V. I may be able to lower these voltages with some additional tweeking. I ran out of patience with microtweeking only to find an unstable system,so I jumped ahead a bit (+0.1 voltages weren't working; "Is there any stable configuration at all?"). CPU temperatures are in the mid-30's idle and just over 60C under load. All of the Intel speed auto adjust/powersavings options are disabled. (In fact, it looks like Linux disables speed step when the cpu is OC'd.)
Stable/unstable are being tested using memtest 1.7 and mprime (which is prime95 for Linux, I believe). I run four instances of mprime to test. I had a hard time getting memory stable. While microtweeking, things would look initially stable, but then memory errors would crop up after about a half hour of memtest; eventually I found stability with the settings above. 5 hours on memtest without error, similar for "mprime -t". The computer is used for scientific work; I can't adopt the "well it works o.k." philosophy (nothing wrong with that, I just can't do that) I also consider the system unstable if mprime aborts on error (or if the system freezes, of course)
I've not been able to find any stable configuration above 3.0 GHz. I tried 8X400=3.2 GHz, but curiously Linux recognized this as a 3.6 GHz cpu (even though bios reported 3.2 GHz at boot up). That failed mprime in a hurry. 9X366=3.3 GHz passed memtest o.k., but quickly failed the mprime torture test. (Yet more cpu voltage? I hesitate.)
Perhaps I just have an "unlucky" Q6600... Still, 3 Ghz is nice. (I hope AMD can get their quad-core situation turned around.)
Thanks again for all the info, and thanks for any tips y'all may be able to offer!
did u set "performance enhance" to standard? That setting gives me instability with overclocks sometimes if I have it on Turbo or Extreme (especially).
Yeah, it is possible you have a relatively poor Q6600. What kind of Vdroop do you see when all cores are loaded with Prime95? At 1.425 BIOS, I see 1.31-1.33v loaded, for example.
Yes, I've always set that to standard. The flashing red warning was enough for me...Quote:
Originally Posted by swaaye
(Did you mean 1.325 BIOS, per your signature?)Quote:
Yeah, it is possible you have a relatively poor Q6600. What kind of Vdroop do you see when all cores are loaded with Prime95? At 1.425 BIOS, I see 1.31-1.33v loaded, for example.
The bios setting is 1.40V. At idle, "sensors" (the linux utility for reporting on hardware state) reports the cpu at 1.36V, I believe, although that's a bit of a guess (Its labeled "in0"). Under load it drops to 1.31V. A curious thing vdroop. Although having tested this now, idle is now 1.34V, and 1.31V under load. Perhaps I do have a bit of leeway to tweek the core voltage up a bit.
I would guess that RAM voltage, set to 2.0 in BIOS (+0.2V) is being reported as 1.95V. (labeled as "in1")
I'm never quite sure if things are being reported correctly, though...
I forgot to mention that I have a version 2.0 GA-P35-DS3R MB.
Added later: I've just tried 1.425 V and 9X366=3.3 GHz. That locked up under mprime in a few seconds. Too bad. I wonder if the OEM versions of the Q6600 overclock less than the boxed versions. The former perhaps being intended for computer makers who would not overclock, the latter intended for afficionados.
You might be a bit limited by your cooling. The ACF7 Pro was great back in K8 dual core days...but with the quads, something a little beefier would be more appropriate for high OCs. Watch your temps, and make sure your staying under 80C at load. I've run my Q6600 with up to 1.575V without problems (got me 8hrs prime95 stable at 3.85GHz) with "only" air cooling, so you still have plenty of headroom with your voltage....so long as your cooling is up to the task.Quote:
Originally Posted by LasVegas
FYI - my "only" air cooling setup was at ambient temps around 8C and 110CFM blowing through my TRUE.
I see...you are suggesting that I might get stability at higher speeds if I boost my cpu's voltage up more, but that my cooling is marginal for that. That was also my impression of the AFC7 Pro, according to reviews, but it seemed serviceable enough and cost effective. I was not really aiming to stress my system out too much to get that additional 5-10% cpu speed, so I'll have to think about whether I want to continue boosting the cpu voltage up. I'll let you know if I get a little faster and remain stable with greater voltage. I was just a little surprised to be limited to 3.0 GHz, seeing the speeds that others have gotten on "stock" voltage.Quote:
Originally Posted by miahallen
So far, I consider mid-30'sC idle, and ca. 60C under load to be just fine. Heat doesn't seem to have been a problem so far. If you are mentioning 80C as a practical upper limit, then perhaps my system will support the additional voltage; I have a ways to go to get to 80 C! (Or perhaps I should just keep my computer in a freezer...)
I'd like to report that BIOS F12 fixes the Memtest errors on the GA-P35-DS3 aswell. I'm currrently running 8GB DDR2-800 4-5-4-15 at 1.95V and the board handles it perfectly. Quite impressive.
Board revision of these P35-DS3R's doesn't seem to mean all that much. They all use the same BIOSs even. You get some more USB ports with 2.0 though, instead of serial and parallel. Rev 2.1 has some new power saving tech though so it might be more "different".Quote:
Originally Posted by LasVegas
And it seems you figured out what I meant by vdroop. My 1.325v in BIOS gives me about 1.23v load volts, which is still below my CPU's rated VID of 1.25v. Ironically, when my board is set to stock settings, my CPU is undervolted by nearly 0.1v.
Of course, you are right about the sensors being off. They are almost certainly not spot on. I don't really feel like trying to use my multimeter on my mobo though. Kinda scary to stick the probes in there and get them on the tiny points people have found to measure vcore.
I'm not sure what to tell you about your chip. It sounds like it is just not going to clock high. Overclocking is a crap shoot and sometimes you end up at the low end of the bell curve...
Are these latest bioses worth testing, I'm still at F5f myself as I found mem performance to be good. Any1 tried comparing memory benchmarks or anything between BIOS versions?
the new bioses still killing ram or no?
anyone have a good tip to clear the CMOS?
I removed the Batt, shorted with jumper, unplugged all power, press pwr button a few times, let it stand for 30 mins...
but the only thing that reset was the clock lol? help please :D
it has cleared before with above method, but not any more :s
All I ever do is reset it with the jumper.... And I've only done that once or twice. Otherwise I just load defaults after flashing it (as Gigabyte says to do).
can i get idea of what G(MCH) and fsb volts need to be set it @
Am I right to say that the BIOS uGuru functions are exactly the same as the IP35 PRO?
P35-DQ6
Just realese officially bios F7C,on giga site.
Btw,just receive new bios P35-DQ6.F7E(build 10/01/2008).For interested link is:
HERE
http://img214.imageshack.us/img214/5403/biosf7eza8.png
By stasio
Gigabyte:New utility
@BIOS-Ver 1.22
Easy Tune5 Pro-B7.1212.1
Face Wizard-B7.1212.1
HERE
Some new software last two days:
EVEREST Ultimate Edition 4.20 Build 1254 Beta with key:
HERE
EVEREST Ultimate Edition 4.20 Build 1256 Beta
HERE
Core Temp 0.96.1
HERE
Have you seen any improvement ?Quote:
Originally Posted by stasio
Could you post your bios settings?
I just got a q6600 L740A630 and can not get it prim95 stable @3.6
I know my ram is good as it is stable at a higher FSB with my E6850. After 30mins or so it reboots the computer. I have not seen anyone with this batch number.
Thanks
hlonipha,what is your q6600 VID.My is 1.31250V and I need 1.5 V in bios for 3.6GHz.
Bios F7e just start to test.
CPU Fan Control-change from(Auto-Intel@QST-Legacy-Disabled) to (Enable-Disable).
Loadline-(Auto-Enable-Disable)
Notice my CPU Fan speed 1800 RPM(800 RPM in F7c)idle
Mine is the same 1.31250v I'm running it at 1.48 right now on bios but the vdroop brings it to like1.43-1.44v on windowsQuote:
Originally Posted by stasio
What are your other settings
pcie
mch
fsb
another thing is that I see posts around with a +.2 does that means .200 or .02?
http://img86.imageshack.us/img86/2789/36tk0.png
By stasio
My Vdrop is 1.49v in windows.
pcie-+0.05
mch-+0.1
fsb-+0.15
+.2 is mean .200
ok thanks will try the new bios today with new settings.
Another thing I took my cooling pipes off and cleaned the NB and applied AS5.
My cpu temps drop was 4c
Quote:
Originally Posted by stasio
Does it fix the bios reset overclocking on reboot problem ?