Questin 1. Leave the strap on auto and then choose the DDR1000 option for memory speed.Quote:
Originally Posted by douirc
Question 2, you can't, there are no sensors on the board to get a reading from.
Printable View
Questin 1. Leave the strap on auto and then choose the DDR1000 option for memory speed.Quote:
Originally Posted by douirc
Question 2, you can't, there are no sensors on the board to get a reading from.
Quote:
Originally Posted by Pntgrd
hm...tried that. I get options 901, 1081, 1128, 1199, 1353, etc. it's like it's not reading the fsb at 450 eventhough bios and cpuz say 450.
no nb/sb temps? that's weird :(
P5Q doesn't have a ratio to do that.
Have you tried the 1081? That looks like it is as close to what you want. You couls also try setting the Straps manually and see if you get closer to 1000.Quote:
Originally Posted by douirc
this look safe, or do i need to back down the voltage? i'm on water.
i cant get the dedicated RAID working (drive xpert)
in the bios it greyed out as "automatic",
i went to splashtop and set both drives to Speed and rebooted - nothing, windows says "Drive Xpert Offline", and "show drive xpert" is greyed out
i hooked the drives up to regular sata ports and the data on them wasnt even deleted
these are 2x seagate 320gb 7200.10
edit:
lol, i disabled the IDE marvel adapter under integrated peripherals :doh:
:mad::mad::mad:Quote:
Originally Posted by Tulû
I ran blend again ; it was stable for about 210 mins. yesterday but now i got error after 30 mins. on core0. :confused:
Edit:
Increased vFSB : 1.34 to 1.36
Increased PLL : 1.56 to 1.58
Quote:
Originally Posted by Tulû
Tulu, if you are sure that your vcore/pll is stable then raising it should not solve your blend problem. Also you should take note of the packet size that is failing, if you are failing on larger packets like 1024K then you have a ram/nb issue but if you fail on smaller packets like 8-60K then you have a CPU voltage problem. If you can pass at least 6 hours of small ftt's with that vcore/pll then your problem is with your ram/nb and I am pretty sure at 1.34nb you have plenty for 425mhz FSB. You might want to check both your sticks of ram one at a time because you might just have one bad stick. I would also disable static read control and read training then try to pass at PL10. It seems you have plenty of Ram volts with 2.1 for 1133mhz with PC 9600 kit. If you can not pass a blend or ram test with SRC and Read Training disabled at PL10 then you have a problem that is beyond my knowledge of how to solve. Maybe Leeghoofd or one of the super clocking guys can help solve your problem. Hope you find your answer quickly, this is the best place to come for help and these guys here really helped me out a lot in the last few months.
Tulu - you are going to have to test you ram and cpu separately. Set the FSB:ram ratio to 1:1. so use 425MHz FSB, 851 MHz ram. Set ram timings to 5-5-5-15, and use 2.1v vdimm. Set static read control to disabled, disable read training, set transaction booster to PL = 10. Use 1.34v vNB (should be plenty for that FSB).
Use the vcore you used yesterday for 210mins stable run, and the same 1.56v vPLL.
So:
Try small FFT test for at least 3 hours, then if you pass try the large FFT for at least 3 hours. Good luck :up:PHP Code:FSB: 425
PCIE: 100
RAM : 851 @ 5-5-5-18 (trfc 55)
DRAM STATIC READ: disabled
DRAM READ TRAINING : disabled
MEM OC. CHARGER : ENABLED
AI. CLOCK : LIGHTER
PL : 10
Vcore: 1.38750
GTL REFs : AUTO
PLL : 1.56
vFSB : 1.34
NB. 1.34
DRAM : 2.10
LLC : ENABLED
If all goes well, next we will work on your ram.
Quote:
Originally Posted by CryptiK
Yes! At least one full loop with small fft's from 8-64K which takes a little bit over 3 hours. I just like to run two loops for 6 hours because one time I kept having problems with OCCT ram test and found out that I needed just one more notch up in vcore because I could pass one loop but not two :eek: Honestly though for now 3 hours is probably just enough. I suggest getting OCCT though to run larger fft's instead of prime95, for some reason large fft's in prime95 sometimes fail even though I am perfectly stable. With prime95 I only use small and blends. My stability process is usually something like 6 hours with small fft's followed by a 2 hour OCCT Ram test then a long 12-18 hour blend with prime95. I don't really like my Q6700 and if I degrade it by so much stress testing then I just get to buy me one of the new Q9650's with E0 stepping:up::up::up:
Furthur to the voltages I posted earlier, Seems like vista 64 is a lot harder on your system that xp 32. In order to get blend stable on my rig under vista64, I have had to up vDimm to 2.22 and vNB to 1.34. My memory temp is 1c degree higher under idle but loads to the same 45c as before. NB runs 1c hotter.
Another issue too, I need 1.288 vcore to be orthos stable at 4.25 Ghz. I was previously setting 1.30 with LLC on which is giving me 1.288 idle and 1.296 load. I have now changed this to LLC off and 1.36, which fives me 1.320 idle and 1.288 load. Rig is totally stable. The strange thing is with LLC off I get lower temps under load. Before I was hitting 66c load, now with LLC off and same volts, i am hitting 61c load. My idle temps with LLC off are also around 5c higher but im not bothered about idle temps, load is all that matters. Dont know if this is always the case for everybody but surprised and pleased with my new lower load temps with LLC off
To answer why your idle temp is higher:Quote:
Originally Posted by Mumid
LLC on = 1.288 idle
LLC off = 1.320 idle
more volt -> more temp ;)
I wasnt questioning why my idle temps were higher with LLC off, thats obvious. I was wondering why my load temps were lower with LLC off even though with the same volts as before!Quote:
Originally Posted by JHS
Im guessing different ambient...
same ambient mate, i tested it. Really strange, i cant get my head round it!Quote:
Originally Posted by jas420221
P5Q vanilla new bios v1104 dated 8/19/2008.
i havent completed all my testing with it but so far its similar to 1004, slightly better performance (clock for clock), 500fsb stable, and nbs400 is working albeit slightly slower then nbs333.
download 1104
Quote:
Originally Posted by PortCitySlim
Quote:
Originally Posted by CryptiK
My reply delayed ; because i was running blend and here is the result :rolleyes:Quote:
Originally Posted by PortCitySlim
vFSB : 1.36
PLL : 1.58
Vcore : 1.38750
DRAM : 1133 @5-5-5-18 trfc 55 @2.1 V
NB : 1.34
GTL REFs : Auto
Static read : Enabled
Dram training : Enabled
Mem OC Charger : Enabled
AI Clock : Lighter
PL : 10
http://i37.tinypic.com/f34xus.jpg
Try my specs in my template i listed in my signature .. i am using BIOS 1201 and i suggest trying that BIOS as well.
1200 mhz memory is overkill and the only way to get it 1:1 is to run a 6 multi x 600fsb STABLE, but not many quad-cores will run 600FSB and mobo's/bioses.
So you will have to sacrifice memory mhz for CPU GHZ.. i running 8 x 450 ram is at 901 mhz ... even though it is rated 1000mhz memory ... even if i had 1100 mhz memory and wanted to sue it all . i would be trying to push all i had in hardware to its limits and that is not good for 24/7 daily usage even if it was stable.
Another thing is HEAT which is major factor in Ooing and stabiltity, you can only go as fast and high as the heat will allow and if to hot when full load .. then you have to sacrifice GHZ for heat reduction or get better cooling and try fir higher stable ghz with cooler temps.
Anyhoo .. look at my specs and download my template .. print it off and adjust from there.... 901 mhz on memory at 3.6ghz is fun and good enough for me.
Quote:
Originally Posted by BeastNotro
Thank you for your advice.
But i learned that (from the clock masters) i think i 've cpu issue, not RAM ; because as you see the prime screenshot i got error on core0 on 40k test so i think it's a Vcore issue.
Looked for your template ; but you have your GTL REF 1/3 is greater than 0/2 ; but mine must be vice-versa because i always get error on core0 so it seems my core0 is weak.
Btw , my rams are rated as pc2 9600 @ 5-5-5-18 @2.1 V
The 0.63 and 0.67 are the most common settings for the Q6600's, but not written in stone, but are the most common used and i suggest putting them to that.. basically use my template settings and try all of them out and you may get stable, but your VID is a higher VID then mine and you may have to go even higher voltage then what i am using and that causes a HEAT issue, you maybe one of those who has to settle for 3.4ghz @ 1.42v or 1.43v or 1.44v or 1.45v.
Look at my voltage = 1.43 full load and in Bios 1201 it is 1.425 .... so try it out... is all . keep an eye on your temps ... under 70C is fine .. 60-63C is average on air cooling for CORE 0 full load.
Update BIOS to 1201 and try settings as no one here can absolutely state 100% what you should run at, but my settings have helped many as they had to adjust this or that for themselves to get stable, but at least they have a very good reference point from my template.
You may have to settle like i stated with 3.4ghz and also maybe memory @ 901mhz to 1080mhz.
Also one note .... CLOCK MASTERS .. dude i ain't no newbie when it comes to overclocking componets either and i like to state everyone has things to suggest or use as every componet can be and maybe different from others.
Another NOTE my ram is rated at 2.0-2.1v at 1000mhz i am running @ 901mhz, but if i run it at 2.0v i would error out, so i dropped to 1.96 and i pass whatever can be thrown at it now.
How can I disable SMART function?
I'm with a RAID configuration, when in BIOS the SATA are configured as RAID there are not the SATA 1/6 and each settings for every port so I can't disable SMART; tried to set IDE instead of RAID, re-entered in BIOS and disabled SMART on SATA 1 and 2 (where I have the HDD), saved and re-entered, changed in RAID but at the POST i see that SMART is still there.
How can I disable it defenitively?
Under the MAIN tab in the BIOS look through that section of stuff as there should be SMART MONITORING (AUTO) set it to disabled.
edit:
Dang it i looked in MANUAL page 3-11 it states that they are not user configurable.
Yes and smart is under each drive, but ti will state N/A if not IDE drives are plugged in.. go figure. lol
But on PAGE 3-12 it states it can be set Auto, Disabled, Enabled.. i can't go my bios to check as i am folding right now.. so have a look around in thats ection of bios and look in the manual.
Yes, I have looked the manual and the bios section, but, as i have written, when you chose RAID the configurable SATA port 1/6 disappear and you can't chose anything of the settings... is for that reason that i tried to configure first the drive when they are in IDE mode but when switched to RAID my chenges had no effect.
Hey there .Quote:
Originally Posted by BeastNotro
Thanks again for advice ; but i still think i have a cpu issue not ram .
"clock masters" , don't take it personal please :( , you missunderstood me i think . I know that everybody here knows something better than me.
I 'll try a one notch higher Vcore . Will disable Dram static read and mem oc. charger .
Btw , my temps were max 68 68 62 62 at prime blend for 6 hours and my room is abaout 30C.
Are these temps normal ?
Thank you very much .
Quote:
Originally Posted by Tulû
30C ambient is really high, but I feel you there since I live in the southern US my room temps get almost 30C at times. Those temps seem high to me still even considering the ambient temps, under load I do not even break 60C at 1.38vcore and that ws this summer when my ambient temps were close to yours. Disabling Static Read control/read training is a good idea to start with, later after your stable you can try enabling then. I have no idea whether the MemOC charger has any effect at all, I use it when I use Static Read Control but have never tested without it.
Can I say "It's rock solid stable" now ? :hm:
http://i36.tinypic.com/2n23amf.jpg
nope, it has to run 10 hours, start over and try it again.Quote:
Originally Posted by Tulû
j/k :p:
Stability is up to you. Personally, I usually run 3 hours of Orthos small FFT and then 3 hoursof Orthos blend. If it passes that, good enough for me.
As long as I can game, do folding@home and general desktop usage without BSODs, it's stable enough.
This is what mine does with the 0803 the highest revision I can go right now...
http://img398.imageshack.us/img398/5...sseskk9.th.jpg
Does anyone have settings for a Q9550 doing 3.8? I just got my system bled and reloaded and here is where I am starting. I have read 106 pages and this is some thread. Full specs in sig.
PHP Code:ASUS P5Q DLX Template : SiGfever
Ambient: 24.4c
CPU: Q9550 vid = 1.25
Memory: G.Skill F2-8500CL5D-4GBPK
JumperFree Configuration Settings
AI Overclock tuner: Manual
CPU Ratio Setting: 8.5
FSB Strap to North Bridge: 333
FSB Frequency: 415
PCI-E Frequency: 101
DRAM Frequency: 831
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: Manual
1st Information :
CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: 55
Write Recovery Time: AUTO
Read to Precharge Time: AUTO
2nd Information :
READ to WRITE Delay (S/D): AUTO
Write to Read Delay (S): AUTO
WRITE to READ Delay (D): AUTO
READ to READ Delay (S): AUTO
READ to READ Delay (D): AUTO
WRITE to WRITE Delay (S): AUTO
WRITE to WRITE Delay (D): AUTO
3rd Information :
WRITE to PRE Delay: AUTO
READ to PRE Delay: AUTO
PRE to PRE Delay: AUTO
ALL PRE to ACT Delay: AUTO
ALL PRE to REF Delay: AUTO
DRAM Static Read Control: Disabled
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: Manual
Common Performance Level [12]
Pull-In of CHA PH1: DISABLED
Pull-In of CHA PH2: DISABLED
Pull-In of CHA PH3: DISABLED
Pull-In of CHA PH4: DISABLED
Pull-In of CHB PH1: DISABLED
Pull-In of CHB PH2: DISABLED
Pull-In of CHB PH3: DISABLED
Pull-In of CHB PH4: DISABLED
CPU Voltage: 1.36250
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.54
FSB Termination Voltage: 1.32
DRAM Voltage: 2.00
NB Voltage: 1.32
NB GTL Reference: 0.68
SBridge Voltage: Auto
PCIE SATA Voltage: AUTO
Load Line Calibration: Enabled
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : Normal
NB Clock Skew : Normal
CPU Margin Enhancement: Compatible
Advance CPU Settings
CPU Ratio Setting: 8.5
C1E Support: Disabled
Max CPUID Value Limit: Disabled
Intel® Virtualization Tech: Disabled
CPU TM Function: Disabled
Execute Disable Bit: Disabled
Sorry can't really understand what your problems are??Quote:
Originally Posted by SiGfever
You mean your systems fail to POST using these set of setting?
Quads usually should stress NB more so try adjusting NB volt a bit may help
try like 1.3v NB and 0.61X for GTL Reference.
PL use 11 first 10 for CL5 is a bit stressful just try to isolate the problem
(Sorry I'm a newbie from Hong Kong and I suckz in English)
Thank you for your reply. The system posts fine at 3.52 but I want to take it to 3.8 if possible. I understand that every setup is different but if someone with a 45nm quad like mine could help with some setting I would appreciate it.Quote:
Originally Posted by ci2kla
I will make the changes that you listed.
Heres what I have been working on, board runs fine so far.
http://img178.imageshack.us/img178/3...occtld5.th.jpg
Trying for 475 stable tomm.
Larry
edit; I forgot, it runs 1.24 or so Load LLC disabled.
Actually from what I've seen is 9550 should be able to reach 4Ghz @ around 1.3v without any issues.Quote:
Originally Posted by SiGfever
By the way, is the max multiplier for Q9550 8.5? If so...you may need 470 FSB to make it and that won't be easy to go above 450FSB as it requires futher tweaking. Moreover for OCing quad processors CPU GTL Reference plays an very important role. FYI:up:
Yes the max is 8.5 on the Multi. I will do some testing at these settings and see where it goes, thanks again for your help.Quote:
Originally Posted by ci2kla
Having a wee problem here, not with my 3.6 clock - thats been running steady and zero glitches
Its a bios problem; I noticed that my cpu was not speed stepping
Tired 1201 and manually re-entering my OC settings and even with cpu ratio Auto and speedstep enabled, its 3.6
Even tried just all default and its always 2.6 default
I've gone back & forth between 803 and 1201 and notice 803 does not have the speedstep option at the end of that section
Tried an internal reset, but not used the jumper to clear cmos - thats next step, its just my GPUs are in the way & lazy
I'm not worried about the extra power or heat on the cpu as its a crossfire rig, but although I think its just buggy bios, I thought I'd post what was happening and ask: do you guys generally leave the overclock set 24/7 or do you use speedstep yourselves?
cheers
You're welcome.Quote:
Originally Posted by SiGfever
Now...for my own question...I think it's a motherboard issue...
Recently I'm seeking for a higher memory bandwidth so I'm playing around with memory frequencies, CL staff and I discovered a problem. With CL4 or CL5 this motherboard makes zero difference on memory read bandwidth at all! I wonder why...
(Also for memory OCs I've found that with CL4 I will be able to POST at PL7 even @500FSB)
Here's my spec:
Asus P5Q Deluxe BIOS 1201 (actually 1201 has made an effort on OCing memory)
Intel Core 2 Duo E8400 Q807A19* 500*8 W/ TB120
G.Skill 6400 PHU2 2GBHZ @ 1200Mhz CL5 2.16v BIOS W/ TT.V1R
GeForce 88GT Forceware 177.79
If anybody have P5Q-D would you guys please help testing if CL5 and CL4s make any boost on read bandwidth? Thanks in advance...just really confused:(
Here's the Screenshots in Everest memory benchmark:
http://i317.photobucket.com/albums/m...e/4-4-4-10.png
http://i317.photobucket.com/albums/m...e/5-5-5-15.png
It says in the manual that CPU Ratio Control needs to be on Auto for it to work.Quote:
Originally Posted by mac_69
Larry
I love how I get my shiznit stable at 3.8 easy and run about the equavalent of 80hrs of prime and no one says congrats.
Here is a cookie.
*Cookie*
Congrats. :clap:
3.8Ghz isnt a feat. 4Ghz is the magic number that gets you an XS reacharound, though I must say that Vcore is awfully low! :up:
Yep it is larry, both places, its strange bios bugQuote:
Originally Posted by Big Lar
I'll bite the bullet and rest the jumper tomorrow
:rofl: Just jealous here. Can't get 3.6 stable at these volts you use. Can you post your template for 3.8 please :up:Quote:
Originally Posted by truehighroller
I will say thank you for that.. I am on air here man, 3.8Ghz 24/7 is pretty good for air and at the voltage I have for that,,. I could prob go lower on the vcore I'm thinking. I was in the Gigabyte overclocking conetest not to long ago and was in second place on air "beating a person on water" then a bunch of phase and ln2 people jumped in and I was bumped to 7th place at the last second of the whole thing. I would love to test new stuff for vendors.. All I need is some water cooling and I'm at 4.0Ghz easy.Quote:
Originally Posted by jas420221
ASUS P5Q DLX Template :Quote:
Originally Posted by mac_69
JumperFree Configuration Settings
AI Overclock tuner: MANUAL
CPU Ratio Setting: 8
FSB Strap to North Bridge: 333
FSB Frequency: 475
PCI-E Frequency: 103
DRAM Frequency: 1141
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: MANUAL
1st Information :
CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: 50
Write Recovery Time: AUTO
Read to Precharge Time: AUTO
2nd Information :
READ to WRITE Delay (S/D): AUTO
Write to Read Delay (S): AUTO
WRITE to READ Delay (D): AUTO
READ to READ Delay (S): AUTO
READ to READ Delay (D): AUTO
WRITE to WRITE Delay (S): AUTO
WRITE to WRITE Delay (D): AUTO
3rd Information :
WRITE to PRE Delay: AUTO
READ to PRE Delay: AUTO
PRE to PRE Delay: AUTO
ALL PRE to ACT Delay: AUTO
ALL PRE to REF Delay: AUTO
DRAM Static Read Control: Disabled
DRAM Read Training: Disabled
MEM. OC Charger: Disabled
dAI Clock Twister: stronger
AI Transaction Booster: manual
pl set to 8
pullins set to whatever is default.
CPU Voltage: 1.25
CPU GTL Voltage Reference (0/2): .63
CPU GTL Voltage Reference (1/3): .63
CPU PLL Voltage: 1.56
FSB Termination Voltage: 1.22
DRAM Voltage: 2.02
NB Voltage: 1.28
NB GTL Reference: AUTO
SBridge Voltage: 1.22
PCIE SATA Voltage: 1.6
Load Line Calibration: Enabled
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : AUTO
NB Clock Skew : AUTO
Advance CPU Setings
CPU Ratio Setting: Manual
CPU VID: Default
C1E Suppport: Disabled
Max CPUID Value Limit: Disabled
Intel® Virtualization Tech: Disabled
Vanderpool Technology: Disabled
CPU TM Function: Disabled
Execute Disable Bit: Disabled
I had 3.8Ghz on air as well with a 780i (Tuniq120) months ago. I thought it was fairly easy to get to 4Ghz as well considering it took me just about 3 nights total (about 8 hours not including Prime) to get 3.8Ghz stable on the P5Q-D. Not so. Im on water and its still proving elusive.
EDIT: I am however amazed at the low voltages all around really. I cant get 3.8Ghz stable without 1.34 NB...especially with that ram running (for no real reason) so high.
What is the best program for reading temps?
Thx, This template works nice, Just need to test it and mesure temps. What programs are best for testing stability? P95 or that new Intel linpack?Quote:
Originally Posted by BeastNotro
The new real temp, using the correct Tjmax is the most accurate.Quote:
Originally Posted by gallag
You should really use both, I use Linpack to get some initial stability (30 runs), then use Orthos to run at least 12 hours of small FFT, then 12 hours of large FFT.Quote:
Originally Posted by gallag
Then gaming hard is a good test to make sure it can stand up to 3D work.
You can go for shorter or longer if you like, everyone's definition of stability is different.
Thanks for the reply, I ran the intel burn test and within a couple of seconds real temp was reading 82degres.Quote:
Originally Posted by CryptiK
beastnotro, or anyone that can help, I used your template to get to 3.6, Any ideas what I should change to get a cooler 3.2-3.4?
Your cpu temp is basically totally dependent upon vcore and vFSB. Just keep lowering vcore as far as you can whilst remaining stable at the given speed. However from your temp it sounds like you're still on the stock cooler, and if so, upgrading to an aftermarket cooler would be a wise investment.
UMM --- he ran Linpack and it will push the CPU to 80-84C and thats normal for it to do so, depends on cooling... you don't want to see 88-90C-95C as that is not GOOD.. but 78-80 and 84C max is normal for many ppl using Linpack.. now the temps will fluctuate when testing too like go from 68-78-80-82... just keep and eye on it.
Mine hits 80-84C... air cooled with thermalright Ultra 120 Extreme, when running Intel Linpack .. thats only on 1-2 cores the other 2 cores hit 75-78C.
Nothing to worry about as Linpack will push the temps that high.... only suggestion is don't run more then 10-15 loops at a time as running at 80C temps for TO LONG (like weeks or days) will degrade CPU over time.
As for running lower GHZ .... why? .... if 3.6ghz is stable and you have effective cooling = running Linpack for 10 loops (3-6 times) and Orthos Blend (12 hours) and temps are around 60-58-57-58 or 65-62-58-60 (WHEN RUNNING ORTHOS)...... then all is within good performance ..... mine hits those.
So use Linpack as a FAST stability test and make sure all numbers are the same for both sets of numbers . if so you passed, if not then you not pass ... and if you not sure about things POST here a screen shot.
By the way WHAT cooling are you using?
Quote:
Originally Posted by BeastNotro
OCZ vendetta 2 cooling, Running p95 large now and temps are 73 73 72 72. Is this normal or would you recomend going to 3.4?
What program you using for monitoring temps?
Is it realtemp if so set it to 95C tjmax.
I am folding right now on my GPU and CPU and my temps are 62-58-53-55 using realtemp at 95C TjMax..... now there is the arguement of it should be set to 100C TJmax... but many have tested and stated 95C TJ max is more accurate .. its up to you.
And if it is at 95C TJmax now .... then 73-73-72-72 is getting into the temps i would seek better cooling paste like thermalright chill factor (apply thin layer on CPU only) and/or i would also lap the heatsink and drop that temp down at least another 5C-6C.
So post back what you use and such for paste and programs..... etc.
Wow I didn't think they would get that hot at all. I have max temp of 52*C when running Linpack and 48*C when running orthos.
Your temps are not normally going to be that high during normal use, so I'd stay at 3.6GHz.
Yeah thats because your at 1.32v and its a 10C-15C difference from 1.32v to 1.43v... :)Quote:
Originally Posted by CryptiK
Room temps will make a difference as well, mine is usually around 24C ambient.
Plus your chip is a 45nm and newer in specs that it runs lower.
The main reason to is yours is a Dual-Core not a Quad-Core.... :)
Quads run hotter ... especially a 65nm Q6600...... :)
You have a 45nm chip, we have a 65nm chip... these heat quiet more and also need higher voltage...
LMAO -- we posted at same time.. we have same brain wave dude.. :)Quote:
Originally Posted by RejZoR
Can anyone verify that with the new bios the memory overvoltage doesn't exist ?
Thats a nice clock :) Whats the Vid on your chip? I have the neutered version of your chip at the same speed and testing.Quote:
Originally Posted by truehighroller
Larry
[img=http://img381.imageshack.us/img381/1981/3800occtjf9.th.jpg]
It does exist 1201 bios i use 1.96v and at full load i get 2.00v.Quote:
Originally Posted by mathewgx
Non-Full Load it is around 1.97v-1.98v
Its not like the other bios before where 2.00v was 2.04v non load, then 2.1-.2.4v full load.
So basically if you ram MAX is say 2.1v like mine for example ... you not want it overvolting it .. FULL load past that... so stability also is the issue.
When i was at 2.00v in bios linpack would fail .... so i set it to 1.96v and i pass with flying colors.
Yeah 3.8ghz is not a uncommon task to achieve, but only a few or so have done so on a Q9450 100% stable and for 80 hrs.... do you have screen shot of 80 hrs... :)Quote:
Originally Posted by truehighroller
Whats the equivalent program you used to make that assumption it was equivalent to 80 hrs of prime.... umm let me guess linpack? . if so did ya run it 100 passes or something.. :)
If so thats awesome dude and i have milk and cookies sent to you VIA pm rail.
Seriously good job dude .. now try for 4.0ghz stable and i will very very very impressed..... :)
Why casue i want 4.0ghz stable and can't cause tis Q6600 is a HOT mother when trying to do 4.0ghz STABLE testing.
That is an awesome result, don't know how I missed that.Quote:
Originally Posted by truehighroller
:cheer::cheer::cheer::cheer::cheer::cheer::cheer:
:cheer::cheer::cheer::cheer::cheer::cheer::cheer:
:cheer::cheer::cheer::cheer::cheer::cheer::cheer:
:cheer::cheer::cheer::cheer::cheer::cheer::cheer:
And yeah, I remember how hot my old 65nm used to run. Even using a TRUE and you're still hitting 80+ is crazy. Poor bastards :p:
Yeah, I am useing real temp with 100TJmax 2hours 15 mins stable now with prime95 large fft with max temps 77 77 75 75, SO should it be o.k then?Quote:
Originally Posted by BeastNotro
Well no ... as that is to hot for daily usage or folding and such ... i use 95C TJMAX.... as that is more accurate based upon many testing it .. including myself.
Even at 95C your temps will be around 72-72-70-70... you want UNDER THAT like 60-65 MAX... like my temps.
What thermal paste are you using?
Also when priming i use Orthos Blend as i hate Prime 95, but with orthos you need to run 2 instances at once to get all cores running. make 2 folders and call one of them Orthos 2, and you have to set priority to 2 cores and then again with the other two cores within Task manager .... if you not know how.. let me know.... but prime 95 is OK, but i not like... :)
.. and you need at least 12 hours to be stable in many ppls books... and run BLEND .. not large, not small.... but BLEND.
Ok so if you use better paste and lap heatsink i bet you can get 65C temps.. NP's if lapped right and using Chill Factor paste OR MX2 paste... and if properly applied = a thin layer spread on CPU .. like maybe a paper thickness.
I should also note that even at 3.4ghz at around 1.36-1.38v-1.39v is need for stability (DEPENDS ON CHIP) .. and your temps maybe still around the 70C range with your setup as it is now.. you want 65C max. well give or take a few degrees.
Quote:
Originally Posted by BeastNotro
Thanks for all your help man, Going to see if I can drop the voltage a bit, Would 1.425 down to 1.4 make much difference?
I edited my post before you posted .. check the last part of it out.Quote:
Originally Posted by gallag
Bascially you need to do something about cooling right now.
And if you drop to 1.40v.... then stability maybe an issue for you, but temps will still be around the same as you have now.
I asked you twice now ... what PASTE are you using?
Also what is the CPU's VID?.. click on the info button in real temp till you see VID... like 1.2875 or 1.32 or 1.35. what is it?
Running p95 blend test now and after about 10 mins max temp is 65, This is @ 1.4v EDIT this is with 95TJmax in real tempQuote:
Originally Posted by BeastNotro
Useing the paste that came with the cooler, OCZ white grease type of stuff, VID is 1.25
Ok the paste from OCZ is pretty decent stuff if applied right.
Now knowing your VID is 1.25 is good a little lower then mine .. so 1.4v 65C max is ok in ten min, but watch it as in about 30 min to 1 hour into prime it shall rise a little.. and that will be the true temp MAX.. not ten min.
Check temps within the first hour.... and if stable and temps still around 65C max let her run for a minimum of 12 hours or unless it fails before then.. :)
Now with that lower VID you may get away with 1.38v-1.39v 100% STABLE
Sorry for hyjacking this thread and demanding more of your time but could you give me a quick lesson on what the VID reading means, And just so we are on the same page, My VID reading of 1.25 is vith a 1.4v not 1.425v.Quote:
Originally Posted by BeastNotro
Agian, Many thanks:)
The VID that shows in realtemp no matter what VOLTS your at is the TRUE VID of the CPU.
VID is the STOCK VOLTS AT FULL LOAD with or without SPEED STEPPING enabled, but with speed stepping enabled the VID shows lower say in a program like CPU-Z, but in RealTemp or even CoreTemp they will show the real VID.
Lower VID is better, but does not always mean that ppl can HIGH OC with a lower vcore as some ppl with a VID of 1.22-1.25 can't even get 3.4ghz stable cause of CORE issues or other things.
Generally 1.25-1.2875 VID is common and most generally can do 3.4-3.6ghz with decent Vcore range of 1.35-1.42, but some have had to use 1.45-1.49v again depends on CPU and other factors.
If anyone knows anyone selling a used P45 board at a decent price send them my way. i have a P5Q3 DLX but i want to get a DDR2 variant as well.
My VID is 1.2250 in real temp. I think this chip has a fsb wall but, I will try to see how far I can push it today for the fun of it and post my results.
All most there I think,
http://valid.x86-secret.com/show_oc.php?id=408268
Edit: That was as good as she would do and I couldn't run the burn in test without crashing there. I did get it to run stable at 485 x 8 3.88Ghz, but I had to pump 1.35v through it to do it.
This is my new sweet spot now though.
http://img368.imageshack.us/img368/2...spotcj0.th.jpg
This chip just doesn't want to go over 490 at all with air. I think when winter comes around I will stick it outside and run my cables inside and see how high I can get with good cold air. I can't afford a water setup.
I want to see 4.5ghz on a Q6600 and a Q9450.
Hey guys, does anyone know if you can modify the fan profiles for the chassis fans on the MB? I have it set to the silent profile but I want to slow the fans down further...
That would be nice to see on air.Quote:
Originally Posted by BeastNotro
Unless he is folding, it's unlikely the cpu would get above 65*C under 'normal' usage - folding all day at 100% cpu usage is not normal usage ;)Quote:
Originally Posted by BeastNotro
Also, blend is not good enough, once the FFT's get longer than 64K (takes about 3 hours) it starts to load the ram more and the cpu less, therefore not stressing the cpu alone for 12 hours. I have seen many systems fails small FFT after 8 hours or so, and blend will simply not push it hard enough for long enough. You may pass 12 hours blend, but then when you run 12 hours small FFT, it may fail after 3 hours. I've seen it happen more than once.
To do it right, is to use small FFT for 12 hours+, then large FFT for 12 hours+, then a blend run of around 12 hours to cap it off.
Who's to say what is "right"?
If it's stable enough for what he uses it for, why run small FFT for 12+ hours? That's just obsene in my opinion... I have 3 hours worth of small FFT stability and I fold 24/7 without issue.
I have never and i repeat never have had any issues after running Orthos Blend for a minimum of 12 hours and thats on several OC's that i have had over the past 5-6 years.
As for daily usage yeah i should of corrected that as yeah for normal surfing and what have you the CPU will not hit 65C .. heck even gaming i am lucky to hit 43C in Cod 4.. so yeah if folding like i do most of the day 12 hrs or so... then 65C is my max limits .. give or take a few as the one core will fluctuate at times from 63-67C depends on what it has to do, all other cores remain around 60C.
Does anyone know when did ASUS add support for E0 wolfdales? Specifically, starting with which bios revision? From the ASUS website, it seems that only 1004 rev from 07/29/08 supports these new stepping cpus properly, which probably means that most mobos in the retail channel will have an earlier rev of the bios and do not support E0s...
Well, there is no 'right' per se, just an approximate maximum stress test length that will satisfy most people that their overclock is truly stable. I have read of systems running orthos for days on end, and after the 24th day, it failed. That system was not truly stable, as a pc should be able to run something for any length of time without failing or crashing.Quote:
Originally Posted by Polizei
12 hours is just taken as the gold standard of stability, as most people need to use their system daily and that is the longest you can really run a test before it starts to impede your use of the system.
As I said, I have seen lots of overclocks fail orthos after 8 hours, and I like to know that my pc is as truly stable as I can make it, as I use it for work often and cannot afford it crashing in the middle of something. It gives me a sense of security. What you do is totally up to you.
It seems it's a good idea to set AiBooster to some fixed value. When i left it set to AUTO my PC generated errors in Orthos after 15 minutes. When i set it to Moderate, Orthos worked fine for 6 hours mixed mode. Looks like AUTO set it too high by itself and caused instability. So don't forget to select one fixed mode for this feature.
You gotta be kidding lol... even if the CPU is megacooled, the board wil give in first prolly at 500FSB plus...Quote:
Originally Posted by BeastNotro
Mobo will post and seemed to work fine with Cero's rig ( E8600 on the P5Q dlx) with the stock bios... but it's mostly better to update the biosQuote:
Originally Posted by anzial
for full CPU feature support
You mean clock twister ?Quote:
Originally Posted by RejZoR
Hi there!!
I have a problem with my setup, and i´m getting very, very upset :mad:
My system:
Q6600 cooled by Tuniq --> L806 B346; VID:1,2625
P5Q Deluxe (BIOS 0803)
GSKILL PC2-8500 (2 x 2048) F2-8500CL5D-4GBPK
Powercolor HD4870
2x Seagate 250GB (ST3250410AS)@RAID0
Corsair HX620
I´m now at 3GHz, with a 333FSB and the memories at 1066Mhz and my voltage definitions are:
CPU GTL Voltage Reference (0/2)---> 0.630
CPU Voltage-----------------------> 1.268
CPU GTL Voltage Reference (1/3)---> 0.670
CPU PLL Voltage-------------------> 1.52
FSB Termination Voltage-----------> 1.28
NB Voltage------------------------> 1.18
DRAM Voltage---------------------> 1.96
PCIE SATA Voltage----------------> 1.50
SBridge Voltage-------------------> 1.10
NB GTL Reference-----------------> 0.630
My main problem is the fact that the system only boots if a specific divider (FSB/RAM) is used, in my case, FSB=333Mhz and Ram frequency=1066MHz, so the ratio is = 3.201201, i guess.... I´ve tried every one of the other dividers, at the exact same settings, and the system won´t boot. I don´t understand why the hell the memories won´t seem to allow me to use another FSB/RAM ratio.
With this divider, i can achieve ~3.15GHz (RAM~1120MHz), but above that frequency the system becomes unstable. The problem, i think, is the memory frequency (won´t pass memtest but seems stable at the CPU test~Prime95).
Could anyone help me?
set multi to 8X plz and FSB to 400 --> ram at 1000 or 1066Mhz then...Quote:
Originally Posted by Blasar
You got already a nice kit of ram there, my Gskills don't even breach 1100mhz... don't waste them for a few mhz more...
With intel boards dividers are locked, with nvida boards you could alter them to plus minus your ram specifications... if you want high speeds on the ram 1gb kits are the way to go or go flat out on 4gb PC8800 rams...but I have my doubts on them for 24/7 ops... I use 400FSB with 1066Mhz on the rams... pretty happy with that...
PHP Code:JumperFree Configuration Settings
AI Overclock tuner: AUTO
CPU Ratio Setting: 8X
FSB Strap to North Bridge: AUTO
FSB Frequency: 400
PCI-E Frequency: 101
DRAM Frequency: 1066
DRAM CLK Skew on Channel A1: AUTO
DRAM CLK Skew on Channel A2: AUTO
DRAM CLK Skew on Channel B1: AUTO
DRAM CLK Skew on Channel B2: AUTO
DRAM Timing Control: MANUAL
1st Information :
CAS# Latency: 5
DRAM RAS# to CAS# Delay: 5
DRAM RAS# Precharge: 5
DRAM RAS# Activate to Precharge: 15
RAS# to RAS# Delay : AUTO
Row Refresh Cycle Time: 55
Write Recovery Time: AUTO
Read to Precharge Time: AUTO
DRAM Static Read Control: Disabled
DRAM Read Training: AUTO
MEM. OC Charger: AUTO
AI Clock Twister: AUTO
AI Transaction Booster: AUTO
CPU Voltage: 1.30
CPU GTL Voltage Reference (0/2): AUTO
CPU GTL Voltage Reference (1/3): AUTO
CPU PLL Voltage: 1.56
FSB Termination Voltage: 1.30
DRAM Voltage: 1.96
NB Voltage: 1.22
NB GTL Reference: AUTO
SBridge Voltage: AUTO
PCIE SATA Voltage: AUTO
Load Line Calibration: enabled
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew : Normal
My bad, yes i meant Ai Clock Twister.
any one here with a 3.8 or higher Q6600? i have mine at 3.780 :( i cant get over 420 fsb if anyone has gotten the 3.8 show me your specs plz!!! no matter what i do i cant get it to boot at 425 bus! btw my vid is 1.225
here are my settings
cpu voltage:1.5
gtl refs :auto
pll:1.6
fsb term:1.56
nb voltage:1.38
llc enabled
everything else auto
I´ve tried the settings you´ve said, and the result is = no boot :mad:Quote:
Originally Posted by Leeghoofd
I don´t know why, but the only way for my system to boot, is to select the 3rd option (starting from the end of the list) for the ram divider, any other option = no boot.:rolleyes:
I´ve tried to give a little more volts to all the fields and the result is always the same. :(
Is it normal? My P5Q Dlx only accepts one of the options to the ram divider, and my question is why?
I´ve tried ti update (EZ Flash) the BIOS to the 1201, but after a couple of tries, the BIOS was corrupted :down:
I just don´t get it... I´ve a CPU that does 3.0GHz at 1.2625, and i´m stuck because the mobo won´t boot if i don´t choose the DRAM frequency that "she likes". Damn..
I´m open to suggestions...
ASUS P5Q 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5Q Pro 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5QC 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5Q-E 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5Q Deluxe 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5Q Premium 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
ASUS P5Q3 Deluxe 1306 Beta Bios
[ 1306 ]
----------------------
Enhance the memory compatibility.
:ROTF::rofl:Great can I actually boot into windows with this one lol..
Cool. Got to try it when i get from work. Last beta enhanced my overclock by 10MHz on FSB. 10-15 MHz more and i'll be very happy hehe
Quote:
Originally Posted by RejZoR
I can try this one now as I am working at home today . :up:
2nd OC attemp :) My personal goal is 450FSB ... Higher than that is a plus already for me :D woot good thing that I did not exchange this for maximus 2 formula :D I opted to buy a better ram than add cash just to get the MF2 :D
http://img.photobucket.com/albums/v6...73/bartOC3.jpg
Replaced my G-skill DDR2 800 with Patriot Viper Fin 1066
@truehighroller ....is it possible for you to post or PM me your BIOS settings? I would like to try if I can reach that with my current setup :)
Can someone tell me how does this board behave with 4 sticks of memory? I've got 4x1gb 8500-rated memory (D9GMH), just want to make sure that P5Q will be able to handle them at top speed...
In my case for some weird reason it clocks higher with 4 sticks than with just 2. I have 2 generic PQI sticks and 2 better ones from TeamGroup Elite with heat spreaders. And it works better with all 4 of them than with just TG Elite ones. But hey if 6GB clocks better than just 4GB, why would i complain lol
:rofl::ROTF::clap::down:I still can't boot into windows with this new one either.
No matter what settings you use? Even at stock? If so maybe it's time to RMA the board.
Sorry to post this in here,but it might not get seen in the p5q-e thread.
I have just flashed my p5q-e to 1603 bios,my 6750 is now running 3.6ghz at 1.375vcore(cpu-z says 1.33xxv) seemingly fine.I have just ran 3dmark06 and it never failed.I dont stress test,i just run the pc and see if it errors or not.
I can run on the 0803 fine and this is what my board that I rmaed was doing as well. I talked to the top engineer at ASUS here in America, and we both think it is tied to the new CPU Margine Enhancement feature they have added to the newer bios revisions. They might be buying me a new cpu if they fried out whatever the setting is on the cpu itself.Quote:
Originally Posted by CryptiK
and your point is.........:shrug:Quote:
Originally Posted by TI66ER