My Vid is in my signature now.
Printable View
My Vid is in my signature now.
My chip hits 80 degrees on Load @ 3,8GHz @ 1,392V :S
Want to know the funny thing..... my chip is a (look in signature i update it) batch number.
Very few people get that lucky with a L803-L807.
The chips I've had are:
L804, L733, L743, and L737
no thats under prime small fftsQuote:
Originally Posted by BeastNotro
I've got 41C on cpu today in idle @ 2.4GHz. Was 104F outside today, a bad day to do any benching.
I just got a quad up to 3.6ghz with 1.4v in the bios.. 1.38v in cpu-z
after 2 hours in prime core temp was showing 66ºC and real temp 61ºC
which one should I believe?
oh and by the way, the core temps were so different one from another that i ended up using a wire to pull up my TRUE a bit and now the temps are even across.
I get the same thing as u do, real temps allway being 5C lower. I trust core temp and everest, its safer. Plus 2 programs vs. one.
Set your TJMax in Everest and CoreTemp to 95 Deg to match RealTemp. See what happens.Quote:
Originally Posted by fatguy1992
At 3.6GHz/1.4Vand full load under Air, RealTemp should be reading accurately regardless of whether or not you did the idle calibration. So should everest and CoreTemp for that matter. In this case, if you have changed TJMax to 95 Deg in Everest and CoreTemp, all three programs should be reading within a degree of each other. Idle may be a different story.
http://img504.imageshack.us/img504/892/allthree2wx9.png
http://img504.imageshack.us/img504/7...ee2loadck3.png
Now I'm a fan of Everest, I use it every day. And coretemp has always been a standby. But to date, the soundest theory and best explanation and documentation come out of the RealTemp corner IMO.
Ok thx, so real temp is wrong to start with?
I wouldn't say that necessarily, it depends on who you ask. If anything, I guess I would be saying that Coretemp was "wrong" to begin with. The numbers for Tj Max are not formally documented, but if you read the setup and Theory and Calibration blurbs for Realtemp supporting using a 95TJMax instead of 100, it does make sense when you consider the validity of the testing done.
If you read on and follow the calibration steps for Realtemp, it's not really wrong until you see that your processor(in this case our Q66's) is significantly more than 6-7deg higher than your ambient with solid air cooling like an TRUE. Then you should make a small -ve offset to have it read accurately at IDLE. This should guarantee that it reads accurately over the entire range of temperature, and not just the higher load ranges.
In my case, running a fairly high end Water setup, my temps were 3-4DegC. higer than ambient at 1.6GHz/1.1v, which seemed reasonable to me. Hence I run no offset.
This all assumes that I am understanding the theory correctly. Maybe webb can step in and verify.
Thanks a lot.
Yesterday was a very hot day here. Today I'm much happier with the temps...
3.6Ghz / 1.38v / 56º Full Load...
I just bought this processor yesterday, big improvement over my old e6420.
You know what i say ...... you should research a lot better instead just staing things you find on the web about batches can and can't do as many ppl don't have the Knowledge and cooling to get 3.6ghz stable ... so automatically they say IT CAN'T do it, guess bad VID or whatever story they come up with.Quote:
Originally Posted by Loser777
Many can do .. no i will state THAT MOST Q6660's can do 3.6ghz no matter what batch it is ... all it matters is on Voltage needed, motheboard used and USER experience to try this or that in the bios.
Now i am not going to sit here and argue with you about what batches and what VID's do better as like i stated MOST can/will do 3.6ghz under the right conditions, cooling, users paitents to try this or that in Bios and so on.
Here is another recently bought L805 batch number (LINK BELOW) ... whether he is 100% stable in Orthos and Linpack is another question, but look at his Volts 9 x 400 1.38-1.39 which i have done that, but i wanted higher FSB .. So i chose 8 x 450 and to be 100% linpack and Orthos stable (11hrs 5min before i stopped Orthos) i had to use 1.43 in Bios.
http://www.xtremesystems.org/forums/...postcount=3312
I'm perfectly sure all Q6600s can do 3.6GHz, my point is is that there is a larger percentage of them that cannot do it COMFORTABLY... there is a point on every chip where the scaling is so bad that it's not worth it to pursue higher clocks.
I would say anything over 1.5v not very "COMFORTABLE".
There were lots of threads about people buying L80x batches and getting absolute crap chips that needed insane voltage for 3.6GHz up until L808.
There is no 100% again, as there are good chips in generally bad batches and horrible chips in generally good batches.
What I am saying is based off of personal experience and the experiences of others here... And I can assure you that I am VERY patient in BIOS. I test out every variable possible to narrow it down to the chip.
I think that you need to read some more about the experiences of Q6600 users here before citing lack of patience, experience, or adequate cooling as the reason people can't get 3.6GHz. I bet my first chip could do 3.6GHz @ 1.55v... would I be willing to run that even if I had water? No.
If you want examples of bad chips that would refuse to clock, look here:
http://www.xtremesystems.org/forums/...d.php?t=196981
http://www.xtremesystems.org/forums/...d.php?t=196648
Me neither.
Just wondering what do you guys think about a "safe" voltage to run at 24/7. My Q600@3.6Ghz temps are 50-50-45-45 at idle and 70-75 at full load. I heard some Q6600 degraded after a period of time so i'm just curious to know if I'm safe with my current voltage @ 1.44V in BIOS.
You're fine. Most say stay under 1.5-1.55v.
The reason I decided TjMax=95C for the Q6600 - G0 was because of what I saw during real world testing. Here's a couple of pictures I took 2 seconds apart.
http://img172.imageshack.us/img172/3...eter51cxy4.png
http://img111.imageshack.us/img111/8480/snapshotoc3.jpg
Every other program has decided to assume TjMax=100C. There is no documentation from Intel to support that number and from what I saw during testing, that number is certainly debatable. This test was on my brand new Quad so I didn't push it as far as I would have liked and trying to get some good pics while juggling a camera, a temp gun and 4 cores without a heatsink wasn't easy.
CoreTemp and Everest using TjMax=105C for 45nm seems to be dead wrong but everyone still assumes that they are getting TjMax right for 65nm. Why?
Intel plans to give full disclosure on their 45nm sensors later this week at the IDF conference so I'm hoping that they also talk about TjMax for the previous 65nm generation to finally get this cleared up. If they have nothing to say about 65nm then next weekend I will pull off my heatsink and do a more thorough test.
Hey everyone,
This is a little off topic for you guys but I need some basic help. So I overclocked my Q6600 to 3GHz (did not adjust the voltage). Booted up into windows no problem. I then ran a couple of benchmarks and did a little gaming to make sure everything was running smoothly (Boy was I enjoying the extra juice:D). Later on in the evening I decided to go out so I left my system running prime95. Upon returning about 4 hours later I turned on my monitor and got no display (case fans and lights were all still on). So i cut the power and booted back up again and did not pass the POST. I switched my jumper and recovered the bios. I am running in stock conditions now.
Any idea what could cause the system to crash during prime95? Should I try upping the voltage to my CPU?:shrug:
Your overclock to 3.6GHz probably wasn't stable... upping the CPU voltage is the way to go if you want 3GHz.
Run Core Temp and check your VID also.
You may also try upping PLL and FSB.
Thanx again Loser. I'll give your suggestion a try, post again later.
The new 1.2000 VID chip arrived today, and so far it's a real Gem.
Temperatures should smarten up when the ambient goes down, it was 30degC. in here tonight. Bleck!
I'll let it run Prime over the next 24Hrs to make sure it's stable, but it's a darn good start.
http://img125.imageshack.us/img125/2...w32idlelh5.png
LOAD:
http://img125.imageshack.us/img125/4...w32loadpd7.png
Awesome, 1.2 VID is best isn't it?
Yea the lowest Vid, it should require even less volts then mine ( @ signature), cheers with such a rare chip
:D I used to trust Everest or CoreTemp against RealTemp about my Q6600's temp'...Quote:
Originally Posted by unclewebb
That is to say that in fact, my temp' are better than I though???
:D:up:
Lowest VID normally provides better OC or less vcore.Quote:
Originally Posted by CERO
Unfortunalty, it's not that simple.
A friend sends me a Q6600 G0, VID 1.2125v->Needed 1.5125v (Cpuz) to be stable @8x500.
My Q6600 G0, VID 1.2250v->Only needed 1.488v (CpuZ) 8x500...
BTW, With a P45, I only need 1.30~1.31v to run @8x450.
VID helps but Batch is quite important too.
(don't remember the batches... :p: )
Started playing with mine, needed to reinstall windows a priori to o/c.
Batch L801A203, VID 1.28750V currently @ 3.00GHz default voltage, orthos stable, temps in idle 39,40C and in full load 56,57C at 27C ambiental.
Hope more fun to come. :)
I think its time to get rid of my Quad or my motherboard or both. I used to run my Q6600@3.6Ghz 1.4750Vcore on Water (dedicated Thermochill/DDC/Apogee GTX loop) for 9 months and it was rock stable no matter what i did with it ant that is on an EVGA 680i with 8800GTX SLI. Suddenly I had some random freezing while "unrarring" large 10GB+ files or Encoding video while of course other tasks where running as well.
I immediately tried Orthos, Prime OCCT and the rest and of course there it was not stable. I checked the BIOS redid some settings but it seems my chip does not want to continue its career at 3.6Ghz but it works fine at stock.
Is it possible that I am experiencing chip degradation? Do chips degrade and only work at stock? I thought they either die or work at whatever you through at them.
Another explanation would be that my motherboard has degarded. What would you do? Should I grab another Q6600 since they are dirt cheap now or even a Q9550 (is it working on 680i?).
The real sad think is that only a few days ago I put an Asus P5E Premium with E8500 in my HTPC and it is doing 4.3Ghz on AIR! LOL U am so tempted to just put an 4870X2 in there and turn this to my gaming rig as well on a 52" HDTV.
Which BIOS are you running?
You need p32 to run 45nm chips. I'm not sure if that includes quads, but duals for sure.
What is the VID of your chip?
How long did it prime test at that freq. and voltage?
You should not have any sort of degradation me thinks.
My BIOS is P30 and my VID is 1.3125 according to Core Temp 0.99.2 at stock settings.
When I first set up this config 9 months ago I left Prime, Orthos and even Memtested thoroughly for 12+ hours at 1.4750 bios Vcore.
I had a much better experience overclocking my q66 after the flash to p32. I've never been able to get such good clocks from so little voltage. My X48 doesn't make sounds like it's gonna die like the 680i did, but it takes more voltage to stabilize the same overclock. I'll add that I totally skipped over the p31 bios and went straight to the p32.
Here's an example:
25Hrs Prime Stable, Initial test for stability. Ambient 25DegC.
http://img142.imageshack.us/img142/1...000v25hja6.png
Nice Cool Day Idle. Ambient 21DegC.
http://img507.imageshack.us/img507/5...2vcoreisr3.png
Load. Amient 21DegC.
http://img507.imageshack.us/img507/6...2vcoreltm6.png
Your core temps are lower than your ambient? Doesn't that, like, defy the laws of physics?Quote:
Originally Posted by Ol'Baditude
This is mine on its current all day every day setting:
http://database.he-computer.de/Bilder/41ghzprime.jpg
Not quite satisfied yet, I always wanted to get to 4,2Ghz for 24/7 use and I will. Just have to coax my DFI to run a tad more FSB, got 462 already but I need 467...
Oh yeah, it's an older L737 (don't remember the exact batch but can look it up), Vid 1.225V.
Does 4Ghz on water with around 1,45V too.. problem is FSB/the mobo. Some unresolved issues with DFIs X38 and Quad FSBs, no one got over 460 stable (while I could prime the CPU at 480 on another X38 board, will probably go even higher on P45).
3ghz @ 1.160 vcore
http://img292.imageshack.us/img292/5310/ttttyy4.jpg
In a word... No.Quote:
Originally Posted by DKK
If your freezing a CPU then your temps will be a lot lower then ambient.... :)
Well i have not posted in here for a page or so.... so i am using Bios 1201 now and all is well so far.. check my specs and template for what i have and running.. Orthos 11hrs before i shut her down and Linpack 1.4 stable .. 10 loops.
Yup, the answer was pretty much as simple as using DICE or LN2. But an undervolted chip cooled by a high end WC loop can hover right around the ambient room temperature.
What you state has nothing to do with your snapshot. I have no idea if you're using a high quality WC, but you're definitely not undervolting to get your OC clock. Even a high quality WC will not give your core temps below ambient. This is not an opinion but a fact.Quote:
Originally Posted by Ol'Baditude
You should have had an idea of my cooling from looking at my sig. It's staring you right in the face. And yes the chip is in fact undervolted if the VID is 1.200 and the Vcore after .05 droop is 1.184v That means I have to set BIOS voltage above VID at 1.225v to get the required ~1.175v that it takes to run the chip stable.Quote:
Originally Posted by DKK
Anyone who knows anything about water cooling should know that you can't go sub ambient without a chiller. Thx but you're stating the obvious.
At no time did I say that my idle temperatures were correct. I was just giving a screenshot of what I had so far. Those numbers were not calibrated in RealTemp which should have been obvious by the fact that they matched perfectly with Everest. Anyone who says that their temperature monitoring program is exactly right anyway is deluding themselves.
Anyone who's ever taken a snapshot of a temp monitoring program knows that:
a) the numbers fluctuate and it's anyones guess where they're going to end up at the instant you take the screen shot. In this case it was 32,28,28,28. They can and will fluxuate ~5 deg in either direction.
b)thermal sensors in a chip aren't always exactly stable, consistent or as good at reporting temperatures as people think.
It just so happened the Idle core temps happened to read a little low at the instant I took the screencap that's all.
I can see now that your original question was aimed at me in particular, but it was ignored due to the obvious reasons above, and because it came off as ignorant.
I addressed the obvious and general inaccuracies of your question by indicating sub ambient temperatures are of course not against the laws of physics.
And the point you should pay particular attention to is this: nobody posts a screen shot to show their idle temperatures, because they don't matter. People post two screenshots. One to show idle and one to show load so that people can get an idea of the two important things: The max temp, and the delta between idle and load.
I beg to differ - idle temps don't matter, ever. Delta is between load core temp and air (aircooling), water (watercooling) or, in some cases, between evap and core (I get around 0 C coretemp at -20 evap and approx 260W heat load, for example, hence Delta = 20 C).
Ok, that's one way to look at it. And you are right. Thank you.
Another would be to refer to the change btw. idle and load as the delta (since delta can be used quantify the amount of change in any situation). For me, monitoring this delta is merely for diagnosing problems with my setup(ie: HS/WB mount, TIM application, rad efficiency, etc.), and to show this when displaying screen caps to support my OC reports.
I can see your point. Rad efficiency is best measured with the delta between ambient and water temp, not CPU core temp though. Too many variables there, like mount, TIM used, pressure applied, flow etc.
You know why idle temps aren't of any importance to me? :D Because we DCers don't even know a state called "idle". You might wanna have a go at it, too - head over to the distributed computing forums and pick your addiction (I'd favor WCG :wasntme:)
This is called active member acquisition, you know :D
Bah thats it i moving to the North pole anyone wanna join me as that is a COLD/FREEZING place.. heck i bet we won't need anything on the CPU at all to run it... :)
Hehe, I was wondering what DCers meant. I fold for team32 already and have a dedicated rig for that, and yes, It does not have an idle state. Sry, no room for more in my small one bedroom condo.Quote:
Originally Posted by jcool
Also, yes. When I said rad efficiency I meant more along the lines that it doesn't take too much to dissipate the heat of an idle processor, and if you just look at that, you're not even coming close to getting the whole story. But if my delta between CPU idle and load temps skyrocket after I turn on prime, and I've verified the function and performance of the rest of the loop, it might be time to change part of my WC setup. In particular the Rad and or Fans, or just make adjustments to my fan controller. I think that in this case, the way to make sure if your rad was in need of replacement would be to do what you said, and check the efficiency by monitoring the Ambient air and water temperatures.
If my idle temps show -1C (core sensors won't go lower than that..) and temps move at all as soon as I start up prime, my evap mount is bad :D
Phase change can be so simple sometimes :ROTF:
Good to see you're folding already :up:
You have pictures of that setup?
Not really... I am waiting on some final WC parts to finish the rig, for now it's just running, cabling is a mess etc.
You can catch a glimpse at it here, note these pics were taken while the unit was down, hence the "emergency CPU waterblock" :D
Hi all...
I have just joined the q6600 club and wanted to report/share my findings, especially after reading about the effect on of low/high VIDs on oc. Namely, when i first run core temp and saw that my q6600 is g0 with the VID of 1.325 I said: ":banana::banana::banana::banana: i am screwed". And I was right. When I tried to boot at 3ghz with everything at stock, pc did not even boot. I said ok, this is some serious crap. But then I went in to bios and changed some voltages (had 939 rig before so i really did not know what i have to change :D) and eventually ended up booting and priming stable at 3.2ghz at 1.35v. Then I lowered cpu voltage to stock and primed again...stable :). So I am running at stock voltage at 3.2ghz at 60/60/54/55°c under load (according to core temp) on p5q using artic freezer 7 again having hope that I did not get a complete dud.
What batch?
hmm...i do not know if this is the correct No., but under fpo it says:Quote:
Originally Posted by Loser777
L807A422 (SLACR, Malay, pack date 5/9/2008). Is this good?
will go for 3.4ghz@stock and report back.
3.4 @ stock might be a little tough for you. I can boot my 1.2000VID at 3.4GHz at stock settings and run applications , but forget about Prime stability.
When you say you are priming stable, how long are you priming for?
Here's a clip from last night.
http://img365.imageshack.us/img365/6...hz3dm06up5.png
It all comes down to scaling... 3.4 will probably need 1.4v+
I was just going to post that ... yeah it may need 1.40-1.42v to get 3.4ghz on that VID ... another issue will be cooling as at 1.40-1.42v the temps will rise to at least 65C-70C full load on his cooling, priming in the first hour to get MAX temps.Quote:
Originally Posted by Loser777
So maybe MAX on his rig will be 3.2-3.4ghz with his cooling setup, but more likely 3.2ghz cause of cooling issues maybe all that will be 24/7 on his rig, but you never know... good air cooling is a must when hitting 1.42-1.43v on the Q6600's.
I am priming for 3 hours. I know, it may not enough for many , but it is ok for me as in real life i am not using my comp for anything heavy, except for gaming which still do not stress cpu as much. So my benchmark is to be 3h prime stable. ;)Quote:
Originally Posted by Ol'Baditude
Anyway, I went for 3.4 and core 3 failed within a minute. raised voltage to 1.35 and it failed after 3 minutes. :( so, no 3.6 for me i guess. will try to raise voltages more but, i do not want to go over 70°c. is there anything else beside cpu voltage that i should raise?
I am still new to 755 platform so excuse me if asking stupid questions.
edit: I set it for 1.360v and now it says 1.352 in cpu-z/ai probe and it is been stable for 30mins, temps are 66/66/59/59
Update: It seems prime stable at 1.36 (cpu-z/ai probe) at 3.4ghz which I guess is not to shabby for 1.325 VID chip. :) Temps are 69/69/62/62 (core temp).
Are those ok for 24/7 use?
There's nothing wrong with 1.325VID chips. Mine made it to 3.6 at 1.4V, just the same as my 1.2625VID chip. It would not, however, scale to 3.8GHz, and would not give me the nice 3.0GHz at 1.176v like the 1.2625 would. The scaling at the mid level frequencies was good, the at near the end points it was not as good a chip.
Hy ppl, I'd like some advice here.
I'm trying to get stable on air (TR Ultra 120 Extreme lapped and the Q6600 lapped too) above 3.6GHz.
With 3.7GHz (463x8) I can Prime almost 2 hours then system reboots. With 3.8GHz (425x9) it instantaneously reboots when Prime starts.
Temps on the Q6600 @ 3.7GHz are ~75ºC @ load so I think it's not the problem. I'm almost sure it's my psu (Enemax 535W) but can someone else confirm what else could be and what other advanced adjustments I should use?
For 3.7 I'm using about 1.4750 vcore and 1.5250 vcore for 3.8 (both set on bios) with very little vdroop.
Have you tried scaling back to 1 or 2 sticks of ram?
Also, whats the VID of your chip?
At some point you have to get temps down if you want to OC higher. 75C at >3,6Ghz is very high and very likely to cause instabilities.
Depends on your CPU, but general rule is, the higher your frequency, the lower your CPU temp needs to be.
Rule of thumb for Kentsfields, <70C for 3,6Ghz, <65 for 3,8 and <60 for around 4Ghz. Assuming your chip can handle the Mhz in the first place.
Xeons can usually take a few degrees more, but in the end it comes down to the individual chip (as always).
Bottom line is, 75C is probably too high for that kind of clock to be stable, sorry.
VID 1.250v. Actually, I tried booting with only two sticks but system couldn't post at all (very strange).
Well, it tough to tell if that's the PSU or not, but it definitely couldn't hurt to upgrade to a more powerful unit, perhaps a corsair or PCP&C in the 6-700 range. Also, There are plenty of Silverstone units in the Strider or Zeus series that won't cost alot of money and will provide clean power for the life of your machine.
As for the overclock itself, what are your FSB and MCH voltages? RAM frequency and timings etc?
Do you have a listing of BIOS parameters which includes the parameters that you have changed? I work with the award BIOS on my Gigabyte X48T, so I'm not too familliar with what you are dealing with.
I could do 3.83GHz stable on my 680i, so I'm sure you could do it on your board, assuming your chip will handle it. I was under water though.
Those temperatures are more than I would like to see however, and that is definitely not helping your cause.
Quote:
Originally Posted by Viper.Br
I guess temps are not the real problem here (but it is one thing to consider) because in the past I could prime with temps around 80~85ºC for more then one hour (I know it's crazy but I tested anyway :D). And actually 75ºC was the highest temp on the hottest core. the other ones were around 70~73º.
I have all settings back in my home, but as far as I remember I was using:
JumperFree Configuration Settings
AI Overclocking: Manual
CPU Ratio Setting: 8 (or 9 for 3.8GHz)
FSB Frequency: 463 (or 425 for 3.8GHz)
PCI-E Frequency: 103
DRAM Frequency: DDR2-??? (don't remember exactly but I guess it was 1:1 or 5:4 divider, above DDR2-1100)
DRAM Timing Control: Manual
CAS# Latency: 5
RAS# to CAS# Delay: 5
RAS# Pre Time: 5
RAS# ACT Time: 15
TWR: Auto
TRFC: Auto
TWTR: Auto
TRRD:Auto
TRTP: Auto
DRAM Static Read Control: Enabled
Transaction Booster: Disabled
CPU Voltage: 1.4750v (or 1.5250v for 3.8GHz)
CPU PLL Voltage: 1.60
FSB Termination Voltage: 1.40
Dram Voltage: 2.25
NB Voltage: 1.55
SB Voltage: 1.05
Clock Over-Charging Voltage: 0.8
Load-Line Calibration: Enabled
CPU GTL Voltage Reference: 0.63x
NB GTL Voltage Reference: 0.61x
CPU Spread Spectrum: Disabled
PCIE Spread Spectrum: Disabled
CPU Clock Skew: Auto
NB Clock Skew: Auto
FSB 460ish might be getting up there for a Q6600. I know my Chip doesn't like it much, and I would probably have to do some serious tweaking to get it stable.
So I assume you are doing 425x9 for 3.8Ghz?
My BIOS shows two GTLREF values. GTLREF1 which defaults to .67 and GTLREF2 which defaults to .63 I believe. I can only assume that 1 corresponds to the CPU GTL and 2 corresponds to the NB GTL.
As far as I've been able to understand, the value for the CPU GTL ofg a Q6600 should be 67%. Perhaps try upping that one.
When I get up into the higher FSB's it helps to tweak these values and find a good combination. Anything past 450FSB and it helps me to up the GTLREF1(CPU GTL) +3%.
I'm not saying this will work for you, but it couldn't hurt to play around with these a little bit.
Are you sure it's not your RAM holding you back??? What are the frequency and timing values when you try your 3.8GHz overclock?
Nice I just go and installed X3210 (Q6400 or , 8x400)
http://img58.imageshack.us/img58/586...6400pc9.th.jpg
In my experience there is a point with Q6600 where you have to keep the temperatures low to gain more mhz... normally at the point where you have to give it a bit of a big bump in vcore to obtain a higher clock..
On my L737B192 I can run 8 hours Prime stable @ 3.745ghz @ 1.424v however to get to 3.825ghz 8 hours prime stable requires an extra .072v (1.496v) for 80mhz compaired to 3.825ghz prime stable - 4.023ghz primestable which needs extra 0.040v (1.536v) for 198mhz..
So at up to the 3.745ghz clock I can prime at temps of up to 75c on the cores (never let it go higher), where after this point and the big jump in vcore if any core hits 63/64c then prime 99% of the time will fail..
Yeah I have noticed that too. If I could keep my water temps at 20Celcius @ 247 I could run this chip easily at 4 GHz.Quote:
Originally Posted by Pneumonic
I have a problem with Prime 95, now I know u all will probally just say its not stable, but it is.
I can run at 4.0GHz with 1.536v (idle) and 1.518v (load) and I can run 3D Mark 06, wPrime 1024m and super PI etc, its stable. Then as soon as I start Prime 95 in one sec it crashes. I get same problem for anything over 3.65GHz. When I was at 3.7GHz and it happened I uped my volts a lot and it still happened. If it was stable enough to run wPrime for 5min +, y can't it even run Prime 95 for 1 sec. Also I know my temps aren't the problem, because I watch them when I run wPrime etc and if it was my mobo would tell me before shutting down.
So WTF is going on?
whats the max voltage you guys recomend to use for 24/7 usage on water for a Q6600 G0?
I'd say 1,5V.. 1,55V if you're bold and can keep your temps low.
Really depends on the chip, mine puts out 260W at 1,45V lol.. so no go (and no sense) with 1,5V+ lol.
i just got my new gigabyte board.
so far its about 15 minutes stable.
http://img515.imageshack.us/img515/103/newboardgj6.jpg
http://img504.imageshack.us/img504/6537/43630705fs6.jpg
Both of my 1.2125v VID chips couldn't touch 3.7GHz... nice result.
What FPO/Batch?
L745A i think
nothing special, but ill post never the less :D, havnt really tried to overclock it to its maximum, but its fast enough for what i need it to do atm :D
http://i276.photobucket.com/albums/k...ra69/G0cpu.jpg
not bad. i guess any thing is better than stock
Wheels, nice clock and temps. I assume you're using 95 as TJMax for Coretemp? What was the ambient temperature at the time of that screenshot? Is the CPU the only component in that loop?
My temperatures after about an hour of Prime are very similar to yours, except with the overclock @ 3.6GHz. But I am running a dual rad, not a triple. I imagine if I ran a triple, I would be mid 40's at 3.8Ghz.
It's either not stable or it's your PSU. My system was getting all sorts of crap until I flipped my Corsair 650 over, the intake fan was being blocked. I got the same behavior... clocks would run stable for hours on end but crash the instant I ran full load.Quote:
Originally Posted by fatguy1992
However, there is such a thing as a non-P95 stable overclock... sometimes you can run stuff windows and PI stable, but P95 will crash instantly.
THX, ok its probally my PSU, dam cheap 650watt, but i have a TT 650watt diver bay PSU for my 8800GT.
I'll do some more testing soon, as in what can it run and for how long.
I would image that its the PSU or the board cant keep the voltages stable enough
U have a Q6600 stable for 8hr Prime 95 so what PSU do u have?Quote:
Originally Posted by Pneumonic
Saying "whatever" to failed o/c stress testsQuote:
Originally Posted by fatguy1992
its a cpu only loop with dual mcr 320's and no i have been using the 100c TJmax for coretemp should i use the 95c TJMax for core temp. its has been about 65f here lately so about 18c to 24cQuote:
Originally Posted by Ol'Baditude
I think you've got the point. First, I'm getting a new PSU (Extream Superflower Aurora 700W) this weekend. Then I'll try to work on the board cooling and finally work with different RAM dividers.Quote:
Originally Posted by Ol'Baditude
Right now I'm trying to find the "limits" of my Q6600 in terms of clock. I know my board can post up to 480~489 fsb, and I'm confortable with 3.6GHz for daily use, but I'd like to know the max clock Prime stable on air so 425x9 for 3.8GHz is my goal right now.
I just found this screen shot with 3.7GHz:
http://farm4.static.flickr.com/3098/...c6f35381_o.png
Too shame I didn't open a CPU-z window with memory clocks, but I'm sure it's 1100MHz+ with 5-5-5-18.
That's exactly what I'm talking about here. Above ~465 fsb whenever I start Prime95 system crashes in one sec.Quote:
Originally Posted by fatguy1992
I can also run 4.1GHz (on air):
http://farm4.static.flickr.com/3286/...d8270f95_o.png
THX I read the thread, but my prob is a bit different, I don't get an error just a crash. Anyway I don't think any real world situation would use 100% of a 4.0GHz Q6600's processing power.Quote:
Originally Posted by Alcibiades
I've found issues with the memory dividers causing stability issues with my overclock. I managed to get 3.8 prime stable, but when I game it will still crash on me. Dropped my memory to 1:1 and it's always fine. Running 4GB G.Skill DDR2-1100 which is memtest stable, so it's not the ram, but the divider on the board.
I've tried the latest bioses for P5K-E but the 0602 seems to be the best still.
http://3dxtreme.net/other/Q6600%20L7...-1.44-11hr.jpg
I get the exact same thing unless my load temps are under 63c on my cores... anything over is instant crash.. I know iceman had similar issues on his 4.1ghz 8 hour primestable clock and a number of other people at OCUK have reported similar problems..Quote:
Originally Posted by fatguy1992
Hmm, well I watched my temps and they did go over 60C :D Also before I got my IFX-14 I had the stock cooler and I had the thermalpaste on wrong and it was running at 60C idle.
60c at idle ,ouch...
LOL i'd never put a HS on before and the stock cooler is pretty S***.
I run my fans off a seperate power supply as they are 24v and eat 80watts each and whilst running some prime runs @ 3.9ghz to test the new setup the other day i forgot to turn them up (dont ask me how i forgot they are bloody loud enough at full voltage to notice they arent on) :D 102c was the only thing i saw coretemp flash before the machine shut down.. luckly no damage
:shocked::shocked::shocked:Quote:
Originally Posted by Pneumonic
Your fans eat more than my phase unit :rofl: :ROTF::ROTF::ROTF:
:rolleyes:
Have you checked whether your ears are still clinging onto your head recently? :D
Quote:
Originally Posted by jcool
They probably do :) but well is more experiment than day-to-day... I really want 4.1ghz 8 hours prime stable on air... and the old extreme air setup just wasnt man enough :)
some info here: http://www.xtremesystems.org/forums/...d.php?t=197519
and a whole build / clocking log here: http://forums.overclockers.co.uk/sho...php?t=17888722
I guest you might want to try some other bios versionQuote:
Originally Posted by dnottis
Dropping to 8x mult helped me pass my wall. Keeping memory 1:1 also fixed the stability issues with ram dividers. Sucks to have DDR2-8800 and only run 960, but at least it's stable at 3.8Ghz now.
http://3dxtreme.net/other/Q6600%20L7....primedone.jpg
Can you post your bios settings for that FSB on that board? What bios are you on?Quote:
Originally Posted by dnottis
nice i have maybe i lll start trying to 3.9or 4.0 stable instead of trying to get 3.6 stable
so far so good at 3.8 voltage is way high ill fix that later but look at my temps.
http://img514.imageshack.us/img514/1123/ohsnapak8.jpg
nvm it crashed...
I got same prob here for 3.7/3.8, 5 mins then crash.
i didnt have this problem my old max extreme or ip35
I'll post pics of my bios tonight.Quote:
Originally Posted by kup
Using 0602, Asus P5K-E/Wifi