Seeing as most ppl seem to run into trouble either with cooling or mobo-issues, this should be a good setup for OC I thought.
Ok, the question: leaving multiplier and ram settings alone, upping volts by 10 or 15%, where would I aim the FSB?
If I am not mistaken, FSB=350 gives 3,15Ghz. 400 would give roughly 3,8 GHz. What will the setup take without additional bridge-cooling? Anyone out there with experience of clocking this CPU on that mobo?
Thanks!
As a sidenote, FSB=400 doesn't even exceed the maximum bus-speed for the mobo, and liquid cooling should fix the upped voltage. That would mean that 3,6 GHz even sounds reasonable. One test i read upped the FSB on a quad to 460, stable.
However, ppl here with supposedly descent rigs have had trouble clocking significantly more modest than this would suggest. It seems a bit too... well, random is the word I guess.
From what you've spec'd sounds good - although don't "expect" a final outcome - you could equally get a crap ocing chip just like a good one, regardless of voltage and cooling.
There's lots of Q6600 ocing and spec threads in the hardware forum - try reading a few of those to see if they have some answers too :)
Thought I'd just follow up on my post. Did the bold (stupid?) move of just jumping straight to 333 MHz FSB (gives 3,00 GHz), and upped the VCore to 1,300 in one go. Works like a charm, np whatsoever. Raised the score in 3dmarks06 by over 1k. Max temp on any core I noticed is 67, generaly loads 60-65 and thats with ambient temps @ 25 degrees celsius.
Tried 350 MHz FSB and had one crash after a couple of hours of gaming, prolly lack of voltage. Since the watercooling is internal (sucks chassi air through the radiator) and the chassi also holds a hot GTX 8800, Im gonna install two new 12cm case-fans today to see if i can up the VCore a bit more and get stable at higher frequencies.
Anyway, the message being, with a descent mobo and cooling, you can get an 20-25% more Hz just by spending 3 minutes in the BIOS. Gotta love it!!
Simply Amazing. I new to most of this, but for it to be written so that a laymen such as myself can learn, much appreciated. I have a question that may have been answered after the first few pages, when I could read no longer! lol. I am running the 6600, 780i, with 8gb of g.skill pc2-8000. I just got done putting it together(my first system), and have a question about the memory. With that amount of memory, is there a need to increase any of the voltages, or does that not matter whatsoever?
I am aware of some 680i chipsets that required a mem voltage boost with just 4gb ram to even post without overclocking, something of which i hope was addressed in the 780i chipset.
However if your board is booting fine, then id leave the voltages if you intend to leave them at stock speed, if you intend to overclock the ram also, then you may find you will need to increase the ram voltage, id personally not recommend increasing the voltage over any more than .1 of a volt max, than that of what they where running at fine before.
Mine will run @ stock on 1.8v fine, however to clock them a bit I found i had to increase to 1.9v
Some people have made larger increases than this, however im more on the side of caution.
alright, sounds good. so what i conclude after reading, is that with my system, these would be good specs to start out at? And yes, I hope to OC the memory, so what do you guys think I'll need to change to do that? (please comment on if i'm correct or if you have any suggestions of different/better suggestions):
system: asus P5N-T Deux 780i
Q6600
8gb PC2-8000
Nvidia Quadro 4600, it' for rendering
2x500gb RAID1
OC specs:
fsb to 333mhz
vCore Voltage to 1.4625V
CPU FSB Voltage from 1.2V to 1.4V
nForce SPP and MCP Voltages from 1.2V and 1.5V to 1.45V and 1.55V
I am not sure how much you have tested already but for that OC on the CPU the VCore seems way too high. I have mine now at 366 MHz FSB with the VCore well below 1,4v. Doesnt it run very hot with that voltage? At 1,4 I could run it at 380 MHz but i stepped it down because it ran hotter then I like. Personally, I have set a limit that no core of the CPU shall exceed 65 degrees after 1 hour of 100% stress.
For temp monitoring, the easiest to use IMO is coretemp (free download). You can minimize it to the toolbar and it takes up like 2 megs if ram while operating. Just check so that the version you get is compatible to your OS. Since you have 8 gigs of ram I am gessing you have Vista 64-bit, dunno if coretemp is availible for that. Also, everest ultimate edtition (Company: Lavalys) has a function of monitoring ALL temps (including GPU, drives, Mobo etc). I am gessing that your system runs really hot under stress (unless you have some killer cooling systems). Hotter temps decrease life span and the temp of the q6600 shall never exceed 71 degrees (if I remember correctly).
Everest also has a stability test. Otherwise, the most popular seems to be Prime95 (also free download) but be sure to run it in four instances (all four cores). There are links to that procedure above in this forum.
Please get back with some temp-results when u have it. I have a Cooler Master Stacker 831 with 4 120mm case-fans and watercooling on the CPU. So if you can get descent temps with that voltage I would be happy to know the secret.
My BIOS allows for all other voltages on auto and I havent clocked the RAM, so I am sorry I can't comment on the other stuff.
As a clarification to the above: Though this OC-test ended with the VCore volts you provided, please notice the mother of all aircoolers they are using. So be careful. Not knowing your cooling, my guess with the suggestions you provided above is that u will have no problem getting the system to work stable, but a big problem with CPU-heat. If your aim is "only" 333 MHz FSB (3 Ghz), begin with much lower VCore-volts. I know more than a few people who runs this CPU at 333MHz FSB at default VCore volts. Practical suggestion: VCore: 1,400 and CPU FSB voltage:1,3 is a good starting point. Than increase FSB a little at a time.
In the end, you have to decide for yourself if you want to do a max OC, set a temp maximum you think is reasonable, a max FSB you think is reasonable, or some other limit. If the goal is 333 MHz FSB, start at the above volts, increase FSB until u reach the goal, then lower the volts until you reach the level at which the CPU is no longer stable. The (obviously) take it up one notch again. If u cant reach 333 MHz FSB at the above volts, then increase them, while keeping track on the temps under load. I think the volts will be enough though.
/J
awesome, that is exactly what i was looking for. to clear things up, i haven't actually tried any of these settings yet, i to run it by you guys first, and i am glad i did. i will not run it at that high of volts, long term durability is more important to me.
i thought it was one of the best, but i am sorta new to most of this. so i guess if it is not such a great one, maybe i'll try to get the one tey had on here and sell mine or something. and then i'm running (1) 200mm fan, (4) 120mm fans (Antec 900 case).
as for the 8gb of ram, i am running XP Pro Corp 64. i needed the ram to help out with rendering. and coretemp worked perfectly with this OS and monitors all 4 cores perfectly.
i think this is what i'm about to try:
360mhz fsb
vcore- 1.4v
cpu fsb-1.3v
maybe don't mess with any of the spp and mcp voltages?
one last thing, i just wanted to make sure that when you initially suggested CPU FSB of 1.3v, that should be the same thing as the settng of "CPU VTT Voltage" that I am changing on my board right?
well, just got done with a third test. hope i'm going in the right direction with this stuff?
360mhz FSB
1.375 vcore
1.2 cpu vtt voltage
average temp was 51 with a max of 57.
i guess i should keep going lower. but the cpu vtt v is as low as it can go, before it goes back to auto? should i keep going lower with the vcore?
Seems like nice results. I too managed to clear a few hours of prime95 at 366/1,375 settings (didn't bother more), though my temps where a bit higher. However, some applications still crashed (Crysis for example), so I am at 360/1,380 right now and I intend to stay there. So just because you clear a certain amont of hours in prime95 doesn't mean it will work out perfectly. How long you should run Prime95 in order to claim "stability" is a matter of taste, some say 24 hour, others say a lot less.
Sure you can try to lower your vcore but not with those temps, no need IMO (You are running 4 instances of Prime95, one per core???). Just stay where u are and enjoy the extra stability. Try to fiddle around a bit with different torture tests on different cores. Do the "In-place large FFT:s" on all 4 cores to max temps. Otherwise, I prefer to do "blended" on 1 core, and the above on the other 3.
My recommendation since you promote stability and durability, don't change anything. You have a nice overclock of around 35% and got a frequency exceeding the really expensive quads at stock, so why bother.
Some computer-wiz may give you a detailed description of why core 1 and 2 are running about 5 degrees hotter than 3 and 4. I cannot, I can just say it is normal (btw, the cores are actually named 0,1,2,3). Your temps seems almost to good to be true (at least to me), please confirm that u get those after a few hours of 4-instance prime. If that is the case, congrats, you have gotten a much better OC-ing set than I have.
/J
thanks again. i didn't get your message until just now, but i have been doing a torture test on some new voltages for the past 3 hours. but i want to make sure that i am running prime95 completely right. i dont know which version of prime95 you are running, but i am running 25.5, and it is supposed to automatically test all 4 cores, and when i start it, it asks me how many worker threads i want to run, so i put in 4. run the task manager, and sure enough, all 4 are running 100%.
anyways, the numbers i gave you before were from everest using its stability test like you suggested. right now i took a big leap down in vcore just to see what would happen. i put it at 1.3375. i have been running the prime95for a few hours on small FFTs, and monitoring the temps with coretemp. i am sitting at 58 58 52 52.
let me know what you think? should i put it back up to where it was at? or retest that voltage with prime95? let me know, later jerk;)
alright, well, i guess i'm where i'm going to stay at unless tonights torture tests prove otherwise, or you think it would be wise to up the voltage to where it was.
360fsb
1.3375 vcore
1.2 cpu vtt volatage
after testing a 4 hour cycle with prime95, temps sat at ahigh of 58 58 54 54
aftr 4 hour cycle with everest, it was 51 51 47 47
so these temps sound good. i'll doa 24 hour test with prime 95, but i thinki'm good to go. now i have to look at overclocking memory!
Sounds lika absolutely fantastic temps and voltage on that overclock. Gratz. You could probably get a lot more out of that CPU if you wanted to, even though the limit seems to come quite suddenly when you reach it. I have the same figures u have exept Vcore is 1,38 and temps 63, 63, 58, 58 (watercooled:( ). Again, I clear Prime95 with lower voltage but not Crysis while gaming.
If all apps works with that voltage there is no need to increase it. If something crashes along the road, a small increase is often enough.
I haven´t worked with OC:ing memory that much but a few rules of thumb is that you cant clock memory nearly as much as cpu:s, and you have to more careful with voltage. I would personally never exceed a +10% increase in neither speed nor voltage. There is also a point trying to get the memory to work at the same frequency as the cpu. That means that it could be a point in dropping the frequency on the memory of a 800 Mhz DDR2 (actual frequency of 400) memory to 720 (360) on a set-up that has a 360 FSB cpu. Or, setting the cpu/ram rate to 7/8 on a rig with a 350 mhz FSB cpu with the same memory. Lower frequencies on the memory generally enables lower latencies as well. I am not sure to what point the trade-offs are worth it though.
Also, see the RAM-manufacturers website. With a little luck, you will see that your ram has been tested at lower latencies than your BIOS has set it to. Be aware though that higher speed often equals higher latencies, so the tested latencies may not be accurat once you have OC:d the RAM. If you are highly dependent on RAM, check out a program called "Memtest". It's a bit of a hassle since u have to boot it from disk (standalone program). Personally never tried it but from what I can gather, it is the best memorytester. Both "CPU-Z" and "memset" will reveal the speed and latency setting of your current RAM. "SiSoft Sandra" can benchmark your memory. You get a couple of diagrams comparing your setup to other typical rigs. This is more to see if anything is really wrong, plus to compare to your own results after you have OC:d.
Though your GPU and mine isn't really comparable, it would be fun to see what u would score in 3DMark06 (free version available). Post the score if you try it will you?
/J
first, i'd deff love to run that 3dmark06, i'll do that sometime today.
one last question, it seems that all the warning signs people spoke of as fas as an unstable system go, were (1) not POSTing, (2) prime95 actually telling you there was an error, (3) not booting up or not booting windows, or maybe some other drastic and obvious way that it is unstable. But what about freezing up? Yesterday while I was starting prime95, i started cpu-z to see something, and bam, it froze up. Then this morning I woke up and prime95 had went for quite sometime, but the comp was frozen. so i just restart it, and it seems like no problem.
i would assume this is considered a symptom of an unstable computer also right?
also, as for the RAM, yes, I would much rather lower the latencies of the RAM rather than OC it. It is DDR2 1000 (PC2-8000). i think that i would much rather try to lower the latencies than oc it. plus, i hear it yields better performance in the trade off.
hey, i just ran the test to see if my computer was compatible with the 3Dmark06, it says that it's not due to the fact that
(1) I have windows xp pro 64bit running, and it needs vista. and
(2) " Your display card is significantly more powerful in raw performance
Your display card fails the following requirements:
Pixel Shader 4.0, you have 3.0"
ok, so now i'm getting worried. i run prime last night, and it ends up freezing at 20hours. so i think ok, bump it up from the 1.375 and 1.2cpu that i had it at, to 1.3875 and 1.3cpu. i start prime95 and walk away for a while. i come back in like a couple hours and find that it had been sitting at 66-67 for a while. i dont know what happened? i only bumped it up a little, and i've had it there before, that was the strange thing. and now, even if i bump it down to like 1.375/1.25 like i had it all last night, it jumps staight up to like 65-66, rather than like 58-59 which it sat at for liek 20 hrs.
so what do you think happened? think i screwed something up permanently? ****, i hope not...
Dunno what happened dude, but those temps seem more like the ones I have. Cant say why it is hotter now than before. Have u changed anything (like changed program for monitoring temps or voltage)?
I really have no idea what happened but for that overclock/volts on air, those temps are pretty normal. Strange you got others before though. Be careful to run the same processes in Prime, as different processes on all 4 cores can easily give a 5 degree difference. Apart from that, I am clueless.
I have moved down even more and am at 350fsb now, with lower voltage. That gives me much nicer temps and it its still faster than a $1000 quad so...
Another thing to put the fiddling in perspective: Having the window a bit open in my room while I game has the same impact on cpu-temps as an extra fan for the water-cooling :), and its cheaper/easier.
I am at 37,37,40,43 idle and 58,58,62,63 under load right now while stable in everything I throw at it @ 3,15 GHz so I will officially stop fiddling now.
alright, that sounds good. guess maybe i'll have to lower mine too. it sucks, but at least i'm average now. i swear there were so many tests that it did great, then BAM! but oh well. anyways, thank for the help, i appreciate it. hows that watercooling doing for ya? i'm so afraid of it leaking??!!
I've been fiddling with my b3 q6600 for a couple weeks now (bought it last summer, only recently started oc'ing it). It's in an EVGA 680i LT board and I've been having a couple of problems. First, it's difficult to get it past ~2.8ghz in a functional state (Boot, etc.) Second, any oc at all and it will run full load for about 8 hours and then the system will freeze. I run seti@home on this comp, and the decrease in workunit computation time is quite noticeable at even more modest oc's (currently 2.66). This 2.8ghz wall, and the system freezing are frustrating. Any advice on other's experiences with this stepping/board combo would be greatly appreciated
My specs
Intel Q6600 B3
Clocked at 2.66 temps:
Cores: ~54C
CPUTIN: ~44C
VCore: 1.240v
I did the pencil mod on the mb, so no vDroop
Thermaltake V1 Cooler in Antec P180 case (cpu fan exhausts straight up into another (120mm) case exhaust fan on the top of the case
FSB voltage set to auto (believe it's 1.2v)
temps, voltages from CPUID Hardware Monitor (Speedfan reports ~12C cooler, going with higher reported temps) CPU-Z and Core Temp
I think I found the culprit! After reading this review at xbitlabs, my TT V1 is apparently incapable of cooling oc'd quad core procs. I wish I'd done this research before I bought it! Fortunately, it's worked fine for the past 10 months of running stock clocks and I believe I got my moneys worth out of it, however, looks like I'll need a new cooler.
Aanaddor, I have very simmilar set up, XFX nForce 680i LT SLI, Q6600, 4Gbs Corsair DDR2 @ 800 MHz, BFG 8800GT 512.
Upped my FSB to 1334 Mhz
unlinked Mem Clock
Mem Bus set@ 814Ghz ,shows 800.4 Mhz
CPU volts set @ 1.42 v shows 1.36v
Mem volts set 1.9
FSB Volts set 1.4
PassMark Performance test 1475 @ above settings
Keep an eye on your temps CPU is up to 68 under load.
Looking for better cooler than my bogo intel one any sujestions welcome.
Comments 151 to 175 of 175
As a sidenote, FSB=400 doesn't even exceed the maximum bus-speed for the mobo, and liquid cooling should fix the upped voltage. That would mean that 3,6 GHz even sounds reasonable. One test i read upped the FSB on a quad to 460, stable.
However, ppl here with supposedly descent rigs have had trouble clocking significantly more modest than this would suggest. It seems a bit too... well, random is the word I guess.
From what you've spec'd sounds good - although don't "expect" a final outcome - you could equally get a crap ocing chip just like a good one, regardless of voltage and cooling.
There's lots of Q6600 ocing and spec threads in the hardware forum - try reading a few of those to see if they have some answers too :)
Thought I'd just follow up on my post. Did the bold (stupid?) move of just jumping straight to 333 MHz FSB (gives 3,00 GHz), and upped the VCore to 1,300 in one go. Works like a charm, np whatsoever. Raised the score in 3dmarks06 by over 1k. Max temp on any core I noticed is 67, generaly loads 60-65 and thats with ambient temps @ 25 degrees celsius.
Tried 350 MHz FSB and had one crash after a couple of hours of gaming, prolly lack of voltage. Since the watercooling is internal (sucks chassi air through the radiator) and the chassi also holds a hot GTX 8800, Im gonna install two new 12cm case-fans today to see if i can up the VCore a bit more and get stable at higher frequencies.
Anyway, the message being, with a descent mobo and cooling, you can get an 20-25% more Hz just by spending 3 minutes in the BIOS. Gotta love it!!
I am aware of some 680i chipsets that required a mem voltage boost with just 4gb ram to even post without overclocking, something of which i hope was addressed in the 780i chipset.
However if your board is booting fine, then id leave the voltages if you intend to leave them at stock speed, if you intend to overclock the ram also, then you may find you will need to increase the ram voltage, id personally not recommend increasing the voltage over any more than .1 of a volt max, than that of what they where running at fine before.
Mine will run @ stock on 1.8v fine, however to clock them a bit I found i had to increase to 1.9v
Some people have made larger increases than this, however im more on the side of caution.
system: asus P5N-T Deux 780i
Q6600
8gb PC2-8000
Nvidia Quadro 4600, it' for rendering
2x500gb RAID1
OC specs:
fsb to 333mhz
vCore Voltage to 1.4625V
CPU FSB Voltage from 1.2V to 1.4V
nForce SPP and MCP Voltages from 1.2V and 1.5V to 1.45V and 1.55V
Thanks guys
I am not sure how much you have tested already but for that OC on the CPU the VCore seems way too high. I have mine now at 366 MHz FSB with the VCore well below 1,4v. Doesnt it run very hot with that voltage? At 1,4 I could run it at 380 MHz but i stepped it down because it ran hotter then I like. Personally, I have set a limit that no core of the CPU shall exceed 65 degrees after 1 hour of 100% stress.
For temp monitoring, the easiest to use IMO is coretemp (free download). You can minimize it to the toolbar and it takes up like 2 megs if ram while operating. Just check so that the version you get is compatible to your OS. Since you have 8 gigs of ram I am gessing you have Vista 64-bit, dunno if coretemp is availible for that. Also, everest ultimate edtition (Company: Lavalys) has a function of monitoring ALL temps (including GPU, drives, Mobo etc). I am gessing that your system runs really hot under stress (unless you have some killer cooling systems). Hotter temps decrease life span and the temp of the q6600 shall never exceed 71 degrees (if I remember correctly).
Everest also has a stability test. Otherwise, the most popular seems to be Prime95 (also free download) but be sure to run it in four instances (all four cores). There are links to that procedure above in this forum.
Please get back with some temp-results when u have it. I have a Cooler Master Stacker 831 with 4 120mm case-fans and watercooling on the CPU. So if you can get descent temps with that voltage I would be happy to know the secret.
My BIOS allows for all other voltages on auto and I havent clocked the RAM, so I am sorry I can't comment on the other stuff.
/J
In the end, you have to decide for yourself if you want to do a max OC, set a temp maximum you think is reasonable, a max FSB you think is reasonable, or some other limit. If the goal is 333 MHz FSB, start at the above volts, increase FSB until u reach the goal, then lower the volts until you reach the level at which the CPU is no longer stable. The (obviously) take it up one notch again. If u cant reach 333 MHz FSB at the above volts, then increase them, while keeping track on the temps under load. I think the volts will be enough though.
/J
and i am not even running a liquid cooling system. currently i'm running a Zalman 9700 aircooler(what do you think of this one?)
http://www.newegg.com/Product/Product.aspx?Item=N82E16835118020)
i thought it was one of the best, but i am sorta new to most of this. so i guess if it is not such a great one, maybe i'll try to get the one tey had on here and sell mine or something. and then i'm running (1) 200mm fan, (4) 120mm fans (Antec 900 case).
as for the 8gb of ram, i am running XP Pro Corp 64. i needed the ram to help out with rendering. and coretemp worked perfectly with this OS and monitors all 4 cores perfectly.
i think this is what i'm about to try:
360mhz fsb
vcore- 1.4v
cpu fsb-1.3v
maybe don't mess with any of the spp and mcp voltages?
thanks guys.
at 360mhz, 1.4 vcore, 1.3 cpu voltage
the average temp was 55 with a max of 63
then i lowered the vcore to 1.3875, and a 1.25 cpu voltage
the average raised? to 56, but the max lowered to 62. idk.
those numbers were the high two cores of the 4.
question: why do you think core 1 & 2 would be the same, but higher than core 3 & 4? by about 5 degrees each.
well, just got done with a third test. hope i'm going in the right direction with this stuff?
360mhz FSB
1.375 vcore
1.2 cpu vtt voltage
average temp was 51 with a max of 57.
i guess i should keep going lower. but the cpu vtt v is as low as it can go, before it goes back to auto? should i keep going lower with the vcore?
Sure you can try to lower your vcore but not with those temps, no need IMO (You are running 4 instances of Prime95, one per core???). Just stay where u are and enjoy the extra stability. Try to fiddle around a bit with different torture tests on different cores. Do the "In-place large FFT:s" on all 4 cores to max temps. Otherwise, I prefer to do "blended" on 1 core, and the above on the other 3.
My recommendation since you promote stability and durability, don't change anything. You have a nice overclock of around 35% and got a frequency exceeding the really expensive quads at stock, so why bother.
Some computer-wiz may give you a detailed description of why core 1 and 2 are running about 5 degrees hotter than 3 and 4. I cannot, I can just say it is normal (btw, the cores are actually named 0,1,2,3). Your temps seems almost to good to be true (at least to me), please confirm that u get those after a few hours of 4-instance prime. If that is the case, congrats, you have gotten a much better OC-ing set than I have.
/J
anyways, the numbers i gave you before were from everest using its stability test like you suggested. right now i took a big leap down in vcore just to see what would happen. i put it at 1.3375. i have been running the prime95for a few hours on small FFTs, and monitoring the temps with coretemp. i am sitting at 58 58 52 52.
let me know what you think? should i put it back up to where it was at? or retest that voltage with prime95? let me know, later jerk;)
http://forums.techpowerup.com/showthread.php?t=32778
360fsb
1.3375 vcore
1.2 cpu vtt volatage
after testing a 4 hour cycle with prime95, temps sat at ahigh of 58 58 54 54
aftr 4 hour cycle with everest, it was 51 51 47 47
so these temps sound good. i'll doa 24 hour test with prime 95, but i thinki'm good to go. now i have to look at overclocking memory!
let me know what anyone thinks
If all apps works with that voltage there is no need to increase it. If something crashes along the road, a small increase is often enough.
I haven´t worked with OC:ing memory that much but a few rules of thumb is that you cant clock memory nearly as much as cpu:s, and you have to more careful with voltage. I would personally never exceed a +10% increase in neither speed nor voltage. There is also a point trying to get the memory to work at the same frequency as the cpu. That means that it could be a point in dropping the frequency on the memory of a 800 Mhz DDR2 (actual frequency of 400) memory to 720 (360) on a set-up that has a 360 FSB cpu. Or, setting the cpu/ram rate to 7/8 on a rig with a 350 mhz FSB cpu with the same memory. Lower frequencies on the memory generally enables lower latencies as well. I am not sure to what point the trade-offs are worth it though.
Also, see the RAM-manufacturers website. With a little luck, you will see that your ram has been tested at lower latencies than your BIOS has set it to. Be aware though that higher speed often equals higher latencies, so the tested latencies may not be accurat once you have OC:d the RAM. If you are highly dependent on RAM, check out a program called "Memtest". It's a bit of a hassle since u have to boot it from disk (standalone program). Personally never tried it but from what I can gather, it is the best memorytester. Both "CPU-Z" and "memset" will reveal the speed and latency setting of your current RAM. "SiSoft Sandra" can benchmark your memory. You get a couple of diagrams comparing your setup to other typical rigs. This is more to see if anything is really wrong, plus to compare to your own results after you have OC:d.
Though your GPU and mine isn't really comparable, it would be fun to see what u would score in 3DMark06 (free version available). Post the score if you try it will you?
/J
one last question, it seems that all the warning signs people spoke of as fas as an unstable system go, were (1) not POSTing, (2) prime95 actually telling you there was an error, (3) not booting up or not booting windows, or maybe some other drastic and obvious way that it is unstable. But what about freezing up? Yesterday while I was starting prime95, i started cpu-z to see something, and bam, it froze up. Then this morning I woke up and prime95 had went for quite sometime, but the comp was frozen. so i just restart it, and it seems like no problem.
i would assume this is considered a symptom of an unstable computer also right?
also, as for the RAM, yes, I would much rather lower the latencies of the RAM rather than OC it. It is DDR2 1000 (PC2-8000). i think that i would much rather try to lower the latencies than oc it. plus, i hear it yields better performance in the trade off.
ok, post that score lata
(1) I have windows xp pro 64bit running, and it needs vista. and
(2) " Your display card is significantly more powerful in raw performance
Your display card fails the following requirements:
Pixel Shader 4.0, you have 3.0"
so i dont think i'll be running 3d mark damn it!
that sucks
so what do you think happened? think i screwed something up permanently? ****, i hope not...
I really have no idea what happened but for that overclock/volts on air, those temps are pretty normal. Strange you got others before though. Be careful to run the same processes in Prime, as different processes on all 4 cores can easily give a 5 degree difference. Apart from that, I am clueless.
I have moved down even more and am at 350fsb now, with lower voltage. That gives me much nicer temps and it its still faster than a $1000 quad so...
Another thing to put the fiddling in perspective: Having the window a bit open in my room while I game has the same impact on cpu-temps as an extra fan for the water-cooling :), and its cheaper/easier.
I am at 37,37,40,43 idle and 58,58,62,63 under load right now while stable in everything I throw at it @ 3,15 GHz so I will officially stop fiddling now.
My specs
Intel Q6600 B3
Clocked at 2.66 temps:
Cores: ~54C
CPUTIN: ~44C
VCore: 1.240v
I did the pencil mod on the mb, so no vDroop
Thermaltake V1 Cooler in Antec P180 case (cpu fan exhausts straight up into another (120mm) case exhaust fan on the top of the case
FSB voltage set to auto (believe it's 1.2v)
temps, voltages from CPUID Hardware Monitor (Speedfan reports ~12C cooler, going with higher reported temps) CPU-Z and Core Temp
I think I found the culprit! After reading this review at xbitlabs, my TT V1 is apparently incapable of cooling oc'd quad core procs. I wish I'd done this research before I bought it! Fortunately, it's worked fine for the past 10 months of running stock clocks and I believe I got my moneys worth out of it, however, looks like I'll need a new cooler.
I recently bought these components:
XFX nForce 680i SE SLI Socket 775 Motherboard
Intel Quad Core Kentsfield 2.4Ghz Q6600 CPU
(4)Corsair 1024MB DDR2 @ 800MHz
I also have from old PC (1) GeForce 8500 GT 512MB PCIe **
** I Will be upgrading this video card soon to the following one:
(2x)EVGA GeForce 8800 GT Video Card - Superclocked Edition, 512MB GDDR3, PCI Express 2.0, SLI Ready, (Dual Link) Dual DVI, HDTV, Video Card
My question is, how do I over clock the Q6600 with the SLI ready ram. I am new to this Over Clocking and don't want to screw up.
Any help would be most appreciated.
Thanks
I made a slight mistake here , it should read >> XFX nForce 680i LT SLI Socket 775 Motherboard <<
Upped my FSB to 1334 Mhz
unlinked Mem Clock
Mem Bus set@ 814Ghz ,shows 800.4 Mhz
CPU volts set @ 1.42 v shows 1.36v
Mem volts set 1.9
FSB Volts set 1.4
PassMark Performance test 1475 @ above settings
Keep an eye on your temps CPU is up to 68 under load.
Looking for better cooler than my bogo intel one any sujestions welcome.