Check tRD...maybe put it up to 9 or 10.
Check tRD...maybe put it up to 9 or 10.
I found that with the 500 FSB I needed to relax tRD to at least 10 to get these sticks to post. On Auto it bumps it to 13.
got the asus p5q-e , see sig enjoy :)
also tryed higher trfc. wont help :(
http://www.abload.de/thumb/unbenanntoy1.jpg
Try tRD=13.... that's what mine default to on AUTO when I put the FSB to 500.
(When I had these sticks, I sold them now)
tRD is the performance level isnt it ? it is on default on 13. same results. i just pushed it to 10 down for the test.
Yeah...tRD is performance level.
Well, then I don't know what to tell you. You might not be able to get 1:1
yay strike :cool:
i set my memory manualy in bios up and it works with your settings you told me, just passing 2 hours of blend test :D BIG THX! 8x500
how should i set it up now to go for more fsb and about 1060 @ ram or 1076 ?
higher trfc / tRD ?
Well, you might want to check your bandwidth and latency in Everest before relaxing the timings too much. I find when I relax tRD too much, the latency really goes up.
dante`afk
On a P5Q series board, almost certainly 9x445 [5:6] DDR2-1068 will be better than 8x500 1:1.
You'll be able to lower voltages for the mobo, & will be able to run a much lower (tighter) tRD (aka Performance Level) & assuming the RAM can do that speed, much faster RAM too.
You'd want to check Everest to be sure, but i'd expect a properly tweaked 9x445 [5:6] DDR2-1068 to be a much better config than 8x500.
1:1 is not your friend on the P5Qs.
Exactly.
Even though a 500 FSB is an impressive feat, unless you can get the timings out of the memory, you often end up with worse performance than a lower FSB with nice tweaked memory. I run 467 FSB with my memory and get much better bandwidth/latency than if I go to 500.
hehe thanks guys, will test this. blend test errored after 7 hours :/ - memory got still errors im memtest as i saw now
running currently at 9x435 - 440 crashes..
trfc 75 - works memtest passes. but i have ai clock twister for mem on "lighter" - seems to be no performance lose if i run everest memorybench.
in another forum i got told that my ram is broken, i should send it back ?
9x400 - 5:6 - ram@500mhz - no errors in memtest
9x500 - 1:1 - ram@500mhz - full of errors
?
I doubt very much that your RAM is broken. It puts a lot more strain on the system to run the 500 FSB vs. 400 FSB.
im very disappointed about this :( chipset is p45 .. asus p5q-e.. should do this w/o problems ?
i run 8x500
small ffts w/o errors
large ffts w/o errors
blend - errors after 1 hour
i rly believe its just about the ram settings.. so i get errors in memtest+ :(
What is your vMCH? You may need to raise it.
1.32. tested till 1.4. think more wont be nice for the p45
Do you guys think it would be relatively safe to run 2.2V through these sticks if they check out stable in OCCT @ 1130mhz? (bios actually reports 1133mhz).
Fairly heavy PC usage (gaming/watching media, ect). PC isn't on 24/7 all the time, but it can be on for long spells consisting of a few days.
Thanks :yepp:
ps. Keep in mind I'm in the UK, so for what it's worth room temperature isn't that high usually :P
Yeah I might drop back down to 1066mhz and run at <=2.1v.
1133mhz stable is on auto settings as well, I haven't tried to tighten. Well except from it's running in 5-5-3-15, instead of 5-5-5-15.
I'll see how tight I can get the sticks, then I'll see if I can drop to 2.15v or 2.1v, but I doubt it.
Just ran OCCT with the memory at 1130 @ 2.1V and it finished the 1 hour test as stable.
Not tried to tighten anything.
With running the memory at this speed is it even worth trying to tighten things?
If so, what would you guys try changing below in memset?
Thanks
http://i85.photobucket.com/albums/k4.../gskillmem.jpg
http://i85.photobucket.com/albums/k4...es/e660034.jpg
My F2-6400CL5D-4GBPQ does 1000mhz @ 5.5.5.13, 2t, 1.8v
and 800mhz 4.4.4.11, 2t, 1.8v
stable for 1 hour test.
What's the max safe voltage for my kits ?
Would anyone know if the TR HR-07 coolers work on these?
http://www.performance-pcs.com/catal...ducts_id=22834
As long as they are doubled-sided they would be ok? I wasn't sure on this point. Thanks
As far as I know though this GSkill runs fairly cool and doesn't take much voltage...so would the thermalright heatsinks be just for looks for you then?
No it is not just for looks, ok?
As far as I know, it does run hot in my configuration.
My mobo requires +.03 voltage adjustment to run properly (per G.Skill specs)
Setting +.03 = 2.08, 2.06 vddrop
Treid +.02 = 1.98-1.96, 1.94-1.92 vdrop, sluggish
If you have any other cooling options, I'm all ears.
@kpo6969
You could just use memory coolers instead. Changing the stock heat spreaders results in void warranty especially if you damage the stickers. Mine runs cool at 1066Mhz 1.9v. I also got an Antec Spot Cool blowing air towards it.
http://i21.photobucket.com/albums/b2...IMG_3394-1.jpg
That looks just like the Antec Spot Cooler I ordered from Performance PC yesterday. Thanks for the reply. Does it help?
Yup it sure does. Even without the Spot Cooler the ram running at 1066 1.9v is already cool. It warms up a bit when playing games or running benchmarks, but with the Spot Cooler cooling it it runs cool most of the time.
This ram is amazing, booted right away @1080mhz@5-5-5-15@2.02V. Just a couple of questions, With memtest V3.4 do i just run the test that it starts right away (i am using a boot CD,not in windows ) and if it completes 100% its stable right?, also what timings should i try lowering 5-4-4-12 etc
Just found something weird about this Ram. It ran stable on my old Asus P5K-E Wifi/AP at 1066 1.9v. Then I bought a new board, the Asus Maximus II Formula. I ran into problems, OCCT and Orthos Errors and failed memtest after 17passes I got 2 errors.
So I had to run them at 1066 at 2.0v on the M2F. Passed memtest and Orthos and OCCT....
Ever thought that ya P5K-E maybe overvolted big time ? With P5Q Dlx we have the same issue as my board overvolts with about 0.08 so with 1.92 in bios I get around 2volts... some boards have this phenomenon others are spot on...
I got my system stable with 4 sticks (8 gigs) at 500 FSB, I had to up the Vdim to 2.17 but I dont think it will hurt.
http://img.villagephotos.com/p/2005-...Gigs500FSB.jpg
5-5-3-13 seems good @ 2v.
This ram is so cheap now - $72.99 free shipping at the egg:
http://www.newegg.com/Product/Produc...82E16820231145
I ordered a pair of these as well, not as cheap as you guys can get it but for Ģ60, should get hold of them tomorrow for some testing...
Because of some funky undervolting / underclocking i am doing for an HTPC rig, I was going to run these sticks at 333MHz 4-4-4-12 @ 1.8V. Would this work?
Mainly because i'm trying to keep VDimm down to 1.8V and lessen system draw. Again, I'm using in an HTPC environment so heat, power, and silence are of the utmost importance. 100mV-200mV doesn't seem like much, but keep in mind I'm running my Core2 Quad at 0.89V....
The power consumption difference and heat difference between running this kit @ 1.8v and 2.0v is unnoticeable. And if heat, power and silence are the most important thing to you, then you should be looking at a 45nm CPU. My guess is that you are going to use a Q6600 no? Good luck, the new ones need tons of volts to do anything (even downclock)
Your assumption is incorrect, I utilize a q9300 yorkfield and before that e3110 wolfdale (sorry but contrary to most peoples impression, real htpc folks don't just slap in piecemeal outdate parts for their systems). Because the environment needs to be as silent as possible, minimal cooling will be used so the slightest increase in power/heat will be greatly exacerbated. Anyhow, I am using a pc8000 patriot viper memory (with comparable specs) from my other system and it works so hopefully these gskills can too.
Never realized these could do much more than 1000. Gonna try some higher speeds and back off the volts.
Still $70 free ship @ the Egg.
I thought the $115 I paid was a deal. If I could see the use for more than 4gig I'd get more, but I've never dipped much below 2g available.
Has anyone had a set of these that took Over 2.1v to keep em stable at say 480 to 485 fsb?
My sticks act that way, they did this on the DFI x48 and now on the P5Q-D. measurements taken with a dmm on both boards? Just curious if I have a Very bad set or if this is something someone else has run into.
Larry
If the set won't do 500MHz+ with 2.1v then I think you should RMA it, its a dud set.
I'm in for 2 of these sets. Can't go wrong with 8GB of memory for $136. Thats unbelievably cheap!
I didn't really want to spend the money, but recently upgraded to an e8200/Asus P5Q SE/R. Coming from an AMD Brisbane all I can say is 'wow' ...was running G.Skill 2GBNQ and could only get 450MHz out of them Prime Stable. I know these sticks are holding the processor back because I easily hit 3.6 with a +.02V adjustment of the processor. With some luck, hopefully this chip will hit 4GHz on these modules on air.
Larry,
I just got my 8 gigs of this stuff in my system and seem to be having similar issues to yours. At 500MHz FSB on an Asus p5q se/r I need to put 2.2V through them to get them to pass memtest without errors. This is with the latest BIOS.
It always makes me a bit nervous to volt over spec. I will test this a bit more and let you know what I find.
I got 8gb too, last night I was trying 7x500 but very unstable.
5-5-15-55
ST RD Disable
the other RD option disable
Mem OC Charger Enabled
AI Twister = Lighter
NB = 1,32V
NB GLT = 0.600
Mem 1.98V, I heard P5Q have a + .10V in memory, so 1.98V gives 2.08V if a increase to 2.00V it freezes when trying to run any 3D aplication.
And some of you are posting with more than 2.10V I donīt get it :shrug:
I have another rig with 2 pair of this ram on an Asus P5K board, 8GB @5-5-5-15 2T 2.1v 500Mhz, running with zero problems, it could easily go higher...
I use dividers, 1:1 gives no performance advantages...
I am running 1:1.
Also, I have no idea what the "real" voltage is I am sending to these modules because I can't seem to find out how to monitor vdimm on the P5Q SE/R. This board is kind of 'eh'. You can't set voltage to the hundredth on this thing.
The only memtest stable setting I have found so far is MANUAL 5-5-5-15 2.2v with ALL of the remaining memory settings on AUTO. This includes the Static Read Control/OC Charger, ETC.
Again I have no active cooling on these modules and am nervous about volting them over spec.
It looks like you have too many settings configured to isolate this to only the memory. Take the processor out of the equation! I would try to set as many settings to AUTO in the BIOS as possible, run on a 6 multiplier and boot into a memtest.
Try setting tRFC from 55 to AUTO on your board. Try running with 2.1v all other memory settings to AUTO with the exception of 5-5-5-15
...by the way, I just purchased these from newegg and received '840' (2008 Week 40) from the Whittier CA warehouse.
Wow...
These are only $68 on Newegg now.
I envy Americans, but I always wondered why the economy over there is breaking :ROTF:
Over here they are Ģ58 (around $115).
Back when I got my OCZ Flex II, they cost Ģ76 and mine were, and still are Ģ99.
:shocked:
Nops, not working, setting 500 and everything else as the manual said, freezes during boot, a littlejuice on the NB help to boot, but its unstable.
Tried 2.0V, 2.02V, 2.04V, unstable, 2.08V freeze.
9,5x333 and memory set do 1002:up: , 500:500 :down:
I dont know if there is much more to change in the bios to make this work, tomorrow I'll try 4GB just to confirm that its a memory problem and not a NB issue.
Right now is running at 490Mhz 1:1 with 6x multiplier, and so far its working.
EDIT:
Houston we have a problem, 490 is a no go.
Just so you know, with 8GB I am running these at 2.2v. 1.3v NB @ 500MHz FSB. I emailed tech support from G.Skill and they said running 2.2v through these modules was fine. Memtest ran 10 hours no errors.
EDIT: What the tech said was that my motherboard *might* be under-volting the memory a little and that if they work at 2.2v it would be fine to leave them there. I honestly am fine with this seeing as they have a lifetime warranty and I replace my components every couple years. YMMV
=]
I couldn't get my pair to get to 500 FSB either 1:1.
In my experience, these sticks will work fine at 2.1 real voltage. I have to set it to 2.15 in BIOS though.
Having said that, I've noticed that when overlocking with 8GB of RAM, one needs to give their board a lot of voltage. Although, even with 1.6 on the NB I wasn't able to get 500FSB stable. I'm starting over again from scratch, because being an idiot I didn't write down my stable settings.
Heh. My 4gb kit need 1.98v for 480fsb, 2.04v for 482fsb and 2.06v for 484fsb on my board! How weird is that?
Larry
Hi
I don't push my PQ's to 1066Mhz trfc 60 5-5-5-18, i'm tryied with 2.1v and the result is the same, BSOD in vista.
Only at max. 1033MHz 5-5-5-15 trfc 55 are stable with 2.1v.
please help me.
My hardware: q6600, 2x2gb gskill pq pc8000, asus p5q deluxe bios 1406, msi 8800gts g92 oc, ocz sxs 600w.
Thanks.
(sorry for my bad english)
Add a 120mm fan to cool the sticks and raise the juice up a little to see if that helps.
Okay....
Is there a difference in the divider between 400/800 and 333/667? Because, to me, logically, they are both 1:1. However, when changed from 400/800 to 333/667 and I can now get my tRD value to 8. If I change the tRD value to 8 when I use 400/800 my computer crashes. I'm confused.
Yes, there is difference between the different straps and what the tRD can be set to for each. Anandtech's article is a good read in that respect.
Same divider, but different strap (333 vs 400). The strap determines what tRD can be set to.
Reporting my results with this memory with a P7N Diamond (signature has full details of system).
4GB:
2.2v, 1050, 5-5-5-15-2T, 45 (tRC)/52 (tRFC)
8GB:
2.2v, 1018, 5-5-5-15-2T, 45/52
1066 is unstable at 8GB with those settings. 1083 is unstable at 4GB with 2.2v, 1050/1066 with 2.1v. Rock solid at the settings listed above, though. This fits the typical profile of G.Skill in my experience - tends to overclock at least a bit, but has a very definite envelope of settings. Not disappointed in the least, especially at this price, and I'm surprised it worked so well with this board in the 8GB configuration.
me and my buddy were on the phone working on his tonight,we were able to get 9800MB/s read and 8200Mb/s write on everest with a 54.1 ns latency at 5/5/5/15 - 1081mhz 2.25volts
Has anyone ran this at DDR2-900 Cas4 2.1v successfully? Im having problems running them at this speed:
Q9650 @ 450x9
FSB Strap to North Bridge : 333
DRAM Frequency: DDR2-1081
Dram Clock Skew CH1 A1 : Normal
DRAM Clock Skew CH1 A2 : Normal
Dram Clock Skew CH1 B1 : Normal
Dram Clock Skew CH1 B2 : Normal
DRAM Timing Control: Manual
CAS# Latency : 5
RAS# to CAS# Delay : 5
RAS# Precharge : 5
RAS# ActivateTime : 15
RAS# to RAS# Delay : 3
Row Refresh Cycle Time : 55
Write Recovery Time : 6
Read to Precharge Time : 3
Read to Write Delay (S/D) : 8
Write to Read Delay (S) : 3
Write to Read Delay (D) : 5
Read to Read Delay (S) : 4
Read to Read Delay (D) : 6
Write to Write Delay (S) : 4
Write to Write Delay (D) : 6
Write to PRE Delay : 14
Read to PRE Delay : 5
PRE to PRE Delay : 1
ALL PRE to ACT Delay : 5
ALL PRE to REF Delay : 5
DRAM Static Read Control: disabled
Dram Read Training : disabled
MEM OC Charger : disabled
Ai Clock Twister : Moderate
Transaction Booster : Manual
Common Performance Level [8]
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 CHA PH5 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
Pull-In of CHB PH5 Disabled
DRAM Voltage : 2.118v
But im running it stable at DDR2-900 at all the setting above, but changed this:
FSB Strap to North Bridge : 333
DRAM Frequency: DDR2-900
Common Performance Level [10]
DRAM Voltage : 2.118v
Thanks in advance!
Powerchips don''t like 900 MHz Cas4...maybe 850.
bump* (sorry, I know I'm late to the party but this is the ram I just bought and there is a wealth of information in this thread. Didn't want to lose it before I was finished)
Glad that this can still help people. :up:
Hey ENJOY, would you be willing to help me out a little? I must have read this thread at least twice. But there are some things I'm still confused about.
Ok for instance......@ 996 MHZ, (everything auto in bios)
I can pass memtest all day long. Same at 1003MHZ.
< it's my chip, q6600>
However, if I set the memory manualy in bios to 1066mhz, everything else auto, memtest fails at test #7 I believe.
If I set it manualy to 1066mhz and then manualy set the volts to 2.10, same result. Fails memtest.
Any suggestions? Haven't even begun to ask about timings yet. Lol.
ok, i'm not trying to be funny here........did you mean post my questions HERE, or the link in your sig (current) :shrug:
Cause if you meant HERE, I did ask my question. Why is it I can't pass memtest @ manual setting of 1066 mhz with manual setting of 2.1 volts in the bios.
Again, sorry for being late but I've read this thread (all 28pages) twice. :shocked:
I believe "EnJoY" meant to post your questions here. :) It would be helpfull if you made a signature with your Hardware listed in it, will be easier to field questions that way. Also, quite a bit of the PC8000 ram will Not do 1066. Mine won't for instance, so you may be asking for too much out of it.
Larry
Quote
"If I set it manualy to 1066mhz and then manualy set the volts to 2.10, same result. Fails memtest."
See now, with the signature I could tell what board your on, the above quote is part of the problem. The Asus P5Q series boards overvolt a bit. 2.1v set manually to the ram is more than likely giving it about 2.18. Maybe try 1.96 to 2.02v for the dram voltage.
The other part of the problem is that the ram may not do 1066.
Larry
When I posted my reply, you had not yet edited your first post with the question (my post 4:30PM, your edit 4:31PM)...therefore I was telling you to post your question here in this thread. ;)
You're problem is uncertain at this point for several reasons. Basically, you have not given us enough information.
What divider are you running? What speed is your cpu running at when you try 1066, and are you sure it's stable at this speed?
Two suggestions:
1. Start at 1.9v and test 1000MHz, stock timings. Should pass.
2. If you have truly read all 28 pages of this thread, you should have seen several screen shots from various other members in this thread that detail the memory timings used to achieve certain frequencies. I NEVER recommend using AUTO for your timings, because you don't know what sub timings are being applied in this case. With ALL memory, adjusting and tweaking the sub timings are a necessity when trying to squeeze additional speed and performance out of your sticks.
I suggest you re-read this thread and examine what other people are using for their sub-timings with the P35/P45 chipset and this memory.
Once you have your sub timings set, let us know how 1066 does and what those timings are.
Yeah Enjoy, as I was typing my "edited" version, I was thinking that you may have already read just my original request for help. I'm sorry. Like I said, I honestly wasn't trying to be funny or in any way trying to be hard to get along with , with my second post to you.
(^ the only reason I bring this up is because.......for whatever reason.....I seem to find a way to ALWAYS come across to people IRL as well as forums as being somewhat dickish. I'm not that way at all. I just have a knack for coming across that way. :shrug: I've learned to accept that fact, even though it's far from what I am really trying to portray)
To Big.....yeah, I know, posting my system specs was/is a critical factor of information to have if I expect help. To not post it to begin with , especially in my sig was just a bone headed oversight on my end.
To Enjoy. Yeah, I promise you, I have read this thread twice. Let me back up just a second. This may clarify some things for ANYONE wishing to help me.
Enjoy, some of your questions I can answer, some I cannot, because I just don't know. (ie still confused)
I look at you guys as die hards. <not a thing in the world wrong with that, and THAT is why I come to you, the experts> We all have our indivdual talents. PC overclocking is not one of mine. (BBQ is :up:)
So in my mind, my world, I'm a small fry in sea of sharks here at XS.
When you say divider...........? I don't know. I don't know what that means.
Sort of. The only one I really understand is 1:1. As soon as you go past that...I'm lost.
Yes, Enjoy, I could / have looked at the screenshots.....and I have no doubt I could set up mine the same way. The problem with many of the screenshots I have seen is comments made before the screenshots........for instance:
"Ok, here's 4:5 running 5-5-5-12 @ 1.9 ........may as well add "while in a crescent moon stage, and low tide"
I cringe at what I am about to say because I fear it will come across "wrong" again,,,,,,but I don't care about that stuff really. Sure, I can set up my timings just like theirs, and do some testing. But the thing is, when it's over, I have the test results, but have NO clue what they mean. Is Higher Bandwidth better than lower bandwidth? Is faster read/write times better than lower? I realize there is a 99% chance this board (XS) was probably not the best one for someone like me to come to. < XS rocks, I'm just old now, and getting older fast) I was late to this party. Most are moving on to newer sockets and i7, and I'm just now coming to p45 and quad, from a 1.4 T-BIRD.:shocked:
The learning curve this go around has been HUGE for me. Anway, Enjoy, I honestly cannot say if my CPU was overclocked or not. I don't remember.
So, If I may be so humble.......maybe if we could just start from scratch? I undersand the very basics. On one side you have your CPU and FSB. You can overclock just your CPU via FSB. Then on the other side you have your ram. (won't even go there on vid card) (yet:p:)
The CPU / FSB side of the coin tends to be easiest for me to understand. However, I'm using stock intel cooling. I'm worried about heat. So I decided to lay off that aspect of the overclock and just see what my mem can do sinec I have read that for the most part, heat is not an issue with this set.
Tell you what guys. I'll go back to step 1 with the CPU. I'll set CPU at 333X9
which is 3ghz. I'm ok with that heat. So let's test that first. Ok? Here's what I'll do. I'll leave the CPU voltz on AUTO for this innitial overclock.......and then test using Prime 95 (ok?). With each pass of P95, I'll lower the juice one step at a time until it starts to fail. Then I'll raise it up 2 notches and test again. It should pass. Until that time , I won't report anything else...
My only question now, is what should the ram setting be for this? Manual? Auto? And at what MHZ? Fair enough?
Ok, lots to cover here.
I am not going to type up an entire introductory guide to overclocking for you here, as I simply do not have the will, nor the time.
First rule to remember though. Whenever you are trying to overclock one part of your system, you first need to remove the other parts from the equation. This can be difficult at times, but generally it's quite simple.
Here is an example of what I am referring too...
Say you were trying to determine your maximum fsb speed? You would want to drop your cpu multiplier to something low (I generally like 6x), set all voltages to stock (not AUTO), and memory timings to defaults. You would also want to set a memory divider that was very low, 1:1 ratio would probably be ideal and work fine as with DDR2 and DDR3, you would have plenty of overclocking headroom with the FSB before you even reached the stock speeds of most of today's memory.
From there, you would simply go up 5-10mhz at a time, testing for stability at each increment, moving up again until you hit a wall. Once you've hit the wall, increase the appropriate voltages, and continue moving upwards. Generally when clocking fsb and cpu, I prefer using windows based utilities as it is a lot faster. This way you can quickly bump up the speed, jump into super pi or prime95 for a few mins, and keep going until you error out, then back into bios to adjust the voltages. Understand?
Testing CPU, you would follow the same process, but in this case you would simply set the highest multiplier available to you, and adjust fsb up from there. Assuming you tested fsb limits first, you now know exactly what fsb frequencies your board is capable of and therefore it will not be a factor as you increase mhz to determine the limit of your cpu. And of course, your memory is on a low divider and isn't going to be anywhere near it's rated speed at any point in this testing process, so that is out of the picture as well.
Once you have the max frequencies for both your fsb and cpu, the rest is easy. You now know the limitations of both your chipset and your cpu at every voltage step along the way. Now you can set these at safe limits, and if needed, adjust them along the way as you overclock your memory, knowing that neither of them are a factor in your system's stability. To start, set your cpu to a low multiplier, drop fsb to something like 250 or 300MHz, to avoid hitting a wall too soon. And set a divider that makes sense for the memory frequency you wish to start at.
Typically, I always had my best successes in tweaking memory when I started from a very low frequency, testing lower latencies. For example, for this ram, I started at 750MHz, with maybe 4-4-3-10 timings and 1.8v? And from there I would just into memtest and see if it's stable. It is? Great, how about 1.7v? No? 1.75v? Yes, cool. So 1.75v is now my starting point. 4-4-3-10 is stable, how about 4-3-3-10? Or 3-4-3-10? No, ok well then I guess I'll stick to 4-4-3...but how about 4-4-3-8? See what I mean? It's all meticulous trial and error until you find the sweet spot. Once every is stable, you bump up the frequency, 5mhz at a time until you hit another wall? And voltage should always be the first remedy to a wall, if voltage fails, you no it's not a voltage issue but a timings issue, so you go back and adjust the timings.
This my friend, is how we overclock; or at least how I learned to do it. Some may disagree or have other methods, but these methods have always served me well.
Sub-timings are the tricky part these days, however. BIOS' these days can have so many options available for adjustment that it could literally take you weeks until you found the perfect timing combination for your RAM. This is why for you and many others I simply recommend the lazy man approach, and copy off a fellow forum member. See what other people are running for their TRRD, TWTR, TRFC, etc etc, as well as the primary timings and frequency they are running. If you are at 560MHz 5-4-4-12 and you're looking for the right sub-timings, it's probably not a safe bet to copy someone else's sub-timings who is running only 450MHz and tighter timings. Remember, the appropriate timings and sub-timings will change as the frequencies change and as the other timings change around them. As an example, the appropriate TRAS is basically always decided by what the first 3 timings are set too.
I'm already written way too much now, but this should be more than enough of a guideline for you to use to get started.
:up::up::up:
Got it! THANK YOU, Enjoy.
1)FSB.
2)CPU.
3)Memory.
Picked myself up a thick notebook today and some pens. Time to start testing.
I found that connector you were looking for
Cindy, is this the thing you were looking for?
www.liangdianup.com/computeraccessories_1.htm
It's on the list of computer accessories and parts. They have the DVI video thing to convert that jap monitor to work with your other computer. Just about any other kind of wire adaptor, usb connectors, monitor extension wires, ps2 extention wires, and all kinds of female and male swap connectors and things that I think would help your shop. If that above link don't work then goto www.lducompany.com and click on computer accessories. Let me know if that is what you need and give me your email address again.
Enjoy
When you say you like to use windows apps to tweak your CPU and FSB and voltage....if I am reading it right.....you are talking REAL TIME changes, within a single instance of the windows evironment. (ie no changing and then restarting, but changing and testing)
Then when you have it figured out to your liking, THEN you sort of hard set those numbers in BIOS. Correct?
May I ask what application you use to do this?
Not voltage, just fsb/cpu.
You set voltage is bios, head into windows and use Clockgen or a similar utility to up fsb in small increments, quickly testing each one for stability. And yes, once you have a stable setting, you head back into bios, set the new MHz, up the voltage again, head back into windows and continue clocking.
clockgen is available at www.cpuid.com
I just got a set of this off Craigslist for $25. Guy was running it in a hackintosh, so hadn't been OC'd. Have to see what i can get them to do with a P5K premium.
Got it. Thanks again. I'm going to work backwards though. Since I already have a ball park # in my head of my CPU speed, and FSB, ......
^(Read, I don't have the proper cooling yet ;)
I'm going to shoot for CPU / FSB stability @ 3 GHZ with a "set" memory timing.
Like you said, If I have a set memory timing like 800 mhz or 900 mhz, then in theory, knowing this particular memory can easily handle that , it SHOULD take the memory OUT of the picture.
Then I can work backwards, testing for stability, while slowly decreasing my voltage to the CPU, and thus lowering heat. Seems to me it's now just a matter of finding the correct voltage to run my CPU @ 3 ghz. stable.
Once I have done that, then I can focus on tightening memory timings.
Thank you again, Enjoy for your help. I know what I just posted is probably everything anti-overclocking and how to do it, but I already said I'm not much into benches, screenies, or the like. I'd rather be playing, surfing, or rendering in Poser. Anyway, I'm pretty sure I have it now.
3ghz doesn't sound all that hard to achieve for a q6600.. nor does it require high volts. You should do fine