care to clarify? that fsb seems a bit steep. :-)
Printable View
care to clarify? that fsb seems a bit steep. :-)
BIOS advanced/chpset/memory remap feature=disable. If you want to tighten timings you can go 1 of 2 ways - do it in the bios and see if it boots then run superpi to check followed by Orthos or use memset which will allow u to change settings in windows. Using memset you can test your settings without having to reboot each time although not cas setting - tends to crash the system!. Once proven you can change in the bios or setup memset to change on entering windows. Memset offers you more settings to play with than the BIOS but you need to know what you are doing. Tighter timings usually need more volts to stablise if you are o/c the memory - just watch you dont give it too much volts 2.2 -2.4v is max for most.Search what others have used for settings and o/c with your memory. Before you wander into unknown territory check first post on this thread - Read this guide http://forumz.tomshardware.com/hardw...ict197995.html Get some AS5 and decent cooling and away you go.Quote:
Originally Posted by ryanyeo
Sending poor newb away already? :stick: :DQuote:
Originally Posted by sesdave
theteamaqua, we need more cow bell! :D Seriously what you are talking about? How to hell do you hit 478fsb? and is that vcore or mch volts you mentioned?
Lol! he will have fun as he learns by trial and error!. You might find this memory arcticle interesting http://www.xbitlabs.com/articles/mem...ory-guide.html.Looks like a suicide shot and vcore.
Hi Crazyea,
I sort of wonder about the SPD idea. If you run via SPD, CPU-z just shows looser memory timings than suggested by OCZ. Yes, I may have to loosen the memory timings to get higher. Particularily by running two sets of memory. Otherwise, a person should be able to run the same settings manually as what SPD would give you. Does anyone know if the SPD settings affect any other memory items outside of timings?
I also noticed that some of you have moved from the 9x266 to 8x266 and then ramp up the FSB to 400 or whatever. Has anyone run any tests to see if a system is better off using the 8x multiplier and a faster FSB versus the 9x multiplier and lower FSB settings? Something I may try later. I did send a message to the lawyer person on the front page. On another forum I was told that there was no use having a memory setting higher than double the FSB, ie 333 FSB relates to 666 in memory. I did run some tests changing nothing but the memory speed from 666 up to the 900 my memory will support. The argument being that the CPU can only handle the double data rate and no higher. From my tests, the higher memory settings do indeed result in a faster computer tothe tune of 6%. 6300 versus 6700, if memory serves me right. I agree there should be no reason for this faster result, but ther it was.
I am pretty sick today so I won't get to trying out any of the things you guys suggested. I do have a fair bit of AS5 on hand, so I can try using it on the SB chipset. I already have it on the CPU and NB. Has anyone permanently removed the fancy covers over the NB/SB chipsets to see if that relates to any better cooling? It probagly does, but only to the tune of 1-2 degrees.
I'll get back at you guys in a day or two. Thanks.
Having to use SPD is a limitation of the chipset not your memory.
So I have had my board for a while, and was looking to push it further. First page of this thread says you can fit an nb-1c on this board and I have one laying around but cant for the life of me figure out how to go about mounting it. Could some post a pick if they have one installed or fix the thread if they cant be installed.
Thanks
Barrman
Anzial posted the pictorial guide to the NB?SB mod a couple back and yes you do pull off the covers. I have done it and stuck a 40mm fan on the NB/SB - it should reduce more like 5-10c.Quote:
Originally Posted by AHolland
hiho :)
i'm looking for a bit of help maybe because i'm frustratingly stuck at 3954.22 MHz and cant seem to hit 4GHz :(
hopefully this can give most info: http://valid.x86-secret.com/show_oc?id=184077
(ram is actually 6400, is it showing 5300 because of the divider?)
adding more volts to the cpu and ram doesn't seem to make any difference at all.
timings are set to SPD.
chipset is 1.65 (max)
even putting up the fsb by 1, will result in a failed boot :*(
any ideas guys?
haha lol i bought one of the 1st nbatches availible at newegg .. like June or May last year .. and it died about 1 month later ... and about 2 weeks ago i found the CPU cap .. so i was able to rma it... and now i got the 1.04G rev hahah what a wonderful story
anyhow how is 1.04G different than 1.03??
btw with TEC i cant seem to boot past 470fsb x 9 with 1.7v set in bios ... which i was able to do with P5N32E SLI PLUS and it was benchable actually
As its more than likely MCH volts holding you back on that revision of board you could try this pencil mod (and the rest if you needs them) to get more volts for MCH http://sg.vr-zone.com/?i=3876Quote:
Originally Posted by sb85
Have you tried different multipliers ? Mine clocks better at 9 and 10 then 11.Quote:
Originally Posted by sb85
Beside that - you should realy try to find a board with mch 1,85.
Just ordered 4GB Ballistix 667 :eek: such a good deal on them just couldn't resist. Hopefully they'll clock good and boot manual timings - fingers crossed.
2 x 2gb?
Nah, way too expensive for my range, 4x1GB runs well on this board from my experience anyway :)
Hi Crazyea,
I'm basically responding to you because you were the first to respond to my inquiry. Thanks.
The point I was using regarding setting the SPD or not; The SPD is a chip onboard each memory stick that lists the factory default memory timing settings. Factories traditionally set the SPD more conservatively than what the memory will really do, or are rated for. The motherboard then has a nice safe starting point to go from. My argument is that the more intelligent users, like ourselves can ignore the SPD and manually alter the timing specs, something a lot of people do not want to do. If you have found that you need to run SPD to get to a faster setting, than that is because the SPD setting is making the memory timings looser which traditionally makes your machine slower but more stable. I have used the factory settings of 4-4-5-15, or lossened them to 5-5-5-15. The SPD settings are something like 6-6-6-18 (as per what CPU-Z gave me).
With my testing so far, I have found that the SPD settings in no way make my machine more stable. That does'nt make sense to me as the looser timings should make my machine more stable. All my testing seems to indicate that my memory is not my issue. The funny thing is that it does'nt seem to be the CPU either. Here's where it get weird.
I can run my memory up to 900MHz at timings of 4-4-5-15 as long as I don't use the 1:1 setting. If I do, the memory craps out at about the mid 700's. If I use a different divider than I can get much higher...yes weird.
I can also leave my memory at really low settings and run the CPU up higher than the current 333 FSB. What I can't do is run the processor above 3.0 GHz and the memory above 667 FSB. If it's not in the memory and not in the processor than what the heck can it be.
Another question for a few of you. I am using Orthos to pound the memory and CPU. I can set it to hit only the CPU, or the blend test. Can some of you point me to any software that may help me test my system better. I currently use Orthos (as said), Speedfan to watch temps and CPU-Z. What is the program that allows me to slowly raise the FSB while in windows?
I checked out the article on the pencil mod. I work in electronics and I still am unsure what they mean by the "pencil mod". I'm probably missing something obvious. In any event I doubt I would want to change out the resistors for different values, or installing a potentiometer for better accuracy. I can already get an MCH of 1.85 via the bios. I've ramped up most of my bios voltage settings and my system still seems really reluctant to play nice above the 333FSB level. Once I get past there, all reasonable expectations go out the window. A real pity when others are hitting so much higher. I've done all the AS5 on all the different SB/NB/CPU with no luck. It's always possible I have a dog of a cpu but I still would like to believe there is something I am missing.
It is a known issue that to achieve higher fsb on the 975x chipset that in some case SPD must be enabled. I have a similar memory to you and this was the case for me.
Anyway, try two sticks of memory, this might be holding you back.
Also, have you tried the settings I listed in a previous post?
Hi all, thanks for a great thread - just finished reading it. Just one thing that interests me that I'm not quite clear on regarding bios's later than 1602 locking the extreme chips - are those later bios's limiting the upwards multiplier to 10X, or is it 11X? For instance, I am currently on bios 1707 ( with E6600), but hypothetically, if I got say a QX6700 or X6800, and I know that my board will run currently at 410 fsb, with either of those chips, can I run 10X410, or 11X410? Or is the upward multiplier completely locked so with 1707, I would be limited to 9X410?
Question 2 - is anyone here using a Lightspeed phase unit to chill their cpu on the P5WDH? If so, how is the fit of the clamshell to the board? Does it clear the capacitors ok? Or will I need to modify something?
Regarding the speculation by some of you about degradation of the cpu/mobo, I concur with that theory as well, as 6 months ago was stable @ 3.6Ghz & 1.6vcore, but now need 1.625 vcore for the same stability, so I guess I've slightly bbq'd the cpu with too much vcore, stresstest, & heat. When I say too much heat, the most I've taken it to is 75C coretemp with dual Prime 95, but on average, the stresstest runs up to high 60's , when it hits 70C now , I stop the test.
It harder to overclock with 4 sticks of memory - they put a greater load on the memory controller and the motherboard and that will limit the overclock moreso than 1 pair. Most advice has been given from people with 1 pair who like me assume it will work as well. Have you checked all the memory out stick by stick with memtest86. I would start with 1 pair and see how high you can take fsb 1:1 with memory. Also disable virtulisastion and memory remap - if you dont use it disable it!.i Dont forget to puit a 40mm fan on the NB - big difference! Pencil mod was for someone with an early revision of the board where it only has 1.65 on MCH. You could try clockgen for uping FSB in windows.Quote:
Originally Posted by AHolland
Non Beta Bios 2004 was released today
ftp://ftp.asus.com.tw/pub/ASUS/mb/so.../P5WDH2004.zip
ftp://ftp.asus.com.tw/pub/ASUS/mb/so...20DH%20Deluxe/
The beta version of 2004 didn't work all that well for me, so I'm gonna let you guys try it out first before I jump on it. Please let us know your results and if there are any changes like more memory sub-timings.
Same here. With four memory sticks and SPD enabled, my machine keep crashing randomly. Also tried to boost vMCH and/or vMEM but no luck (now vMCH is configured as AUTO, vMEM as 2.1v and no problem whatsoever). But, unfortunately, this compromise my previous E6600@3.5GHz (9 x 390) and the best I can get now is 8 x 400 allowing me me to run my memory 1:1, which looks better for me. Also, tried to run the same four memory sticks as 3-3-3-10 but the machine even boot.
Quote:
Originally Posted by AHolland
X6700 would be unlocked between 6-10 and the X6800 would be unlocked between 6-11 but not higher than that unless 1601 bios or lower.Quote:
Originally Posted by nifty
So did anyone check the non-beta 2004 bios yet? Or is it just the beta bios renamed?
Its the finalized version of the BIOS. Didn-t tried it out yetQuote:
Originally Posted by miochza
Pretty sure it's identical. 2004 was always only on ftp, now they have just updated the site to show 2004 but nothing new.