The Maximus Formula Thread - Help/OCing/Guide/Rampage conversion

[Nuckin_Futs]

I remember somebody asking of, or spaking of USB issues on certain BETA BIOS. I am experiencing issues with USB ports off one of the hubs from mobo internal on a duplex cable and it is so far on only one of the 2 (ie the top is fine well the bottom drops in and out through the day). I only started since testing on BETA v1004, and I noticed when using iPod docked while playing MP3's it continuously droped out and back in in auto play throught the day and I moved it to 2nd of the 2 ports on case hub to mobo and it is fine. I will test for another set of onvboard headers to confirm, but it may be a bad connection or wire. It is on my 2nd case, Antec P-160 w of almost 5yrs old.

[loften]

my board was acting weird on the rampage 308 bios, it ran good and clocked up for a couple days on it.. just enough to fool me. then it all went to crap. it wouldn't even run on auto setting and pass 1 pass or at any setting on any cpu stress or even pass a 3dmark test.. the usb was going in and out. i am guessing it fried my mouse. also a cd/dvd drive went wacky and my sound. it also went to 2.0v twice after setting it at 1.50v. it was like someone/somthing outside my control was opperating it.. i went back to 1004 and started everything over again for the hundredth time and it is running pretty much what it did the first few days on the rampage 308 bios.. every once in a while it crashes out in a game but it runs good untill it does.. as soon as i save up enough money i am going to get somthing other then this board it is not for me.. then i am going to take it out back and pound it into the earth.. it has been the biggest piece of crap 3-4 month experience i have ever had.

[binormalkilla]

Ok so I flashed the BIOS about a week ago and all is well. I'm stable at 475x8 so I'm pretty happy. THe only thing I'm trying to get is my bandwidth back where it used to be.....I got around 9700 MB/s with the Maximus, and now I barely get 9000 MB/s. I haven't really explored the CH pull off settings yet.....
Anyone running a quad at 3.8-3.9? If so, what sort of memory bandwidth did you get in Everest? I got around 51.5ns also...


I'm also going to reinstall Vista x64 here in a few days. Are there different chipset drivers for the X48? I'm assuming they're the same generic compilation.....

Here is the comparison from Maximus to Rampage. I tried setting the clock twister to strong and stronger to see the affect.....not very much. I had the Maximus set to:
DRAM Static disabled
AI Clock Twister: Strong
Trans booster: auto
Here is a major increase with the posted settings notepad. I'm going to test for stability....I'm skeptical.

[beard]

Have you tried manually setting trans booster to 8 or even 7?

Don't know what it sets on Auto because i can't hit 475x8

EDIT: oh sorry didn't notice the 3rd screenshot...

[Jupiler]

Time to step in on this early friday morning.
Some posts (mostly crapping and flame baits) have been removed.

Keep it on topic and no more thread crapping.
Thx.

[Dostoyevsky77]

(SNIP!)
To the guy who is gonna' RMA is hacked MF: is it stuck on RF BETA, or were you able to get back to MF for return? Is or has anybody been hassled from ASUS for flashing to RF then back to MF for an RMA? I mean can they really tell, are there block traces left like BETA v0907 left.

Oh, hey. Thanks. I was actually able to get it back to MF. My problem was that I was using a non-ES version of AFUDOS. Both versions that Kup has linked on page 1 of this thread work great. I don't know where I got mine from!

[binormalkilla]

Have you tried manually setting trans booster to 8 or even 7?

Don't know what it sets on Auto because i can't hit 475x8

EDIT: oh sorry didn't notice the 3rd screenshot...

Yea I posted that one after rebooting. I ran Everest stress on CPU, memory, cache, and FPU........it locked up some time over night. I know it ran for at least 15 minutes though, so I imagine I'll be stable with a lower FSB and tighter tRD or by setting AI Clock Twister to strong.

[beard]

Guys, when flashing between bioses (which is of course kind of desperate), remember to clear cmos properly - i mean turn off the psu and press ccmos for at least 15 seconds.

I just wasted 2hrs because i forgot to do it.

Of course this is like abc of overclocking but when oc gets boring enough (with this board and cpu it definetly does), you tend to do such stupid mistakes... Besides, think about all those times you thought it's a no go when in reality it was just a SUE (talking about myself)...

[Tony]

clock skew will not improve bandwidth...it will allow you to tune dram clocks higher thats all.

With me not touching skew the max was 1040, play with the skews and I hit 1200 for benchin and 1180 stable for 24/7.

[itznfb]

Guys, when flashing between bioses (which is of course kind of desperate), remember to clear cmos properly - i mean turn off the psu and press ccmos for at least 15 seconds.

I just wasted 2hrs because i forgot to do it.

Of course this is like abc of overclocking but when oc gets boring enough (with this board and cpu it definetly does), you tend to do such stupid mistakes... Besides, think about all those times you thought it's a no go when in reality it was just a SUE (talking about myself)...

that's why you use switch /pbnc when flashing

[Dostoyevsky77]

clock skew will not improve bandwidth...it will allow you to tune dram clocks higher thats all.

With me not touching skew the max was 1040, play with the skews and I hit 1200 for benchin and 1180 stable for 24/7.

Do you have a link where I can learn more about this? I read a good article on the P5E at AnandTech, but it didn't go into enough detail.

The way I understand it is that clock skew places less demand on the memory controller by delaying or advancing one channel slightly to compensate for the difference in physical distance between the two banks. Is that anywhere CLOSE to true, or am I completely off the mark?

[Brama]

Do you have a link where I can learn more about this? I read a good article on the P5E at AnandTech, but it didn't go into enough detail.

The way I understand it is that clock skew places less demand on the memory controller by delaying or advancing one channel slightly to compensate for the difference in physical distance between the two banks. Is that anywhere CLOSE to true, or am I completely off the mark?

I am interested too how to use and set properly the skew parameters.
Can you help us Tony?

Thanks!

[Coldfox]

Need some advice, im currently using this board with a Q6600 @ 3.2.

I'm using the bios that it came with atm, (0802). Is it worth me flashing it to the rampage bios?

Thanks

[cadaveca]

I am interested too how to use and set properly the skew parameters.
Can you help us Tony?

Thanks!

Timing for memory bus and FSB are different. Skew allows adjustment oftimnigs within this ratio so there are more "hits" and less "misses".

for example, 4:5 needs less skew than 3:5, as the ratio is closer. Compuns this by 1-mhz adjustments in FSB, and skew may need to be different for one IC to the next, as each behaves slightly different, Hence skew for oen set may not work for all sets(2gb dimms seem to need more skew than 1gb dimms). You need to eb able to kknow acurately how your mem is erroring to be able to adjust with confidence, rather than blindly.


As an aside, 2gb Micron DDR2 and X38 don't work well together. 5-5-5 won't work abouve 1066mhz or so, but 5-6-5 does. Skew makes no difference.

[beard]

that's why you use switch /pbnc when flashing

That's what i thought too, but it seems you have to physically destroy system cmos.

With previously stable settings, after flash, it freezed in the very beginning of loading windows. In some point i realized i may have to ccmos (i usually do it). I pressed it for a sec, loaded same settings - it loaded but freezed in vista logo. Now i knew why it was ustable, i pressed ccmos for a long period, loaded same settings - voila! Still occt stable.

Don't blindly trust such things.

[itznfb]

umm... you flash with /pbnc. turn pc off. turn pc back on, and you're starting from scratch with a default bios. you don't need to manually reset unless something went wrong.

[beard]

umm... you flash with /pbnc. turn pc off. turn pc back on, and you're starting from scratch with a default bios. you don't need to manually reset unless something went wrong.

Whatever Do it your way

I have my "known to be good" settings on camera too, so after flash, i just enter exactly the same settings. So there was no error between flash and the situation i described. This is what happened to me, maybe it was just crazy (un)luck. I'm not an expert, or even advanced with this board. I can't justify my opinion better than in this post, so you can do it either like you said, or with additional ccmos, it's your choice

[binormalkilla]

Well I'm going to test with the clock skew and pull off settings tonight.....more results soon.

[Zucker2k]

Got myself 2 kits of this G.Skills: F2-8000CL5D-4GBPQ, currently running at DDR2 1066 with 2.1v. 5-5-5-12-3-52-6-3, PL 6. I had to bump NB from 1.45v to 1.68v to drop to PL 6 1066 from PL 7 1040. I'm in a warm room (80+f).

Edit: The overclocking headroom on the 2x2GB kits is a shame.... this is what happens when company employees get too involved in the OC community. This actually has its advantages too, eg. Shamino and Tony, to name a few. But the current 2x2GB kits (going by oc headroom) should be rated at DDR2 800 ( yeah right!). But seriously, Tony took a DDR2 1150 rated 2x2GB kit to DDR2 1180 (a 30mhz overclock!), and that was an achievement. The 50mhz overclock, ie. DDR2 1200 was only good for benching! ROFL. Don't get me wrong, for value, these kits are awesome, but oc is a joke.

[binormalkilla]

Ok I'm 32M superpi stable now with performance level 7.....I'm going to test with clock skewing later. I only dialed in my main timings to 5-5-5-15......

[Dostoyevsky77]

Whatever Do it your way

I have my "known to be good" settings on camera too, so after flash, i just enter exactly the same settings. So there was no error between flash and the situation i described. This is what happened to me, maybe it was just crazy (un)luck. I'm not an expert, or even advanced with this board. I can't justify my opinion better than in this post, so you can do it either like you said, or with additional ccmos, it's your choice

I have to agree with you: there is no harm in resetting CMOS after a flash, if it's necessary or not. Even after I use /pbnc, I will sometimes get DETRAM unless I reset CMOS. Regardless if it has anything to actually do with resetting CMOS, I do it now anyway.

[WFO]

I have to agree with you: there is no harm in resetting CMOS after a flash, if it's necessary or not. Even after I use /pbnc, I will sometimes get DETRAM unless I reset CMOS. Regardless if it has anything to actually do with resetting CMOS, I do it now anyway.

DFI has always recommended a cmos clear with the battery removed after a flash.

[Nuckin_Futs]

Need some advice, im currently using this board with a Q6600 @ 3.2.

I'm using the bios that it came with atm, (0802). Is it worth me flashing it to the rampage bios?

Thanks

I'd say if it works for you, and is stable, stay there. Most of the BETA's can often bring more bugs with them then they fix. 602 was the best in my experience as far as the final's go. 907 was bad for me but , v0903 was bewst of the BIOS in terms of daily use. v1004 offered better RAM loosening but S3/STR no longer worked. If we can get soem more real world insite of 1006, I may try it too.

Don't blindly trust such things.

I hear you there. It only takes an extra 15 secs. to clear out caps of stored energy. I also get the hangs on the BETA BIOS' above 0903 (it was the last colse to stock thatr didnt do it). Even 1004 did that untill I had to clear one more time but it's back up and running fine. But in 907, when it crashed, it crashed hard and took many trys of clearing and battery pulling to get up and going again even if just to POST long enough to flash out of it.

So is it so then, BETA BIOS v1006 cannot be flashed with any version of Afudos? I was told by many ASUS BETA testers not to use Live Update cause of some sort of curruption like in 907. Somebody confirm?

[binormalkilla]

All right I changed the skew to 300ps on both channels, same timings, stable 32M.

[ZenEffect]

All right I changed the skew to 300ps on both channels, same timings, stable 32M.

why? skew doesnt do anything for bandwidth. if you were stable before (superpi is not a stability test...) then what is the point? it becomes useful when you ram has hit a wall and you want to go a bit faster. tony explained this very clearly.

shouldnt 475 fsb be netting better bandwidth results? what is your common performance level?