M17 Underperformance issues

If you have the OCZ or Z's BIOS installed, you press F2 at boot up to get into the BIOS. From there you change your multiplier. SetFSB overclocks the whole system, and I think you have to re-load the profile at every boot up. Once you hit a wall, you could start messing with your CPU voltage. That's if you know what you're doing.

If you don't have the OCZ or Z's BIOS, yes, you're stuck with using SetFSB to overclock. And you won't get too far with it.

I'd say 500Mhz isn't out of the question. 250Mhz is very safe.

If you're going to start overclocking things, WATCH YOUR DAMN TEMPERATURES! HWmonitor is the program for that.

 

i have HWmoniter and Z's bios, so 250mhz incremints are safe? and make sure temps dont hit danger zones

 

I don't know if you can overclock an non-extreme processor with your BIOS. I totally forgot to mention that.

Hit F2 on boot up. You'll enter the BIOS. I think it's the 2nd tab over, you'll see all of your CPUs multipliers. Wherever you're currently at, go up by .5. Hit F10 and save. Then boot into Windows and run a stress test like Orthos, Prime95, or equivalent while you monitor your temperatures. If the program doesn't stop with an error and your laptop doesn't crash or BSOD, you can try the next multiplier. And make sure your temperatures don't skyrocket.

If you don't have the overclocking option, you are stuck with SetFSB. Which sucks.

 

ok thanks ill check and well see what happens

 

well if i knew exactly what the root cause was... its so hard to say because a fix that works for one person doesnt work for someone else. know what i mean. i still think its heat releated myself but some have tried the mod and it didnt work. though i didnt physically see thier mod so i dont know what copper they used and if it was flat or what else was involved know what i mean? if i knew it was bios related i could prob fix it. i see no cause for it though and ive been through the bios with a fine tooth comb so many times. i can up the nb voltage a bit more but when i tested that on my whitebook i used to have the nb temps started to get way out of hand and it sttutered even more after that which is what led to my initial thoughts of it being heat releated. as soon as the nb cooled down those times the stutter went away

 

I'll tell you this, Z. I just acquired an OEM X9100 to replace my QS QX9300. The X9100 doesn't hang in 3D Mark, but it runs at 3.33Ghz stable. So I'm getting better frame rates in all of my games. And the X9100 never stutters. I have even purposely tried to get it to. It just won't. So yeah, I can't boast that I've got a quad core CPU in my laptop anymore. BUT, I don't ram my race car in GRID into a wall because my frame rate drops to two frames per second.

I haven't got any idea what causes it either. And the fact that Alienware has even given up on it makes me think the answer will never surface.

 

yea, I think there will be no answer for this cuz, when I'm using Q9000, it stutters in every game, even when i watch movies or youtube. Then, I upgrade it to QX9300 and overcloekd it to 3.06GHZ and the stutters goes off, there's no stutters at all!.... everything just run smoothly....

 

Exactly. My QX9300 was a stuttering machine. Nodeffect is having 'no defects' with his quad.

 

LOL !.. thats kinda hilarious...

 

my qx9300 on my ocz whitebook hardly stuttered. i saw it only on a very rare occasion.. so its wierd some see it and others dont. im trying to get some info from flex now to see if i can make a .29 bios. im still waiting for them to get me the info i need for some things. they dont seem willing to help much though if at all

 

after 1 week of using my new QX9300, this is the first time i;m having stutter. Then after I restart my pc, the stutter is gone. Don;t know what causes it, it's very hard to find out. I gave up looking the at cause now, I think Im satisfy with this condition right now as it does not stutter as much as before, its like 90% better now.

 

If you get a stutter, just unplug your laptop from the power, wait a few seconds, and plug it back in and it should have stopped.

 

Z i dont know if u came across this but there is a 0.29 bios out at ASI mobile the company that took over the Whitebook. I even tried to get my hands on but never gotten a reply on my email that said i wasnt a ASI customer. If anyone else is intrested in trying here is the link http://www.asimobile.com/
Just goto support tab then when u have to choose the notebook, take the IQ17.

Also im pretty sure the main stuttering is comming from DPC latency. I been monitoring this since i stumbled over the DPC thread on the m17x and i have to say im getting loads of em when playing movies or music or even games. Its not like i can tell for sure that when playing movies ill get it, its rather random sometimes i got loads and sometimes watching the same movie its all green.

 

Can someone teach me what is DPC?
I thought DPC is only a checking program/term for stutter or response rate of the laptop

FPS says his stuttering problems come from DPS latency got me confused.

 

No idea on the DPC end =/

But i just removed AW's stock thermal grease and put some Arctic Silver 5 on both the GPUs' CPU, and removed the glob of 'play-do' they call a thermal pad and replaced that with AS5 as well. idle temps seem to be about 5C less. in CPUID is the north bridge labled as TZ00? or something else. no stutter still yet after updating these things

and out of shear curiousity .. does thermal grease expand after awhile? >.> cuz alienware put a heck of a lot on compared to what AS5 says..

 

AW uses thermal grease?
I thought they only use pads.

 

They uses pads, I dont see any grease on the CPU or the GPU when I open it up.

 

Im no specialist on this matter but from what i understand DPC measures the time your drivers use to do stuff. And normally they shouldn't take more then 500 ?? microseconds?? So due to bad drivers or a badly written driver u get DPC latency.

Please if someone who knows exactly what it does reads this and this is completely wrong say so so i can delete this post.

 

it was a beta bios and was not officially released i do have it though but it needs work

 

Greetings, all
(long time lurker, I've been working on this problem at home also)


@ the thread

I'm in Z's old boat (or more accurately, Don's, because I bought his whitebook, thanks Don, it is... 90% awesome.).

History:
I've modded the the northbridge and the plx chip heatsinks using 0.5mm shims that I rolled from stock copper replacing don's old hammered-out-copper-pipe-chunk (my g/f is a metalsmith, her shop is in our basement). I've bricked up the video card sinks with filed ram heatsinks. I've added additional copper sinks to the northbridge section of the shared heat sink. Don cut some pretty savage holes in the bottom cover which I managed to clean up and expand with a jewellers saw and file (over top of the intakes). I've also switched from the ocz bios to zfactor's custom bios (which is actually somewhat more stable, thankyou z).

Results:

Northbridge temperature is in lock-step with the processor cores. While gaming, all temperatures including the video cards sit at 72 with little variance. During a full-burn prime test (small FFT), temperatures fluctuate on the cores between 85 and 90, northbridge sticks at 90. While running the prime test, there are no stuttering effects (sound is fine, mouse is fine, can surf the internet or run programming IDEs etc).

-- these temps/results are roughly the same at stock settings as well as overclocked settings, using the multiplier or using the fsb

As a side note, using z's bios the fsb's max stable setting on my machine is 1208/302 using 1333 ram (so actually underclocking the ram somewhat... my background is in desktops so I don't actually know of a way to change the ram timing ratio on this thing)

When running games however (fallout 3 as a test case) it will still periodically crank down to 2 frames per second with the sound falling apart. Pulling the power and plugging it back in resolves the issue until when next it comes up (average is 1 time per hour)


so, heat is in check and with these mods in place does not seem to be a factor (given that the prime test makes the machine significantly hotter than gaming and does not send it into stuttering fits).

Were this a desktop with a decent board, i'd suggest a meatier power supply and minor voltage additions to the north bridge and south bridge, however as Z said (paraphrasing here) laptop parts are delicate flowers that don't take well to ham-fisted-volt-driven solutions.

the fact that the "pull the power and plug it back in" solution seems to work across the board, combined with the x9100 having the same voltage profiles but not suffering the same problems, and that many have complained that the quads will stutter during normal windows use.... leaves me ENTIRELY AT A LOSS.



@Tubod

Now that you've had your duo chip for a few days, what can you say about the general system performance and stability versus the quad? Can you recommend a seller? Know anyone with an m17x who wants a quad chip?




Cheers,

 

thats obviously not true now is it

do you think we all a bit silly in here or something

 

Maybe not =/ i thought that was grease on the CPU and GPU cores, the GPU has a few thermal pads on something else, but the play-do looking stuff looked like cured grease to me.

maybe not =/

@Nicu15 : if that even is legit, you probably need to guess the single 8 digit number on the single AW laptop their giving away...

 

i have a freinds m17 here i have apart now. ive been taking some voltage readings for the last 2 days on it. he has a qx9300 and dual 4850's and he has bad stuttering ( we are using the aw bios for testing reasons but i will flash my newer .29 test version soon) i was going to make him a shim but decided to test his system out for a few days since he left it for the week with me. so i waited and waited. (i have the system running out of the casing lol so i can access the motherboard) finally i got stuttering. i quickly grabbed my fluke and let 3dmark06 run and continue to stutter and started testing points on the board. i had to repeat a bunch of times to test many areas since the stuttering didnt always last long. i am finidng some curious things. for one imo the power section of the mobo does not seem to sufficient for its design. its voltage is dropping considerably afterwards on the board during extra stressed situations and doesnt seem to be able to recover fast enough to keep up. i saw voltage drops as low as over 1v variance!!!! remember this was very quick drops and they were not a steady reading. for some reason the power section seemed to cut power then run normal then cut etc and this happened very quickly. thus the reason pulling to the battery for a second works. it allows everything to downclock for a second reducing voltage requirements till they build up again. now its not a totally bad design it just doesnt seem to be able to output enough at times for the extra requirements but note its not a great design either i have seen far better on most desktop boards. i also noticed things like usb and power for the hdd are also reduced slightly during this time sometimes dropping past the allowable % rate i saw the usb port in question down to almost 4v at times!!! okay but what was even more odd is i paced a thermal sensor under both the nb and the cpu. during the stuttering times the heat WENT UP on the nb for a few small peaks they were very quicks jumps that hardware monitor and core temp could not detect but using my test equipment i was able to pick up on. i saw jumps as high as 8-10 deg on my scope.

now for another oddity while on battery the voltage was jumping all over the place again i dont know that hardware monitor would see this but i was using my equipment to watch. something was causing it to kinda go haywire during the stuttering. im wondering if the power brick design has some fault. i will further test that tomm by itself with a dummy load on it

im going to modify his power section of the mobo this week and beef it up some. im also going to set up one of my showroom power supplies (large inverters i use to display 12v equipment) instead of his adapter to make sure i have way more than enough current going in.. im also going to upgrade everything to better caps coils etc as much as i can with the parts i have to make sure everything is solid as it can be

im going to see if this helps (he has a spare mobo the tech left with him when he had it replaced one usb port didnt work on it but the tech that came to fix it left it for him and said he didnt need it??? so i have that one to work on) here's hoping i can figure this out. though a mobo mod is way out of most people's area.. i will foward anything i find to flex and also to my dell contacts (even though its the m17). if this works my flex contact said they would like the mobo to be shipped in for them to see my findings also. again this is mainly for testing purposes and he still has his stock mobo to go back to if this doesnt work but i WANT to know what the heck is going on..

ill keep this updated with my findings

 

+REP.

These are interesting findings. Do you think that reducing the power consumption of other components would do anything to reduce the problem? For instance, don't SSDs consume far less power than mechanical HDDs?

 

yeah thats an idea i have also so im going to actually remove a few things from the board all usb ports etc and use a ssd for testing as well as 7200 500gb hitachi.. im really want to pinpoint this its been just ticking me off since i had my whitebooki have been in contact for about a week with someone i know who also wants to know my findings at another company..i do think though this has a lot to do with why the qx9300 stutters more than others. now i raised up the voltage as far as i could in my bios and i may be able to go a notch higher with .29 if i can release it. i have to have permission from flex since this is a beta bios i would be modding to release so it may never go out but i have now asked them to be able to work on it.. we'll see. with my bios i have adjusted some voltage settings so i know that is why there is much less stutter with mine than the aw bios i have been inside thiers and their settings are all over the place...just weird they way they did it.

could also be why a while back some people found using only a single stick of ram helped with the stutter in some machines... hmmmmm

i also can say i have seen different vendors parts used on different mobos now. for instance his board has different manufactured caps and also chokes and a few mosfet's are also of a different brand his original board was from the summer and the new one he just got a couple weeks ago and i did the swap for him after the tech came and scratched his head about taking the m17 apart lol.. so he left the parts and i did it the next day, so this could also be something to look at maybe certain parts used on different revision boards are adding to this..

 

I'm curious does this stutter problem exist when the battery is removed?

 

im not at my shop to test right now maybe turbo can check tonight ill be back at my shop monday morning unless i can find some time tomm.

 

I never thought about it before, but I remember several people with sagers had a stutter problem while gaming. The cause was never found out, but the stuttering vanished when the laptop was run without the battery in. Course thats not a fix but it might help diagnose the problem if the same thing happens for the m17.

 

@Z
Look forward to hearing about it man. +1 rep for busting out the fluke.
I'd totally do a mobo mod for caps and mosfets if it came down to it, so long as i had pictures and serials to go by.

 

Reporting in, my m17 does not have stuttering issues with Alienware B14 bios, it is stable and all green bar in DPC latency check. No stutter at all. However, when I switch to Z's .28 BIOS, i start getting stuttering and it is unstable even in stocked speed. However, with the Alienware B14, it is not possible to use Intel XTU which is in my opinion a very good tuning program. It's just for me or did anyone else try the new B14 BIOS from Alienware?

 

again did you also do the keyboard firmware with my bios?? very important, for almost everyone who uses my bios their stutter gets better or is almost gone.. so you seem to have the opposite effect..has anyone else had the same result as shinj?? the bios is tweaked about as much as i can without breaking it. i have the .29 beta but i dont know if i can release it i have to talk with them first and it is a beta and last time i tried it it was not nearly as good as the 28 version... they dropped the work on it once all the company's but alienware stopped carrying the system honestly ocz was really the only one pushing for anything to be done with it

 

I use the 0.28 bios without flashing my keyboard BIOS and however, my M17 is still working fine. Is it that important to flash the keyboard also? what does it do ?

 

Ya Z not so say your BIOS is not good, your BIOS is absolutely the best in overclocking and you've put all the effort of making them better. Just that I don't quite understand why I have the opposite effect. Ya when i tried to flash the keyboard BIOS the screen told me that I have incompatible board or something I could not remeber

 

well it could be the newer aw keyboards are not compatible its okay to not use it but much better of course if you can. it will be fine either way though i mean without the kbc

 

i will get stutter even when running on stocked using .28 bios. weird!!

 

yeah it is you are having the exact opposite of what everyone else tells me...just curious did you try a fresh install?? fresh drivers etc..

 

yes everything fresh I even install fresh copy of win7 twice and all drivers, any ideas of what drivers may cause the stuttering to happen?

 

what version of graphic driver u using? and did you mod your M17? apply thermal paste on the CPU and other stuffs?

 

9.11 mob mod from driver heaven and ya I installed a Q9200 and did not try the copper mod on the northbridge yet.Will try it once I had time. But I do not get any stuttering using B14 Bios from Alienware so I doubt the copper modded north bridge will actually helps.

 

it may. remember my bios runs more voltage thorugh both the nb and the cpu..the alineware 14 is set to almost as low as the voltage can be set to for the nb

 

Ok thanks for your advise Z, ordered some shims from K-tron, looking forward to try those in my m17 and hope it solve my problem. Now I'm stable just not can't overclock LOL

 

Arghh I'm having static in my right speaker now. Tried changing BIOS but it wont help at all. Anyone experience the same problem?

 

Is it 'static' as in stuttering that the notebook gets with quad core users, or just speaker static? Try unplugging it from power and then plug it back in a few seconds later.

 

It's the speaker static

 

Continuous or periodic pops?

When you isolate to the left or right speaker (say, using an eq slider) is it only on the one side and clear on the other?

 

Yes it is not continuous just periodically. And it is only the right speaker that make the hissing sound. Have tried all the drivers and different BIOS and it turn out to be the same

 

before pulling things apart, I'd try to uninstall your sound driver, restart, then re-install.

Otherwise did this happen after moving the laptop or did it start suddenly?

 

sounds like hardware but could be the driver but if that was the case it should have been both speakers not just one

 

OK so i have kind of been busy the last couple of days and have neglected the thread =[

Anyways, iv read up on the few pages i'v missed... and Zfactor, i would say i'm pleased to hear you have stumbled upon possible further solutions ^^ it'd be great to finally get this issue fixed for M17 users. Keep us posted and if you do create an updated bios you release to us i will most deffinetly be eager to try it!

thanks in advance for your work - once it gets to this its gettng to be out of my zone of knowledge

EDIT: I'v also been looking into copper-modding the north bridge, is the north bridge directly to the right of the CPU where the CPU fan/heatsink had the thermal pad and what looks like a cpu/gpu exposed core thing? ^^; merely asking because i'v been reading up but cant seem to find a page where this opperation is actually detailed (then again, some posts have 50+ pages so i could have easily missed one)

I did just remove the thermal pad from the NB (if thats what it was, and used AS5, no copper, etc)

12/13/2008 - 8:25pm
got a slight stutter.. issue isnt perminently fixed with .28 bios and AS5... but has SIGNIFIGANTLY decreased the issue. also, pulling the power adapter and putting it back in fixed the issue as suggested. which leads me to agree with you, that its some type of voltage issue or whatever you said it was ;p make that 2 stutter incidences, or lag incidences

PS: Zfactor, seeing as i sought out help from you guys on this issue, as did everyone else, if theres anything you might ask me to do/try let me know, my expertise doesnt come close to what your doing, but its the least i can do for what your doing for us is at least put the offer out there.

 

You've guessed correctly!

Here are the coles notes from the work done by halogod, don, zfactor and others:

you can *sort of* do the north bridge mod without a copper shim (as you've discovered). The shim will make is so that the sink meets the northbridge more evenly (effectively making the northbridge and the chip the same height off the board again) - right now you're likely on a minor angle, reducing the surface area of the contact somewhat.

The ideal is a 0.5mm or even slightly thicker (not thinner) copper square, roughly the width and height of the nb surface. It should be flat, straight, and ideally sanded using sand paper for metal.

Process:

1. apply as5 to the nb (dot or spread, your choice)
2. place the shim
3. apply as5 to the cpu surface
4. apply as5 to the nb contact of the heat sink (easier, as to avoid shifting the shim all over the place)
5. seat the heatsink in one, level move
6. secure the sink, being mindful not to wrench it down too hard - though unlikely, you can crack the die and then all of your m17 problems are over

Imagine that here I say all of the standard stuff about 'don't use too much as5' and 'run some prime tests over the next few days to cure the compound'.