First off I want to say thank you to flipfire (and whoever helped him.) It was his guide that I used to make this all happen. Also thank you to Groove65 that reminded me to get off my lazy butt and continue testing. Ok so now the part you want to hear about so I read through his whole guide twice before I did anything. Now I started following his steps. I tested each VID lowering and lowering it each time until I finally found the one that works! If you go any lower than 1.0375 the cpu will not work right! I have ran mine for one hour and 20 min now and I have not had any downclocking the hottest i've seen is 89c. Now Alienware cause I know you'll read this in the morning tell your product group to make us a new bios that has the cpu voltage lowered and the fan speed a little higher (with the old bios max was 6400, x32 bios is about 6300) and we'll be ok. Hopefully some people can help test and we can get our settings just right and then AW can hopefully give us a new bios so that we won't have to use RMclock. I'll post pics later today...
EDIT: Sorry here is the guide http://forum.notebookreview.com/showthread.php?t=235824
-
-
as you stated. thats a temp. solution, because that lappy should be able to handle its heat by itself, but still. thats a good effort from you to fix your problem by your own.
-
If this is a permanent fix (i.e. holds up for hours) with no loss in CPU power, then this is the way to fix it. -
Cool... That's a nice workaround. How does that impact on CPU performance though?
ALso... this is obviously not a solution that we should be happy with. AW has to come up with some sort of solution and should work on a way to allow extreme processors to be overclocked. Obviously their bad cooling solutions (that they were always aware of) put them back to allow people to use the normal feature of the extreme processors to be overclocked. -
it will never drop the CPU clocks, just watch it for BSODs which will most likely not happen with those Voltages, now i think its GAME TIME tacomenace21
-
Well i'm running Orthos and its going at full speed. If you go under 1.0375 it will affect performance but i'm still testing to make sure.
@GRB If they let us swap out the heatsink/fan I would be cool with that but I would be upset with sending it in. -
Ok, well I'm generalizing a bit obviously but I'm thinking most wouldn't want to send them in. As you say, you'd be upset (still willing to send it in, yes, but there will be plenty of other people who'd also be upset and not willing to send it in)
-
Excellent work taco!
Now I wonder how many people in the AW product group have been refreshing this forum daily, waiting for the members to solve the CPU downclocking problem so they can incorporate an official fix for us. Sigh... -
lol, so true....
-
ok guys update the better number is 1.0500 after I ran the test for awhile the test failed upped the VID now everything else is ok if anyone else that had downclocking can test that would be great.
Edit Ok guys after doing some more tests. I think AW should put us right between 1.0500. 1.0500 is not a perfect fix because it also gets hot enough to reach 93c and 1.0375 has problems with performance. Hopefully they are working a solution cause it has been weeks and we haven't heard anything. Still doing some testing guys to try to get the perfect number but we are on the right track -
tanks for testing but every CPU has its own voltage values. Some need more power than others. Intel solves this by using a high voltage value in order to provide enough power for all CPUS. Therefore another CPU might be unstable at 1.0375 volt while some might even allow further downvolting.
So I would say everybody has to look for his own specific voltage value.
saint -
thanks saint i agree with you and I think 1.0375 is to low for the x9000 If there was anyway I could make just a little higher that I should be ok.
-
hey guys i am a noob but what does down clocking mean? why would you want to down clock your cpu power?
-
well what you don't want is downclocking. What happens is when your cpu takes on a full load and goes to 100% (some not all) the cpu will get to hot and when it reaches a certain temp (93c in my case) it will lower it speed to cool down so that way it won't fry it self. The reason don't want that to happen is cause your cpu will go from 2.8 to 1.0 for example. What we are talking about is undervolting which is providing less voltage to cpu so that way it doesn't get as hot. We have another thread that explains how to test your cpu for this issue.
-
Just a quick question taco, how much time do I have to test to be 100% sure if mine is downclocking? I have a 2.5 one and I did a 1hour test. But now I'm not sure.
Hope AW relieases a fix quickly, if no I'll follow the instructions you provided. Oh, and grab a rep for an informative threat. -
1 hour should be enough.
-
Read that threat and I think I'm going to undervolt it anyway. Saves power and less noice
EDIT: I hope to try undervolting in a couple of days, feeling laizy nowwill post the results.
-
yeah one hour is plenty. Good to see yours is working.
-
id rather try gaming over Orthos though, this is where the real test is. i could do orthos on .9 voltage with my Merom CPU, but i had to increase it to 1.1 v cause it does BSOD after 3 hours of gaming
-
i'm just wondering what game where you playing that would give you BSOD after 3 hours?
-
While gaming you will never stress the CPU as much as with Orthos (the same goes for temperatures). While playing I never saw the CPU going even close to the 92C I was getting with Orthos.
-
ok guys I've been testing all day and I have found my most stable VID 1.0625. I tested it for an hour no problems not 1 time did I hit 90c and it was the most solid. I will test again tomorrow (need to give it a rest). Please remember if your doing testing always watch your temps! don't just turn it on and walk away.
-
Does anyone know the absolute thermal limits on the CPUs? I read somewhere that someone said the absolute limit on the 8800M GTX is around 135 degrees C
-
according to intel for the x9000 its 105c
http://processorfinder.intel.com/Details.aspx?sSpec=SLAQJ# -
Hey guys Just been running some tests in Everest and I have noticed that lowering the VID does result in lower performance but the difference is minimal I ran every cpu benchmark and noticed that it was only difference 50pts which is not even a huge deal I would rather lose 50 little points then have my cpu downclock. CPU Queen results normal 11814 modified VID 11763
-
well then you may have the voltage just a tad too low for your CPU. as others have said downvolting shouldn't lead to any performance loss.
-
well the problem with that is if you go any higher that that the temps go way to high average around 91-93
-
-
i hav this downclocking problem on my x9000 and its at 83c then it starts to downclock i tryed undervolting but this just results in my temprature riseing to 89c (at any voltage) then my laptop shuts down. tacomenace21 could u put up a screen shot of ur RMclock settings thanks
-
Oh man... it just gets better then!
-
-
nar i'm using wat the guy said to use in the undervolting guide hwmonitor and orthos and then rmclock to change voltage
-
o ok well there is are difference I use thermofan to monitor temps which its temps are a lot higher than rm clock and hwmonitor thats why your saying that your downclocking at a lower temp. (by the way thermofan is preinstalled on your m15x)
Here is a pic of my profile. Let me know if you need a pic of the other options.
-
masterchef341 The guy from The Notebook
alienware can't release a fix like that...
each processor is unique, some are built better than others, and as a result, different physical processors (even with the same model number) will be able to take different amounts of overclocking and undervolting. its up to you to undervolt your own machine if you want to do that. -
AW will only fix it when they collect all systems and install new cooling systems on them. That's the only way they will fix it.
Why do you think they are not saying anything? This is not like the GPU... Simply put, the fans can be always on that they will still be not enough to keep the CPU temp down.
To put a long story short, M15x cooling system is completely inadequate and not enough to the power it packs inside. -
Technically speaking, voltage can no way alter CPU performance. Im guessing its RMclock's cpu throttling/management for the x9000 is causing the 50pts loss in the benchmark. Its not a big deal.
tacomenace: I dont think your default voltage was detected correctly but anyway you can set all your multipliers to 1.1625v . If your highest multiplier is stable at that voltage then any multiplier lower than it can also run at that voltage or less.
highfly: If your temps are higher than before, then you must have missed something. -
well flip now that you say that I looked in cpu-z my normal VID is 1.313 and when I run with rm clock at 1.625 cpu-z says it running at 1.275 I guess your right, rmclock must not be reading the VID correctly. Thanks for your help by the way. Now I have no downclocking!
-
haya sorted my problem my cpu fan power cable was not pluged in (my stupid fault) might try undervolting but it runs a lot cooler now
-
yeah, that oughta help!
"Well, there's your problem!"
M15X downclocking SOLVED!!!....Sorta
Discussion in 'Alienware' started by tacomenace21, Jun 3, 2008.