The M15x throttle? The final solution!

would "Disabling" battery via fn+f2 work somehow to "unlock" or draw full power brick output to the m15x enough to NOT throttle 920xm + 7970m? I don't recall reading anything of the sort in this thread. thanks.

 

No. The only thing that works is removing the centre pin from my PSU jack. I had very sporadic success with the 210w PSU with centre pin in place but I couldn't find a repeatable way to always get full 3d clocks. Removing the centre pin fixed this. I never get throttling on anything now

 

A DC female to male connector 7.9x5.5mm(like a little extender) with the center pin of the connector removed would also work, but I can't find one anywhere. One would have been very useful, just use it when you want to game, remove it when you want to charge the battery . Anyone heard of anything like this? If not, I'll just make my own

 

Yo chopsy that sounds interesting. let me know if you pull that off. I find that to be much easier and somehow more convenient than having someone make a switch for my charging/gaming needs.

 

This sound interesting indeed. Shame I already butchered my psu already lol


Sent from my iPod touch using Tapatalk

 

Woooo! Finally got round to some overclocking and benching and managed to break the 7k mark finally (with temps staying below 65 degrees)!
27/27/26/26 on CPU and 950/1400 @1.05v. I think my 7970M is a good overclocker



Firestrike results to follow. God I love the throttle fix! I still watch the combined test with dread.

EDIT: Here is the firestrike result

 

Welcome to the 7k club Now push your vram to 1450 and be happy! I can push up to around 960-970mhz at 1.05V what can yours do.

Many though can do over 1ghz even 1050mhz on stock voltage which is crazy!

 

I tried to play Assassins Creed IV at 950/1400 clocks and got a random shutdown. Not temperature related so I wonder if I tripped the PSU?
Anyway after that I got a bit of artifacting on the screen on reboot so I've reverted to lower clocks and voltage to play games at. Bit scared to push my card too far

 

Your CPU doesn't perform as it should. 25x is already good enough for about 7.4k Physics score, seems like you got some throttling there. At least it doesn't seem like that test run at 26x.

See here: http://forum.notebookreview.com/ali...benchmark-thread-part-2-a-31.html#post8981103

 

More than likely. What TDP and TCP do you run at x25 or x26? That is probably my problem.
Also no volt mod here.

Running 900/1400 on GPU and 26x 105/105 on CPU (throttling still evident), I still managed to get 7k despite lower GPU clock speeds
Whats more this is totally stable for gaming

 

100.100 should be enough to run stable. I find 87W tdp suffices to get me stable at 24x in prime 95. The TDC doesn't seem to make much if any difference so I leave it around 80W or less.

What are your temps. I hope I can be of more help to you

 

I read through a fair few pages of the throttlestop guide and unclewebb indicates that you can set the TDP/TCP as high as you like, the CPU won't actually go up to those values. So I set x26 all cores and 200/200 but I still only managed 6965 Physics score, max temperatures were around 85 degrees. Pretty happy with that tbh and I still use x23 for gaming so not bothered about pushing it harder.

I am tempted to try 1.1v or a higher overclock but I had a random shutdown at 950/1450 @ 1.05v which might have been the PSU cutting out (definitely wasn't temp related). Don't have a power meter here so its hard to tell. Again though no point pushing that high for gaming if 900/1400 @ 1v works so damn well and temps rarely go above 65 degrees

EDIT: Just got a new high score in Passmark CPUMark: 6248

Intel Core i5-4430 @ 3.00GHz = 6,298

Intel Core i7 975 @ 3.33GHz = 6,280

AMD FX-6300 Six-Core = 6,385

Single threaded score: 1397

AMD FX-6300 Six-Core = 1,416

Pretty happy that my 4 year old CPU is up there with a frequently recommended modern mid-range CPU - stock FX 6300 (Tom's Hardware).

 

Your system doesn't seem to get enough power. I bet it's not running at 26x with 100% load during the test, make a TS log and check. Also measure the power draw if you can.

 

So what do you suggest I do to make it run at 100%? You're happy to point out my system isn't running properly, that your's is better, but then not offer any suggestions or advice....

 

Dude, I asked for more information in order to confirm my suspicion about the issue and to be able to provide help and you come up with a reply like this? Well, NBR after all, I'm outta here.

 

You can`t run 1.1V in the M15x. 1.05V is the limit. It is pushing it as it is and might not even be 100% stable at least at high clock rates. Also have to consider that no 2 machines are the same. Mine may have a slightly higher tolerance than yours etc.

Anyways I might be getting my hands on a little heatspring. It could lower temps considerably on the cpu. If it is a success then you could try to get some too! Also with my voltmod in mind (If I ever succeed in getting +100mv to work lol) it would probably be necessary for stability!

 

NBR after all? Cheers for that. We won`t miss you then... you have helped out with advice plenty of times and I respect you for that and the work you have put in for others however I find it pretty mean of you to write off an entire community!

 

What is this nonsense? You mean YOU can't run it is the correct and factual answer here.

 

I assume you can then or you wouldn't call it nonsense. If you can back up your words then I will accept your "correct and factual" answer here. Otherwise I can assure you that ALL the people that I have so far spoken to cannot run the 7970M at 1.1V at a high OC while gaming in the M15x. The card doesn't get the power needed to get into full 3D clock speeds and or if it does remain there stably. Besides the cooling is fairly well pushed at this voltage already unless you do carry out certain improvements to the cooling which I and others have done.

Anyways please back up your words before calling something nonsense. I don't base my knowledge solely on the limitations of my own machine.

 

Hummmmm, I was trying to be funny but ok.
1: Here is your proof AMD Radeon HD 7970M video card benchmark result - Intel Core i7-3920XM Processor Extreme Edition,CLEVO P150EM
2: Since you clearly don't know who you are talking to you will never get much farther than you are now.
3: I will call it nonsense because well, it is. Just because YOU can't do it doesn't make it not true.
4: Just because you brag about your 950/1450 1.05V in your sig does not make it optimal.
5: There are many ways to run a laptop and just because you run yours one way doesn't mean the rest of the world has to follow your lead.
6: You back up your words because as far as i see, all words and no action.
7: Here is another example of clock speeds of 1050/1600 which you can't even set and sit at without a driver display crash error.
http://forum.notebookreview.com/ali...benchmark-thread-part-2-a-46.html#post8811745 (M18X - But this example is solely for gpu clocks only)
8: Another list of people doing better than you 3DMark11 - Performance overclocking records @ HWBOT

So, I would choose my next set of words carefully If i were you, because I have been doing this far longer than you have and i never base answers on fiction.

But hey, You seem to be the go to person on over clocking so I'm going to let you continue to have you day in the sun. Don't mind me. I was just babbling fictitious no backed up answers.

 

Have you owned an M15x? Are any of those benchmarks from an M15x? Nope. Other than svl7 (who probably turns tess off), the benchmarks here are pretty darn good. Are you familiar with the throttling problems that plagued the M15x? King of Interns found the frickin' solution to that problem so I'd take his word over yours. When you have some info regarding the M15x I'll be happy to have a read...

 

Then you keep going with that and have a nice day.
And I'll leave it at that.

 

Sometimes people don't give credit where credit is due. Just because someone shares any idea does not mean it was their invention.

Interesting that an observer with great knowledge suggesting a few things to check and report in order to potentially help fix a problem receives a defensive response instead of appreciation.

If Futuremark did not flag the benchmark, there is no reason to assume the score is less than legit. The only flag is the driver version. Turning tess off flags it differently. Considering whose M15x score it is, there is even less reason to assume it is explainable as anything other than skill and knowledge.

 

Tbh I misinterpreted his response, due to past experiences. Before King discovered the throttle fix, we were all plagued by gpu throttling except apparently svl7 who was achieving these insane benchmarks.
I asked several times on T I how he was avoiding throttling and how I could too but he basically either didn't reply or said you need to balance CPU and GPU even though he was pushing both to the max.

I know svl7 has contributed hugely to us M15x owners, including the 7970M vBIOS patcher and the 680M custom BIOS. I've very grateful for that but I thought he was being similarly vague this time around as I have not seen anyone else push 7k on a 920/940xm on M15x other than him. Happy to be proved wrong.

 

It's easy to misinterpret written words. It happens to me often... either I misinterpret, or my message gets misinterpreted.

I cannot speak to your experience about questions not being answered, but I can speak from my own experience. There are some questions I avoid responding to. The reasons can vary... keeping a secret as a tactical advantage in benchmark scores is sometimes one good reason. In some cases the response needs to be detailed to the extent that I don't have enough time to spend providing the level of detail it deserves. In some cases I am doing things to my machine that involve risk or require some level of skill to avoid catastrophic results, and those are the the times where I go out of my way to skirt the question. I don't want to be legally or morally responsible, or feel guilty about, someone destroying their Alienware because of a mistake or not following directions that I provided. A good example of the latter is my dual AC adapter mod. I would love to see every M18xR1/R2 enjoying one. But, I decline requests from folks asking if I will build one for them.

 

Johnksss did you read my posts? This is only concerning the M15x. Hence why I made it bold and why this thread is in the M15x forum section.

I am well aware that the 7970M is capable of achieving far higher stable clockspeeds at stock 1.05V in OTHER machines.

This thread is for M15x laptop owners. The comments I make pertaining to stability only apply to this laptop!

Again please show me a screenshot showing A M15X running a 7970M at 1.1V at 1ghz+ on the core. I would love to see it in fact! I looked over your links and guess what SVL7 also hits a wall around 1.05V in the M15x most probably as his score is recorded at 975/1475. Clocks that I could run just about at 1.05V but I don't for full stability.

Also I never stated my own score in 3dmark11. I have also got a decent way above 7k but I really don't care about benchmarks. I only care about games. Benchmarks I use to simply test an OC prior to gaming to check for any issues particularly throttle issues. You are wrong about me bragging. I don't brag about my machine why would I it is obsolete!

Anyway I would love to know the reason for your rather personal attack which is I think completely unwarranted and unnecessary! I respect the fact that you have been around for some time (the same as me apparently) and have a ton of posts but it doesn't give you the right to attack people in the way you have done here! No I don't know who I am talking to. However you are talking to someone that knows more about this particular machine than you do. FACT. I have spent hours and hours over the last couple of years playing with this machine and I am quite meticulous, I have even tried my hand at overvolting the cpu with success although it is still arguably a project in progress yes I intend to push a stable OC (in prime 95) beyond what currently has been achieved (IN THIS LAPTOP i should add right) and I am hopeful to do so. When I am fully successful I will share my ideas and won't keep them secret that is assured. The reason I do continue to do that mod is for the community. The 920/40xm is old however if 3.6-3.7ghz can be safely achieved across all cores at 100% load and at 4ghz on single core (prime 95 and games not benchmarks) then we can continue to drop in powerful GPU's without bottlenecking them that is of course if MXM 3.0 is still supported in the future.

Please don't continue. I will.

 

Surely the person that shares an idea with others after working it out themselves after many hours of trial and error has as much claim to it as someone else that worked it out too but then decided not to share it and make it public knowledge. How can one share an idea if the inventor didn't share it in the first place? Credit is due to all parties involved I think.

Let's put it this way I agree with both you Mr Fox and fatboyslimerr. More directly I did not share someone else's idea if this is what you mean. I worked it out myself and later on with the help of certain members of this forum that tested with me.

 

Well, I'll tell you what. Let me see what you have on record. Anything ran at your "stable clocks" in a M15X.
Meaning gaming videos that show your game play at these clocks. Since you frown on benchmarks.

Your problem is, you do not want to broaden your scope.
Example:
You are having power/throttle issues with a 7970M. Yet we already ran a GTX 880M with no throttling in a M15X in under a day. Was this luck, skill or knowledge or a combination of sorts?
This card draws way more watts than a 7970M does right?

Sorry if you felt it was an attack, but it was not. It's meant to push you past your current limitations. Meaning I personally think you can try a few more things before throwing in the towel. If I thought that way, over clocking in laptops would not be where it is today.
Just to give you an idea of what your machine is still capable of...
http://hwbot.org/submission/2502691_svl7_3dmark11___performance_geforce_gtx_880m_10070_marks

 

I have no throttle issues at 1.05V in any game. I can take it to the limit at that voltage but the issues begin at 1.075V and above. I didn't throw in the towel I tested to the limit.

The 880M GTX I read is roughly as efficient as the 780M GTX at stock clocks due to better silicon (I think) The 7970M at 1.05V pushed to the limit at that voltage would be running very similar to what the Kepler based 880M in terms of power consumption. It is widely known that Kepler is quite a bit more efficient than GCN. So no it is completely possible to run that card in the M15x although (as like with the 780M GTX) you wouldn't be able to run much if any of an OC as it would be pushing the power limits at stock power draw.

Why do you think I have thrown in the towel. That is not what I do. Also I am not as ignorant as you seem to think I am. I also do endeavour to push the limits and take risks. Also you mention that you think I can try a few more things. You didn't before in a cordial manner. What are those things?

My current limitations in terms of hardware are 1.) I await the true next gen architecture release and 2.) A wife that wouldn't be happy if I spend $1000 on a new GPU.... no other limitations mate.

Nice bench result: I have seen it already but it really doesn't apply to us as SVL7 won't share how he achieves such high overclocks. Skill and knowledge no doubt. In the meantime we will soldier on and do our best to work things out for ourselves as we have done already.

 

Oh yay another svl7 benchmark! I'm only interested in benchmarks from us mortals... Did svl7 even try gaming with whatever OC he had with that 880M?
Ignoring svl7, who has the highest M15x 7970M 3Dmark 11 benchmark?

 

Yay another missing the point comment.

This is to show what the M15X is capable of doing.
Im pretty sure stock would be better than what you are getting now with being over clocked. The fact is. It can run. Is what you should be more interested in.
PS: I'll give you guys your first hint.
It's in your signature.

 

I would have to run it again but my max was around 7300 tess on. Can't be bothered to OC it back up again though as I currently play red alert 1 rofl. Hardly a reason at the moment.

 

You don't even need to over clock to play games...

 

This guy lol! Perhaps he think everyone can go out and buy a 880M GTX.

I am stuck with the 7970M perhaps for good Johnksss. Of course the 880M GTX smashes the 7970M OCed.

I know you are trying to be funny but it is pretty pathetic.

 

Guess what I am not. Stock clocks and undervolted... I haven't updated my sig in months...

what is your point here? We know that 1.05V is the key to success and the limit of the M15x. Of course that voltage can also apply to other cards. I doesn't mean however that we all can simply hop out and get the card. We do the best with what we have. Clearly the 880M has alot of OC headroom at 1.05V. Still doesn't make it any more attainable to me though so I can join your crowd.

 

The point is. You're doing more complaining about what you can't do instead of find the answer that is sitting right in front of you. And if you're to broke to buy a higher card that's fine. There are other things you can do. And i'll leave it at that. And that's my "pathetic" reply to a "pathetic" Situation.
You have a nice day now.
PS: A 680M is better than a 7970M and they are dirt cheap!

 

Nice condescending cryptic response. Thank you for that. I thought this forum was all about helping others. You are the first I have come across to behave this way to others. Your contribution here 0.

Happy?

Enjoy having a laugh with those that know all but care to share nothing.

edit: PS thanks for the insight. Glad you decided to contribute. Thank you!

 

Seriously. did you even read from here to the next post down?
http://forum.notebookreview.com/ali...5x-throttle-final-solution-5.html#post9606666

 

Now the 680M is pretty cheap if you can find one. Although I wouldn't say dirt cheap. I have had this 7970M for nearly 2 years now the 680M wasn't dirt cheap then (in fact it wasn't available). Sure it is the better card and far better in the M15x but I would still ending up losing money purchasing it now. Your point being?

You hint to us changing our GPU for better performance and we will do so when we can afford to so. Again why do you bother with the sarcasm? At the moment without losing money there is no better card for those of us already stuck with the 7970M. The 780M and 880M are better indeed but very costly. The 680M marginally better and noticeably so with an OC however still not enough to warrant selling the 7970M at a loss for it so close to a next gen release.

I certainly considered selling the 7970M and getting the 680M but decided to wait out for the maxwell and whatever AMD launch. Is that your hint?

As for complaining I am quite happy with what the 7970M is capable of thus far. I don't see what you mean by accusing us of complaining what we can't do.

 

Ok, maybe not dirt cheap!

Actually there is a point to this, but depends on how you look at it.
Example:
This M18XR2 started out as a M18XR1. With crossfire 6990M. It's now what it is today.
In your case. You don't have to go buy a new laptop you can still upgrade yours. What was thought to be impossible is perfectly possible. So say you find the right person selling a 780M for example. You spend say 675.00 or so at RJTech. You have virtually a new laptop to you for under 700.00 Dollars. Far better than spending 2k for a new one if you are on a budget or really looking to get the most out of what you spent two years ago. These things are investments and if that investment can get a major uplift without spending thousands of dollars, then why not do it?

Only a suggestion of sorts....
I know it's not the same thing, but check the links to gaming play at the bottom of the post.
This is perfectly transferable to the M15X. And should be consider pretty darn close. So that would go for the 680M/780M/880M since throttling has been fixed in the card itself.
http://forum.notebookreview.com/alienware/746259-my-nvidia-gtx-880m-test-run-review.html#post9577480

 

Again I would like to but want to wait for maxwell. Kepler and GCN will soon be last generation. I totally agree with you is keeping the old beast going as long as possible in fact I have always followed that philosophy. My last laptop was a Asus C90P that I upgraded to X3380 3.16ghz quad core and the GPU I even suceeded installing a 4850M in the MXM 2.1 II slot by shaving away some of the metal housing on the cpu socket. It worked (albeit to vga only lol) This in a old P35 15 inch laptop. I even managed a respectable stable 3.5ghz on the quad. I am very much an advocate in keeping hardware going as long as possible but I try to do so in the most economic way (not because I can't afford it I can afford the 880M GTX) but because I like the idea of longevity and have to balance the investment in my hobbies with the investment in my family (clearly one takes higher priority lol). The same applies to my other hobby in hifi.

You preach to the choir!

 

Ok, i'm retracting the "broke" comment.

Nothing wrong with waiting on Maxwell. and the first samples have hit the market. Just waiting on some "real" data on what they can do in the "real" world. Sort of why I haven't jumped on a 880M just yet, but i fear it may be 6 months before we see the high performance chips hit the market. (Hoping I am wrong)
You are absolutely right! Family takes president over all! LOL. But um, my kids are trying to follow suit with gaming desktops. And they have 480's and 5870's.
Yep, completely understand about hobbies! Benching for kingpin cooling is a pretty darn expensive hobby and so are fixed up cars. To much money spent on those two items.

Now. This is what you need to find out just what you can do.You will need to mod a 330W psu to work with your M15X. Not sure if that is an easy thing for you to do, but it's very do able. Or a dual psu. There is no other way you will be able to get past certain power issues if you do not have enough power to begin with.

 

I have a 330w psu but already removed the data pin on it. Are you referring to solid states mod? It is beyond me the 210 w does however provide more power than the 150w .


Sent from my iPod touch using Tapatalk

 

I have all 3


Sent from my iPod touch using Tapatalk

 

Are you sure it is power related as whether I have my xm oced or not 1.05v is the limit. Surely this shouldn't be the case if it is simply a question of lack or power


Sent from my iPod touch using Tapatalk

 

Ok, let me ask you this.
Do you have a kilowatt meter? Or some way to see what you are really drawing from the outlet.
The reason is to see if you are in fact pulling more than 210W to 240W.

PS: Your system is capable of drawing far more than 150W
I know it's not a good example but the 780M and 880M can draw about 200 to 250 watts each

 

Yeah I know the 150W cuts out at around 200W.

However the 210W with no data pin doesn`t have this limitation it can pull whatever I need. I can run 25x across all cores in 3dmark and 950/1450 all day and have no throttling not even in the combined test. However if I go up to 1.075V and above the card is not stable.

Also I have also tried +200mv on the cpu like SVL7 did and the cpu did not throttle in Prime 95 although the heat was excessive which was not sustainable.

Basically I realise that the 210W can be pushed fairly hard at my maximum stable clockspeeds but the reason for my doubt is that the xm at 2ghz consumes less power than at 3.2ghz across the 4 cores. This however has no bearing on whether I can push past 1.05V or not though. It is a wall.

Shouldn:t that make a difference? I haven:t got a kilowatt meter only a voltmeter on the cpu lol. I would have to purchases one to find out for sure.

You could be right so I will purchase one and find out soon. Both 210W and 330W have no data pin however. I removed them as this to removed throttling issues completely. This will matter? I know the machine can:t charge with it removed but even modded it will still work? They work now but with no charging of the battery.

I did this after realising the complexity of the PSU mods. I am not competent in electronics and playing with a PSU`s internals and making a mistake scares me! Fire!!

 

Hummmm, without a killawatt meter to tell us for sure....I can't really say. I say it's power. And with you not pulling anything above 1.05V because this is putting you at just under the psu's alleged max. And when you go up, it crashes due to not enough power.When over clocked these cards draw far more than 100W and for you to push the card you will need far more watts than 1.05V is capable of giving....

Guess we wait and see. Once you get your meter, then you should have a better idea of where your system stands.

 

Yeah I will try to get one very soon. There is a hardware store around the corner however I certainly won't announce the purchase to my better half before doing so! Let's hope it is cheap enough to stay "under" the radar!

 

I have a watt meter and I found I never got close to 200W when running pretty high overclocks on CPU and GPU. Max I recorded was 185-190W while playing Far Cry 3.
This was at 1v though. Haven't recorded power draw recently since I gave meter to my dad....

King I can't remember what you said about 330w PSU, do you have it working completely throttle-free with centre pin removed?