M15x 7970M throttle FIX !

Yeah I noticed from a cold boot it works better as well.

I can't get anything stable around 50-60fps if I go higher than Low settings.

Running 850/1200 and 0.975v.
I have a feeling it's undervolted too much as well, but it's been days now I've been trying to change it. No matter what vbios I flash the card is stuck at 0.975v. I want to get it back to stock.

 

Sounds like lambda and I are having the same kind of issues. We will work this out. BTW I have been talking about crysis 3 multiplayer, and with the gpuz render I just alt tab out and close it when the match is loaded to about 99%, but it is not as spot on as in single player where you press enter. I will try with single player too, and I'll try from cold reboot.

 

I have updated the first post to reflect our findings. Happy OCing people! I am settling for 950/1450 at 23/23/25/27 from here on out until I upgrade as I believe it is close to the max stable clocks for me at 1.05V.

I see what you mean with multiplayer. Unfortunately timing is VERY important. If the gpu loads to 100% before you can initiate the gpuz render to ramp load slowly the game will already have throttling issues. THat is a difficult one!

 

Oh so your saying that the throttling is actually Directly initiated by 100% load... Ok maybe that will help me try to fix the issue. Because I think sometimes I was letting the gpuz render run like .5 seconds after I initiate the cryengine... IDK this is still a great thread +1 I forgot if i already gave u one but ill give u another one.

 

Thanks. Follow my instructions to run Crysis and you will be throttle free. GPUz render still puts to GPU to 100% load but the GPU is actually stressed much less than in game.

 

I am a bit freaked out this weird thing is occuring. In gpu-z it says my clocks are at 0 mhz. At least hwinfo says they are at 2d clocks, 300/150, but when I play bf3 (idk im playing bf3 for now not c3) it remains in 2d clocks doesnt go up to 3d! This all seems to have started after I downloaded asus gpu tweak tool! I restarted and they still show 0. I'm reinstalling graphics driver then ill update. I hope my card didnt get damaged from asus gpu tweak.

IDK now the issue is resolved at least i learned to never use asus gpu tweak again!

 

AMD cards. They are fun for all the family

 

Ahhhh finally i found you here! Thanks for the hint on my video! I'm still beginner though haha! Do you think 900/1400 will work well on 720QM?

Thanks bud!

 

950/1450 will work fine You need to do you own testing but I reckon you could jump to 950/1450 at 1.05V and then take it up in increments.

My card doesn't like much more at stock voltage so I stick with 950/1450. Overvolting in the M15x does not work. If you try to game at an overvolt the system shuts off.

However every card is different. Maybe your card can hit the wonderful 1000mhz barrier at stock voltage! Only your testing can find out.

Oh and welcome to NBR mate!

 

Just uploaded new video about Crysis 3 by using VERY HIGH settings on all options. So far it is running good while the temp only hit 77c.

Thank you for the warm welcome!

 

Looks good!

 

I can see how the rending tool, game minimised at "press enter" screen, gpu ramp up could work but what about games like tomb raider where you load straight into the game?
Could you minimise as game is loading, use render tool till 100% then drop back in and see what happens?

Just doesn't seem like it would work for all games? Or is it just Crysis 3 that drops you in at 100% gpu usage?

 

I was wondering exactly that too. I've been trying out Tomb Raider and since it loads straight into game it's hard to figure out when to use the render tool.
I'm getting crazy throttling everytime cutscenes start.

 

Nooooo! What if you load into the game then go to ESC to game menu and minimise to windows then use render tool then maximise back to game?

 

At the in game menu. Before you load into the game. Does the card run at full 3dclocks or 2D clocks. If it is like Far cry 3 which runs the card at 3d clocks even at the startup menu then you can alt tab out before you load into your game and run the render. Play about and see what happens.

 

Has anyone been able to make Far Cry 3 or Tomb Raider work at high OC without throttling?
My 7970m arrives in a week and I can't wait to play these beauties at max settings

I know 240w PSU helps but I should be able to run 150w PSU with OC on gfx and cpu?
Or if tomb raider is more gpu intensive, don't OC the cpu?

 

Me. Read the first post to learn how to kill throttle in FC3. I did all my testing with that game

 

Bah! Sorry I must have missed the really obvious FC3 mention. I think Tomb Raider uses 3d clocks at the menu too as it renders a nice stormy ocean so should be the same process as FC3.
Will report back when my card arrives.

 

Yeah please let me know the process so I can update the first post

 

Anyone tried out Bioshock 3 with 7970m?
I'm getting smooth fps but the game will throttle when going into random rooms or turning certain corners.
Tried render tool at menu but it doesn't seem to help at all.

Also I'm still stuck at 0.975v. My card just won't go back to stock 1.050v no matter how many times I flash.
I've unlocked my msi afterburner, so now my clocks are adjustable but they're maxing out at 850/1200. And I can't adjust the core voltage slider, it's greyed out.
I've read some people were able to adjust the voltage through afterburner. This may be my only option to run stock voltage right now.
Anyone have any suggestions or help ?

 

I finished up Bioshock 3 this weekend but I did play most of my hours on my M15x with 7970m. Unforunately the render tool didn't appear to help me with Bioshock, what I did do was powerdrain my machine and NOT run ThrottleStop and gamed at only stock 920xm settings. My 7970 still throttled a little but overall the game was entirely playable which I had set as Ultra.

Unfortunately due to these throttling issues and while I could sell and buy a new GPU. I am selling my awesome notebook and going desktops. Although I definitely plan to get another notebook (just not a gaming one for now) Gotta say, its the only game I've played where the render didn't help. Rendering before play seemed to help GW2 quite a bit when I jumped on that again for a week.

 

When did they game tend to throttle for you? Like what parts in the game?

 

That's a shame daverish. Did you try different methods? Each game seems to treat the GPU differently. Sometimes it takes a few tries to work out how best to use the render to stop throttle.

 

Is there any way to monitor how any particular game is using the GPU? For example in a menu screen or load screen etc. To determine how much of the GPU's resources are being used?

 

Gpuz. It continually logs data. Also if you are quick and alt tab out you can observe about 1 minute of prior activity both load and temperature on the graph. The other option is to consult the log file if you set it to record.

 

Yeah but if a game is running full 3d clocks in a menu it will be running like 850/1200? I can't imagine any gaming running like 150/350 even in a menu?

 

Crysis 3 in one. Doesn:t run　3D clocks until you start gameplay. All games are different.

 

So I managed to put my voltage back to stock 1050 through AMD Overdrive, and my performance has GREATLY improved. Bioshock 3 now runs real smooth. However my temps have gone way up. And I got shutdowns a couple times, hitting over 80 degrees.
Even CS:GO is hitting 75-78 degrees.
Is this normal?

 

Great to hear that you got your stock voltage back. And yes, it's absolutely normal. The increase in temperature will also depend on your ambient temp. Try playing in an air-cond room?

 

You shouldn't have any shutdowns at 80C!

You should be able to run an OC at 1.05V with no shutdowns occuring even with a heavy CPU OC engaged.

 

I think it shuts down around 83-84c.
I thought I remember many people having this problem with their m15x's when the 7970m first came out? Although I might be mistaken. But you think this is definitely not normal?

 

So I just played Bioshock3 for 20mins and got another shutdown. It just gets progressively hotter the longer I play. Stays in the high 70's for a while then get's into the 80's and once it stayed at 85c for 5 secs the laptop shutdown.
Are these temps normal? It seems like such a drastic difference from gaming at 975v to 1050v. My performance is so much better but these temps are insane. I'm not even playing on ultra or very high settings.
I just cleaned out my fan and vent to make sure it wasn't that. There wasn't much dust in there though.

 

What about the contact between heatsink and die on the card? When I first installed my clevo 6990m with clevo backplate the heatsink wasn't making proper contact with the die and I was getting shutdowns due to crazy temperatures even at idle.
Are you using the correct backplate/x-bracket?

Do you have sufficient thermal pads and good thermal paste? These make a big difference.

Also do you have the A09 BIOS for your m/b?

 

Yep running A09.

I'm idling around 40-45 degrees. CPU a little higher 55-60 degrees.
This is my second 7970m so I think I know how to install them pretty well. I'm pretty sure I got enough pads placed. I bought some thermal paste from best buy when I got the card cause I couldn't be bothered to wait for some good stuff in the mail. You think it would make that much of a difference? I figured my idle temps seem pretty normal?

With my first 7970m I used a copper shim. But with this one I grinded down the x-bracket to fit.

 

85c is the thermal shutdown limit on the 7970m. Those temps appear pretty on for going to 1050v on the M15x's cooling. I found at 0975v I'd idle around 34-35c and load max upto 72c on stress tests. The little I tried 1.05v though, those temps take a good jump close to 85c!

The slowdowns I had were

Spoiler

Later in the game, once you go to enter Comstocks Manor and you're free roaming around parts of the city. That exploration is where I had some decent FPS and performance drops

. Now it may have been better if I upped my GPU voltage while keeping my XM as stock.

 

Has anyone managed to get Bioshock running smoothly with overclocked cpu and gpu?
The menu definitely uses 3d clocks so running the render tool then immediately starting the game should stop throttling?

 

Fair enough daverish. I only had a shutdown when I went for 1.1V at 1000/1500 MHz lol. Now I know why that happened! We need to improve our GPU cooling. 85C isn't very high!!

 

Hmm just had a shutdown whilst playing Bioshock Infinite but not sure why. OCed 7970m to 950/1400 and cpu to 3.3Ghz 86/65.
gpuz log shows gpu was only at 71 degrees before shutdown so that only leaves a cpu thermal shutdown (if that is possible) or some peak power usage thing?
Possibly my 150W psu tripped?

 

It's not your 150w power supply. I had something similar happen to me with my 240w power supply. Mine seems to shut off at around 80-81C. It doesn't get hotter than that and my cpu is fine because max is around 70C nothing more at all, normally around 68-69C is what I get max. I'm not sure why our systems shut off at 85c but it happens on the M15X and M17X-R2.

Does anyone know if it also happens on an M17X-R3, M18X-R1 or M17X-R4/M18X-R2?

Would our systems shut off with a stock 260M or 5850M if they reached 80-85C? I've never really gamed with my 260M just swapped it with a 7970M as soon as I could.

 

I have never had m15x shut down at 85c, I have had 5870 and 260m and used the laptops in Rajasthan (45c ambient in desert).

Now have a m17xr4 with 7970m and will test in few hrs by blocking intake to heat up things, lol

 

Thanks, that will definitely help us find the problem. If the oem cards didn't shut the computer down then it's either 7970m specific or it could also happened on upgrade cards ex: 680m, 670mx, etc.

If it doesn't shut down the R4 @ 85-86C then it has to be very specific to the m15x/m17x-r2

 

Arrghhh another shutdown on Bioshock but this time without any cpu OC. Gpu was at 77 degrees when it shutdown according to gpuz log.
That isn't enough to make 7970m shutdown is it? Not sure what is causing this but its quite annoying. I suppose I should lower my overclock a bit since it was stable at stock.

EDIT: After some reading I think my shutdowns are caused by cpu overheating, not psu cutting out or gpu thermal shutdown. Going to invest in some ICD 7 (can't find anywhere in UK selling ICD 24) and repaste cpu.
I notice moderate cpu usage whilst playing bioshock infinite but some cores are maxed out with high temperatures. Hopefully this is the solution.

 

How high were your cpu temperatures? CPU shouldn't shut down until 100C. All I can think of is the vram or some other component on our 7970m is overheating before the core does.

 

Interesting! I'm going to repaste cpu and gpu so will see how well thermal pads are applying pressure to VRAM.
I'll get cpuz and log to file so I can look at cpu temps if I get another shutdown. Interestingly 7970m ran at same temperatures at stock and at 950/1400 OC lol!

 

That's not bad, I wish mine ran at the same temperatures when overclocked.

What voltage are you running your GPU? Is it the stock 1.050v? Also I may have missed it, but which method did you use for the x-bracket? (Copper shim, grind x-bracket, or buy the clevo x-bracket with proper screws?)

 

I used standard amd x-bracket and removed the original C-screws from heatsink and used screws I bought from a hardware store.
Made a post about it here
Running stock voltage.

 

Well I took a knife to my wallet and bought a M18x. I'll probably put my 7970m in there and test it out or keep it in there.

I definitely want to try the 660M in my M15x and overclock my CPU and see if it throttles or shuts down at 80-85C.

It can't be a power issue because it happens at a higher temp. It can't be the card because supposedly they don't shut down in the M17xR3/R4 and M18xR1/R2 [I'll test this out later when I get my M18x]

So something to do with temps and the M15x/M17x-R2 computers shutting down when the GPu reaches this temp.

There is a possibility that the bios shuts off the computer if the GPU reaches 85C or more, but this would only be a valid assumption if the 260M and 5850M also shut off at that temperature (I've never stressed my 260M that far).

The last possibility is that our vram or vreg or some other component on the GPU itself is overheating causing the computer to shut down. Since we are playing more demanding games [Crysis 3, bioshock infinite, etc.] it may be possible, but finding out what is overheating might be a bit hard without hardware monitors.

 

I got throttling when playing battlefield 3 with a GTX 460m and a big OC 940xm.

Does VRAM get really hot? I read on another thread that svl7 posted VRAM doesn't get that hot. Elsewhere I read that VRAM gets hotter than the chip so if chip reads 80 degrees, VRAM could be 90!
That could definitely cause thermal shutdown!

In the end running an OC on the CPU made no difference to my shutdown issue. Its purely gpu-based and I think its to do with the VRAM.....

 

Yeah even a 260M can cause throttling with an overclocked cpu.

VRAM shouldn't get too hot, but it can always be a possibility. If I remember correctly, wasn't there an unlocked bios for the M15x? We could try that and try increasing the temperature values a bit and see if that helps.

 

When I installed 7970, I noticed with stock heat sink, VRAM doesn't make full contact so I doubled the thermal paste and it seemed to make proper contact. It's only the far side that doesn't make proper contact, so maybe try that and see if it solves the issue.