Dell Precision M6700 Owners Thread

Something like that?
http://forum.notebookreview.com/del...ision-m6700-owners-thread-70.html#post8957278


Best Regards,
Alex.

 

No, thank you for the tip, i'll do that and report back with the numbers

# Alls
Honestly for such an expensive machine I thought that the match boxes are built in
Thanks for the tip

# Krane

Thank you for the help man, appreciate it.

Looks like we are very close on temps meaning that we have the same crappy heatsink. I am saying that it is crappy because from the looks of mine the GPU does not make a good contact with the heatsink and it goes nowhere near those 70 degrees Celsius in Furmark as some people have written around here... more like 80-83 with a 25 Celsius ambient temp.

Quite dissapointed of the Dell quality of the hearsink AND the Dell heatsink mounting system.... and this is my SECOND heatsink... the first was even worse meaning the GPU had worse contact, the some VRAM modules had no contact at all with the pads...


I have to ask though if the Coolmaster was on or off during testing?

Thank you!

 

Those high (but still safe) temperatures would appear admirable for an extreme test, don't you think? Also, what are you comparing it to; and what were you expecting?

Its a dual heatsink isn't it? But the quality control issue does concern me.

The Coolmater was on low. But in my experience the fans make no difference but raising it off the surface does.

First, , the audacity.



And second, the cooling results can be had with a cooler if you prefer a less homemade look.

 

Ok guys,

As promised I returned with the numbers after i raised the laptop's rear end with 2 cm.

I've used the same Heaven benchmark settings as I posted at #2378

The ambient temperature was 26 Celsius
I've let the benchmark loop for 15 minutes and the GPU reached 69 and remained there... then while it was still running I raised slowly the laptop's rear end and put under it 2 small boxes, each 2cm in height (under the laptop's feet) and let the benchmark loop for another 10 minutes. The temperature dropped to 65 Celsius pretty quick, in a couple of minutes (at some scenes it dropped to 64 Celsius and even 63 for a few seconds).

# Krane
"Also, what are you comparing it to; and what were you expecting?"

I was expecting to see a K5000m max temperature of 70-75 degrees Celsius (given the low clocks and pretty good GPU cooling, at least on paper so to speak) but the Furmark takes me to 83 (in a short 5 min run).... and even if I abandoned testing with Furmark playing Sniper Ghost Warrior 2 from Steam constantly takes my GPU to 76-78 Celsius.

One side effect is that the laptop become quite noisy (compared to my previous Ati M6000) and that bothers me more then the GPU temp.

Bottom line... the heatsink is to blame for everything.

Later edit:

Strange thing, with the M6000 I was able to play games even if the laptop was on Quiet or Cool power profile, with the K5000m I am forced to used the Balanced or Ultra Performace profile because with the other profiles the GPU clock is stuck at 135Mhz.

Any workaround available to make the GPU work with proper clocks (601Mhz), when needed with the Cool/Quiet profile?

Thank you guys for all the help!

 

I've posted a side by side picture of the GPU print from the Ati M6000 and the nVidia K5000m.

ATI has better contact with the heatsink. With ATI I had great temps, In Furmark it settled at 65, tested for 10 minutes. In games it was around 60-ish... cool and quiet.

The K5000m does not make a proper contact with the heatsink as you can see.... the die makes a very good contact with the heatsink only on 40-45% of the total surface, hence the higher temps, higher noise of the laptop.

The first nVidia heatsink that i had was worse, it had a very good contact with the hearsink only on 12-18% of the die surface

The problem is that you CAN'T fine tune it and believe me that I've tried everything.... I could tell you the whole story but now it's not the time.

 

Well there's no excuse for defects is materials and/or workmanship if that is the case.

But remember, always compare apples to apples. The k5000m IS after all a Quardro, and therefore tuned specifically for professionals in the creative and data crunching fields. IT MUST BE STABLE above all else.

So it doesn't run circles around other gaming graphics card. That not what its designed to do.

Anyway, I didn't get any noise during my tests, but perhaps that was due to the shorter duration or intensity of the test (I set it for extreme). Perhaps I'll run another test with Furmark later, just to put your mind at ease and confirm or refute your suspicions. Stand by.

One more thing: I forgot to turn off my screen save during the test.

 

Here's something simple I have working. Seems the fans come on less and actually takes in less dust over time - probably because they are not running as much and dust is further away (less vacuum).
It is the corner packing support for an LCD TV (or can get molding crown from hardware store for a buck or two). It can be a little slippery so I put a piece of cabinet liner underneath (the semi sticky perforated kind), which I don't care for in cabinets btw

Edit: The peak height is about 1 5/8"
Edit 2: Oh, and find something to support the power cord so it doesn't dangle and ruin the socket.
Edit 3: If there is a HDD in the system you'll hear it a little more after elevation. I think I'm done editing, geez, sorry!

 

Can't comment on specific GPU heatsink contact issues, but how can you really expect the same thermal performance between M6000 (somewhat downclocked 7870M rumored to have the TDP of around 45W) and K5000M (somewhat downclocked 680M whose TDP is binned at 100W which means it's definitely above 75W)?

On the other hand, if you were more satisfied with the AMD card, you could check the experience of owners of the new M6100 card (used on M6800 line) - it seems to offer the same base level of performance as K5000M 3dMark11-wise, though since it's the similar core as M6000 with somewhat more processing units and a 40% overclock, the TDP seems to be increased to about 75W. Plus, the memory bus is only 128-bit so that might also be significant based on the usecase.

 

EDIT: double post

 

If you are using the PremierColor software, you can try disabling the split screens functionality (which seem to be thoroughly broken), by exiting the program and renaming the 'C:\Program Files (x86)\Common Files\Portrait Displays\Plugins\DP' folder to, for example, DP.bak.

Hope that helps!