Khenglish

TS will only update the TDP values if you delete the config.ini file, otherwise TS will enforce whatever TDP value you wrote in it previously. I found XTU TDP changes to only work after reboot.

Yes clevos fan profiles are poor. The GPU fan does not run high speed until 90C, which is the same as the throttle temp. The CPU does not run full speed until 90C as well, but at least it has 10C left of throttle room. The main problem I have with the fan profiles is that the fan speed lags, and it's possible to skip over trip points if the CPU heats up fast enough. There's no reason why clevo can't make the fan speeds ramp up smoothly so that there are no audible fan speed changes like some laptops do. I suspect Prema can get the Fn-9 shortcut for full speed to work for your system though. I just looked up some stuff on that laptop and apparently the GPU cooling is very good, so you should be able to run high voltage BIOS's. In regards to CPU, a 3970x was overheating a little, but I suspect a 3930k will be fine. A problem with overclocking clevos is that since there are no BIOS overclocking options (same with the P570 for some heinous reason), there is no proper CPU voltage control, so the VID will ramp up stupidly high. Ex for me my CPU will run 1.351V at only 3.4GHz, which is fine for 4.35GHz. I'd love to knock .05V off my CPU voltage and run say 4.1GHz to keep power draw and heat in check, but clevo doesn't let me do that because their BIOS sucks. For your CPU you would likely want to drop the voltage slightly then overclock a little from there or else you'll be overheating, but the auto voltage control will ramp up the voltage if you want to or not. There is no excuse for not being able to set fixed voltages on such a high end system. I can't really see how you can improve the CPU cooling much. IHS removal won't be very helpful since the IHS is needed to spread heat to all 5+ heatpipes. Perhaps clevo bent up the heatsink plate where it touches the CPU like they do on all P170/P150s and you can get 10-20C by lapping it, but I doubt clevo messed that up on the P570 since I feel that mistake would make a 130W CPU impossible to cool. Even if they didn't though lapping the IHS and heatsink should get you a few degrees improvement. I saw the CPU MOSFETs on a picture with the back open and I counted 24, with possibly more on the back. My P150EM has 9 total and they're good for ~80W sustained, so I think your only concern is cooling. Looks like a fun system to play around with. Hopefully clevo's idiotic BIOS limitations will not be an issue.

It only has a 300W PSU, which is severely underpowered for that system. Even a 330W M18x PSU (the strongest I am aware of) will also likely be too little so you likely should do a dual PSU mod to overclock that thing significantly. Getting a system like that and NOT overclocking it a ton to me is just insane. Lol that laptop is like 3 inches thick BTW. It better have good cooling. Update: It looks like there is actually an adapter made for this system to use 2 PSUs, and the PSUs are now 330W.

It should be fine. Fans are controlled by a standardized 25KHz PWM signal, so if one fan works then any should work. The only problem could be if you're putting in a 3.3V fan and the motherboard is supplying 5V, but this mismatch is unlikely. Usually fans are rated for 5V.

Your core looks average and your memory looks slightly below average compared to a clevo card. I think Dell and clevo cards clock the same on the core, and maybe Dells are usually slightly higher on memory.

Does the fact that my CPU's name string is null make it so I can't get any points?

Latitudes will dynamically adjust the TOLUD if they see an eGPU on startup. The problem is that they also disable the iGPU. Because of this it is necessary to hotplug the eGPU after BIOS POST so that the BIOS does not disable the iGPU, but this means that the TOLUD will not be automatically lowered. When booting with the eGPU on and connected with a E6520, I found that one of the eGPU DVI ports would output a video signal with a monitor connected or not. If you tried one port and it doesn't work, then try the other. Unfortunately you cannot stop the system from disabling the iGPU so you will not get optimus with this method, but your eGPU should be capable of running this way until you get the DSDT override working.

I agree. The performance gain is not worth the $600+ cost of the 680m. I suggest waiting for the 780m and either getting that or getting the 680m then since it will be far cheaper.

You can give this a shot. ASUS does not seem to lock their firmware ROMs, but their downloadable BIOS images are incomplete, so I don't know if the firmware's format on the flash ROM will match what Intel's editing tool expects. Remember 1 small mistake can brick your laptop so be very careful. If you'd like I can check over an image for you to see if it is OK prior to flashing. Remember that at best you'll get a 10% overclock, so keep that in mind when flashing your system and putting it at risk of a brick.

That's nuts. I was hoping that since haswell will also use DDR3 that laptop makers will put out better memory. No better memory came out with IVB's release though so I'm concerned. Looks like that memory is hard to find too. Amazon wants $30 each used.

Wow your memory can handle CAS9 2133MHz. Mine can't even handle CAS9 1866. And is that 4590MHz on the CPU stable? Mine can only do around 4.35, with 4.4 OK if under 80C, and 4.5 OK if under 60C.

I'm not really sure what you're concern is. All a BIOS update might do is offer a new microcode for the retail stepping if one is ever released, which is unlikely this far past launch. There will certainly not be any more microcode updates for your stepping 8. The final stepping 8 microcode was probably out even when your laptop was launched, so I highly doubt svl7 needed to update the BIOS's microcode database, so an unmodded BIOS should load a microcode update perfectly fine.

XM and non-xm, as well as dual core use the same microcode. Different CPU revisions do however use different microcodes. The retail ivy bridge revision is E1 (stepping 9). I suspect that the 3920xm you received is an ES, likely revision E0 (stepping 8), while your previous CPU was a E1. In this case your system is running the most updated microcode. You can check your CPU revision/stepping in cpu-z.

How big of a performance hit are you getting for running T2? I've been trying to run minimum latency so I run 1866 10-11-9 T1. Also noticed that my memory is very temperature sensitive. Going to attempt putting copper flashing on it and see if it helps.

Is MSI afterburner not working? It's the number 1 overclocking tool by far. And please tell me you don't use the M16 like 900% of the players in that game do. G3A3 FTW

Those GF110s still have all their resistors! Nice job! Now the next step is to find some titans off ebay and solder them in so you have the world's first dual titan card. Both have 384bit memory so it HAS to work!

I find laptops to be hit or miss if the applying a small dot in the middle and letting it spread itself method works. You need to have high even pressure on the die for it to work, or else it will produce inferior results to the manually spreading thinly method. For example on my last laptop the dot method did work well, but on this clevo tightening the cooler down puts relatively low pressure on the middle of the die compared to the edges, so I found that the paste would be staying thick in the middle of the die. By spreading the paste myself I get around 7C lower CPU temps. For spreading the paste myself, I never got using a razor blade to work. There's always some bump on the blade that ruins the spread, so I just use plastic pulled tight around my finger. What I found really make a difference is that when spreading the paste, if looking straight on at the die it will look like you completely covered the die, but if you look at an angle you'll see the die shining through in several spots. Processor die = shiny while paste is not, so you need to make sure you don't see any shiny spots when spreading the paste yourself. I have definitely put scratches in the die with PK-1 and Shin-itsu, so I think IC diamond is not special in this regard. I don't think the tiny scratches have a significant impact on cooling and they are too minor to damage the CPU. If not then people would be destroying processors all the time since I feel like most pastes scratch the die some.

You want the 4MB chip. 4MB chip = BIOS, 2MB chip = ME FW

Even though the size is the same, your BIOS really is formatted much differently from standard. I can't make out where the ME FW is in a v223 BIOS image with a hex editor. A ME FW image has a characteristic header that I simply cannot find.

Unfortunately I cannot extract out the 4MB image. FITC is not recognizing the format, meaning that ASUS is not following intel standards with the BIOS image. It might be possible with a hex editor, but I think that method would have high odds of resulting in a corrupt image.

FITC is saying it can't recognize the image format, so unfortunately I cannot split the flash images. I even tried old versions from before my FITC was released and still no go. I'm not sure why Asus deviated from the Intel standard. Sorry.

ME FW = Intel Management Engine FirmWare. It's an extension of the core BIOS for more advanced BIOS features and hardware setup. The ME FW is what Intel XTU interacts with to change settings. It can cause your system not to boot, but I think the modded BIOS you flashed did not touch the ME FW, so you only need to worry about the core BIOS chip. Interesting that the BIOS memory chip is 8MB. Flashing on a 4MB image only should not matter. Manufacturers often put on larger flash chips than they need. I think HP even uses 16MB chips. On this laptop after a BIOS flash and resetting BIOS settings, it actually won't boot (freezes on POST screen) unless I start it off battery once. This laptop will also fail to boot on first startup after a BIOS flash. I'm not aware of any BIOS writing company (Insyde, AMI, etc) that doesn't write terrible BIOS code, and pretty much anything can go wrong for no good reason. I would not be surprised at all if there is really nothing wrong with your BIOS image, but things just break down after a reset if things were set a certain way. I would have the company reflash a stock BIOS image despite an 8MB image not being available. Make sure they only flash on the 4MB BIOS image and not the 6MB combined BIOS and ME FW. If that does not fix it then reflash the ME FW chip too.

6MB sounds correct if you have a 1.5MB ME FW. The BIOS will be on a 4MB chip, and the ME FW will be on a 2MB chip. There is a larger 5MB ME FW version that will result in around 12MB total image size (4MB + 8MB), but I think ASUS only uses the 1.5MB version.

I love what the term "bench" means here, and what it means instead in other places (benchpress). 99C on CPU? Does catzilla run particularly hot or is 1.396V at 4.5GHz just like that? Prime95 hits 90C after a minute for me at 4.3GHz 1.351V. I can't get it any hotter since laptop throttles if over 80C for too long. Already modded cooling some and about to do a whole lot more (add 8mm heatpipe, make plate more rigid so die pressure is more even instead of pressing much harder on edges of cpu die). Waiting for solder paste to arrive. Would be interesting if I could make a 15" laptop's CPU cooling outperform an m18x 3 pipe cooler. It'd be amazing to get a giant 12mm heatpipe so that the width covers the whole CPU die so that a copper plate + solder is not needed, but enertron will only make one as part of a $300+ order :( As the cooling is now I get an immediate 30C jump when going from idle to ~100W. I'm hoping to reduce that by around 15C. Update: Ran the kitty test and max CPU temp was 77C with CPU at 4361MHz ([url]http://www.allbenchmark.com/showresult?lp=26540[/url]). Backplate was on. No laptop cooler, just fans forced high and back slightly elevated. Not really a fair comparison though since your CPU was working much harder during the graphics tests to feed the 2nd 680m. I know I don't have an m18x, but I feel like if other laptop users post here we can get nice comparisons between systems with real data. Update 2: Mr. Fox you were throttling quite a bit. I beat you in the cat physics test despite using an SB 40MHz slower. Again I only hit 77C on CPU: [url]http://www.allbenchmark.com/showresult?lp=26552[/url] 1172 vs 1166

The tape will dry out and stop taping after a while due to the heat. Most people do what you did with aluminum tape.