Khenglish

I don't think anyone besides dell calls that a laptop.

That's only a bit over a 4% gain in fire strike for a 20% clock increase. What's funny is I see bigger gains with memory clocks in the combined test compared to the graphics test (680m = 670). Other tests/games really like more memory bandwidth too. Only way to find out is to run a bunch of stuff and see. I definitely do expect to see titan to be more memory clock dependent since the core is exactly 2x as strong as the 670 core, with only exactly 50% wider memory connection.

Gonna need more memory bandwidth than that to carry your team

I'd stick with this laptop. Clevos most certainly have their issues, which are annoying, but I am a nerd and I waste my time on this stuff as a hobby anyway. Unfortunately significant modding is required to make cooling competitive with top AW laptops, but it's doable. Only MSI makes a 15inch laptop that can also use the top end cards, but while it can be made acceptable, MSI laptops just seem to have a bit underpowered cooling. I need to take a laptop into school almost every day, and I am not aware of any 15 inch that has the power and modability of this laptop. 17inch just seems a bit too big to me, and I see no reason for the jump when the screen resolutions are the same anyway. Nvidia doesn't allow optimus on SLI laptops for some reason, so the battery life is horrid. My favorite is that clevo provides complete voltage regulation schematics. For a nerd like me that provides endless possibilities. I need to practice heat gun skills on this dead 580 I have As for what I would like to see, I would love a metal chassis. My last laptop had one and it just made the laptop so much more durable and just feel high quality. No one makes them though for high end non-business laptops. Some people might not call it a laptop, but a P570WM would certainly be interesting to mess around with. Update: found out that the m18x does in fact have a metal chassis. I'm jealous

Saw the new star trek. It was pretty damn good. Like Brian said there is the one part towards the end that is painfully predictable, but it is still a great movie. I really wish that abrahms would drop the silliness that's needed to keep a movie PG-13. No swearing. Tons of people getting killed, but that's OK as long as they don't swear! Plus the people getting shot just falling down thing is odd too. I can never tell if they are stunning or killing people since the effects look the same. Just make it R, or ,maybe make the theatre version PG-13, but have that version and an R version for DVD? It wouldn't take a whole lot more work, but it would make things seem much more "real". I still find it amazing that the matrix got an R rating, but the dark knight was PG-13. What is 17+ in the matrix, uhhh.... they say shit in the matrix! R! Dark Knight has tons of terrorism and other crazy shit, but no one swears so PG-13!

I was thinking his GPU performance seemed a bit low for a 1/7th increase in core performance. It looks like maybe the power monitoring chip really does do something (bad) on the 780m, unlike the 680m. Hopefully this does not require physical mods like the desktop 680 does to circumvent. I assume that a 680m BIOS will not work since the core VRM and power monitoring chip were changed?

When detecting high power draw the P150EM will begin clock modulation throttling on the CPU. I have run throttlestop for months to kill the throttling and the laptop still works fine. As for temperatures, clevo has a poor manufacturing method of both the CPU and GPU heatsinks that results in them being uneven where they must make contact with processor dies. You can knock over 10C off temps by lapping the heatsinks. At 1.05V 1006Mhz with the fans on auto I only reach 84C in heaven with a room temp of 72F. I did some other cooling mods too, but the majority of my improvement came from lapping the heatsink.

This is my best vantage: http://www.3dmark.com/3dmv/4682745 3dm11 won't complete due to PSU shutdowns on the combined test with the memory overvolt despite a 330W psu. Working on a way to get the laptop to stop telling the PSU to turn off. My memory could only do low 4600s with no modification.

How much paste did you put on? It sounds like you need to raise die pressure, or your heatsink is uneven. Or maybe MSI laptops just have underpowered GPU cooling?

I didn't expect to see modded BIOS already to overclock those things. What GPU voltage are you running, and I assume 1000MHz is really 993MHz, or is the 780m not limited to 13mhz increments? I'm surprised at the memory clocks. I assume your memory is running 1.5V, but when I start approaching that voltage with my 680m, my memory starts overheating and clocking worse even with the cooling pads replaced with copper shims. Even with overvolting the highest I have seen anyone get a 680m memory to is just under 5400, and you're at 5800 with no mods I assume. Well I think this might mark needing to do more extreme mods to my 680m... (more capacitors). Also your CPU physics scores are very low for 4.4ghz. I get over 10k at your clocks. Maybe it's the 1333 memory but that seems like a lot to be caused by memory.

I don't think you understand what I am asking. When using the dedicated gpu (680m) the HD3000 controls the display, while the 680m sends it what data to show over the PCI-E (optimus). The p150em service manual has the display port circuit diagram specifically labeled "dedicated GPU displayport", and that along with hearing of issues with displayport and intel IGPs makes me think that it is possible that the displayport bypasses the HD3000. You can tell if it is bypassed or not by what driver lists the displayport display. For example here the nividia driver lists no displays, but the HD3000 driver lists the internal LCD. If the HD3000 was bypassed, the nvidia driver would list the display.

I find it interesting that your minimum framerate did not improve with the BIOS change, but max framerate did. You really are noticing a smoothness improvement though?

I once messed around with an LCD's refresh rates. I could only hit 64Hz before I started losing frames. It was a toshiba laptop though.

I thought about getting an extra gpu fan and fitting it in on the CPU side. I think it's just a little bit too big and the cutting down that would be needed combined with being lower rpm would lead to no benefit though, so I never tried it. Other than that the fans already move a lot of air. My only issue with them is that the CPU fan needs higher RPM to do it, and it blows air pretty unevenly.

Again I am not very familiar with the cooling of your laptop, but it should certainly not be hitting the throttle point on stock volts. Repaste it and maybe up die pressure and see if it improves.

Because with the default BIOS Kombustor will not push the card much. Games will push it even harder. I'm not very familiar with that laptop's cooling effectiveness but 90C is very hot for default voltage so I suspect a die contact issue.

What laptop is this?

It sounds like you plugged the external display into the laptop instead of the card.

svl7's bios disables the throttling for kombustor causing it to get the card hotter than anything. Games should be several degrees cooler.

Max temps before throttling are 105C for ivy bridge CPUs and 90C for nvidia GPUs. For AMD APUs max is 100C. IDK about their descrete GPUs. CPUs can run hotter since they are socketed. Heat can't cause solder balls to flex and break like they can for GPUs. I used to run a sandy bridge over 90C a lot and I never had any problems. The only question is if the CPU cooler also cools other chips like FETs for voltage regulation, in which case you probably should keep temps under 90C. Higher temps do mean there is a higher likelyhood for parts to fail, but many people have laptops that hit throttle temps regularly, and that hardware is still alive today. There really are no "killer temps" since built in throttle points kick in before then. The only thing to look out for really is if a part is known for being prone to failure, such as back in 2008 for the 8600m gt. People saying they prefer lower temps and mostly just being conservative and their hardware would be fine if they ran hotter anyway. Personally I like to stay over 5C under throttle points.

Wow fps dropping into the 40s with 2 titans at 1110. High screen res but still. This reminds me that I never finished 2033. And stop trying to pick up virtual chicks.

Has anyone tried using their P150EM or P170EM's displayport? I'm thinking that it may actually bypass the integrated GPU and is hooked up directly to the dedicated graphics card, thus giving higher performance. If anyone is using the port, could you please check to see which graphics processor it connects to? If it shows up on the intel GMA control panel then it connects to the iGP, but if it shows up in the nvidia or amd control panel then it's connected to the dGPU.

So the HD4000 is permanently disabled? That's very odd. I was not aware of manufacturers still doing that since it can be used to extend battery life. Are you absolutely certain that it is disabled? If it is then it will be difficult to get good eGPU performance, especially on internal LCD. @Quadrider10 No it's really only 3dm06 that performs terrible for nvidia compared to ati/amd, so don't switch cards. Still looking into your PLL.

His scores are in line with no optimus. 3dm06 performs particularly awful with low PCI-E bandwidth for nvidia cards unless optimus is working.

The PE4H runs pci-e 1.1 just fine. When I used a PE4H I was getting 22-24% PCI-E overclocks before I got instability. Maybe a PE4L could have gone, higher, but that was already pretty good. Post your PLL number and I'll see if it can overclock.