S-XPS 1645 AC Power Throttle Issue Investigation

[Zlog]

It just occured to me that my UPS has a little display on it that will show real time power draw. When i get home tonight I'm going to unplug everything from it and just have my laptop plugged in and I'll run some tests as well to see what the wattage draw is.

If the draw is a consistant 75 watts plus in most situations then it'd be a good assumption to think that the adapter does not have very much headroom for any little spikes in draw, which could cause stuttering in games at times and lower scores.

The way I see it (if it IS a power draw issue):

Game Scenario: You're in a hallway indoors in Left 4 Dead 2, you open a door leading outside thats full of zombies and a lot of things happen all at once:
-Zombies attack, AI processing via CPU
-textures loaded from RAM, maybe HDD
-GPU suddenly goes from rendering a very small area to rendering a very large area full of monsters, plus effects from gunfire, fire, explosions, etc.
-You're moving your view/camera around a bunch which is forcing all of the above to be loaded and unloaded and altered based on perspective

Any single item happening is going to cause at least one component to suck some extra power, and may not trigger the issue. The scenario above all of these things happen at once causing your GPU, CPU, RAM, HDD, and everything in between to suddenly say "MOAR!!!!" which could cause a temporary down-throttle that stutters the gameplay and sound.

Or maybe its really just a driver or power management feature causing the CPU to not work properly. We still have yet to get a real confirmation in the form of some one testing in North America with a larger-then-90watt adapter.

[Ollie222]

OK, one more 3DMark result and that is

AC adapter - Battery in laptop 62% at start, 63% at end - Max brightness - Wifi on
3D mark – 6882
SM 2.0 – 2465
SM 3.0 – 3026
CPU score – 2795

So slightly lower scores but not by much. I think there is still a good that it's the power supply and it's just that my machine has the wled, ssd and no bluetooth so is probably about the most power efficient 1645 you can buy off the shelf. Perhaps I also have a good power supply although slightly contradictary to that is my power meter has showed a maximum supply to the ac adapter of 86w in that last test although it's possibly a few watts higher on average than when the battery was charged.

Also my power meter is only a cheap £20 Maplins item so it could be that it is not that accurate.

[khaledseif]

Originally Posted by Ollie222

OK, one more 3DMark result and that is

AC adapter - Battery in laptop 62% at start, 63% at end - Max brightness - Wifi on
3D mark – 6882
SM 2.0 – 2465
SM 3.0 – 3026
CPU score – 2795

So slightly lower scores but not by much. I think there is still a good that it's the power supply and it's just that my machine has the wled, ssd and no bluetooth so is probably about the most power efficient 1645 you can buy off the shelf. Perhaps I also have a good power supply although slightly contradictary to that is my power meter has showed a maximum supply to the ac adapter of 86w in that last test although it's possibly a few watts higher on average than when the battery was charged.

Also my power meter is only a cheap £20 Maplins item so it could be that it is not that accurate.

Exactly 86 Watt maybe the maximum power provided by the power supply. Even though 3Dmark is not sucking the maximum power, the charger provide its maximum and hence will not be able to provide more power to the extra work.

[khaledseif]

Originally Posted by Zlog

So Khaledseif when you ran this set of tests:
http://forum.notebookreview.com/show...25&postcount=6

was it just Crysis looping the bench, or did you have other things running to max everything out? Which programs?

This test was performed on Crysis alone.
Further tests, the one with the pictures was performed using Crysis and prime95

[ikjadoon]

So neither Ollie nor Khlad (who both have the WLED) experience the 3DMark '06 issue.

Khaled, however, does have CPU throttling when running Crysis/Prime95 when on the adapter.

OCCT PSU test won't work on ATI GPUs: it requires CUDA, a nVidia feature.

Oh, BTW, I've run some additional tests and I don't think turning EIST actually lets you hit 13x. The C0 state (how hard the processor is actually running) is pretty low and CPU-Z confirms multipliers low as 5x when i7 Turbo.exe reads 13x constant.

[Ollie222]

Originally Posted by Zlog

So what it boils down to is you don't honestly have the issue that we're having.

Dammit. I'm lost then. The only difference should be that you're AC Adapter is getting a different input power then north american adapters would, right?

So it doesn't entirely rule out the adapter theory, but it still makes my brain explode :P

Can you install Crysis or Far Cry 2 and run the benchmark loop on and off of AC power and look for stuttering? Man im confused

[EDIT] Never mind i forgot that you said you didn't have Crysis... does the demo have a benchmark loop?

Based upon 3dmark alone I don't seem to have the problems that some of you are encountering however Khaledseif makes a valid point about 3dmark not stressing everything out.

I don't have any games here (I'm not a gamer) that I can install but if someone can point me in the direction of a downloadable demo or something else that'll stress the GPU while something like prime95 stresses the cpu and which would give us useful results then I'll gladly do so.

[ikjadoon]

I mentioned FurMark to stress the GPU....

[Zlog]

Originally Posted by Ollie222

OK, one more 3DMark result and that is

AC adapter - Battery in laptop 62% at start, 63% at end - Max brightness - Wifi on
3D mark – 6882
SM 2.0 – 2465
SM 3.0 – 3026
CPU score – 2795

So slightly lower scores but not by much. I think there is still a good that it's the power supply and it's just that my machine has the wled, ssd and no bluetooth so is probably about the most power efficient 1645 you can buy off the shelf. Perhaps I also have a good power supply although slightly contradictary to that is my power meter has showed a maximum supply to the ac adapter of 86w in that last test although it's possibly a few watts higher on average than when the battery was charged.

Also my power meter is only a cheap £20 Maplins item so it could be that it is not that accurate.

If you're hitting 86watt with an SSD and a WLED then we have some solid evidence. Even if the RGB had the same draw, the SSD is going to draw considerably less then a HDD will when active.

Good reports man

[Zlog]

@Ikjadoon: I dont think CPUz is showing the proper states on i7 yet, unless they've updated it recently.

@Ollie222: Crysis has a demo you can DL that does include the benchmark loop. Google ought to pull up the demo DL and the instructions to run it. (can't get a link as I'm at work and that stuff is blocked :P)

[DeathWalking]

Also, UK (that run on 220V) power adapters tend to offer better efficiency in the AC->DC conversion by about 5-10%. You see it all the time on desktop PSUs, so I can think of no reason why the same isn't true here...possible the 90W part is either not delivering 90W for us, or it's delivering slight more than 90W for him.