S-XPS 1645 AC Power Throttle Issue Investigation

well its final fedex will be here tommorrow to pick up my 1645 for a full refund..
no more headaches yay.. I will wait until the power issue is resolved and prolly reorder more then a month or 2 and i will be goin to hp for the envy

 

I hope you have better luck with the Envy 15 than I did. I returned mine because the slice battery had enough play in it to disconnect, or at times, cause the computer to shut down. If you get the Envy, check to make sure the battery locks securely.

 

Stay on-topic in this thread, please. I had to delete a number of condescending and irrelevant posts. I would prefer not to revisit this thread.

Thanks.

 

What? Are you saying they can't cancel orders? I placed my order couple of days ago and was thinking of canceling.

 

They say once it's in production you can't. Some people are getting delays in shipping and have been able to cancel. I got a good deal on this laptop, so I will most likely keep it. I hope this throttling issue gets figured out soon. I feel I should get what I pay for, even though I probably won't stress the system to notice the throttling.

 

B.S. I canceled my order 2 days before it was due to ship.

 

Yeah, I was just reading their policy on the website. I guess one can always return before 21 days, but perhaps with re-stocking fee (unless they admit there is a defect).

Are there steps or procedure posted somewhere about what tools to use and how to reproduce this issue (evidence of throttling)? If somebody could post a link it'd be great. I'd like to run the test as soon as I get my system and make a decistion.

Thanks a lot

 

i just canceled mine yesterday

10 days into the order

 

First off, please don't top-post.

To reproduce the issue, the easy way: Play Crysis.

The technical way:
Turn both battery and AC power settings to max performance. Run Prime95 and Furmark, while monitoring max multiplier, on both AC and battery power. Before running on battery, ensure battery is fully charged. Monitor max and average multipliers using i7 Turbo.

 

Will this problem only show up for during gaming or will I get it during something like photo editing w/Photoshop?

 

Only during gaming, it's not actually a serious issue just the performance isn't as good as we hoped under certain circumstances.

 

It is a serius issue for gamers as we need to play at low light or on external monitors to get the full performance from the laptop.

 

I don't know if it has been mentienned or not, but does your XPS16 successfuly activate turbo mode ? I saw many mentions to "x12" or "x13", but none of that correspond to the turbo frequencies. Does it mean your only benchmarking multi-threaded apps or that you desactivated turbo in BIOS or that the throttleling issue prevent the CPU to reach the turbo mode ?

 

The X12 reference is 133Mhz X 12 which is 1.6 Ghz which is the speed the cpus should run on a 720QM when all are being maxed out.

The problem people are getting is that when the GPU is used at the same time the CPU multiplier is brought down to approximately X 7.

 

Thanks for the info. That helps. Not sure what you mean by top-posting though. Sorry.

That confuses me. How is gaming different than any other application or group of applications pushing the system to max performance? Let's say you'r running a virtual machine with vmware, building/compiling software in the background, running eclipse, running photoshop, all at the same time. Isn't it likely that the issue will be seen in this case as well? This is the info I am looking for. I am not a gamer, but I am a developer and do typically use all resources from the machine. That's my reason for running a quad core with 8GB RAM. What are people's thoughts on this?

 

I have seen this throttle in photoshop CS4, although it is not as easy to get into this mode using photoshop vs gaming. In general CS4 runs good, I had to zoom in and out continously, and generally try to make the video work hard. There may be some filters that use the GPU that might get it into this situation.

I'll run some compilers later this week and see how that works out. Have other things to work on today and don't have my machine with me.

 

This is top-posting.

A. Because it reverses the order in which people naturally read information.
Q. Why is top-posting bad?

Iff you tax the system to it's limits, you will see throttling. That can mean maxing out the CPU with CPU above 90%, in addition to what some people have reported, which is throttling with minimal load on the GPU.

 

Oh, got it. Interesting. Never thought of that. So I guess email replies should go after the quoted original text, and not before. Most emails I see it's actually the other way around. Any way this is a topic for another discussion.

 

This topic is written very poorly due to lack of centralized details, or even a clear description of what the problem is. It needs to be noted that the primary part of focus is the CPU, and the problem are that the CPU is throttled to 931MHz when running a program that is CPU demanding. Gaming and GPU use isn't a requirement, although it qualifies as a CPU intense task.

 

If it is CPU intensive only, the CPU runs fine at maximum. Any GPU 3D load directly affects the CPU and makes it throttle according to how demanding the GPU load is.

 

There hasn't been any substantiated evidence of that, similar about it being linked to power utilization. I've done my own tests to prove it and I'm working with DELL technical engineers on this is directly. Everyone has jumped the ball assuming it is power supply related, it looks more like thermal control. Everything that's been used as evidence could just as well have alternate causes such as more common driver crashes and whatnot to explain the system reboots, and in the case of the 130-watt power adapter, it uses a different profile that might have different thermal throttling thresh holds.

 

You have to put a significant workload on the GPU at the same time the CPU has a high workload to hit the issue, otherwise its not going to draw enough power to cause a problem. Besides none of the scenarios you listed sound particularly stressful for the CPU, especially on a quad core machine. Compiling is mostly disk limited anyway (especially Java I assume), and photoshop is generally idle 99.9% of the time.

 

you obviously havnt read this post fully. but to reiterate again. on a 130w or 90w adapter with a watt meter in the wall. the cpu maxed alone takes 88-89 watts from the wall. what do you think will happen when ya add the gpu in the equation? When firing up a gpu test along side, ive seen spikes over 130w from the wall, on most extreme occasions before throttling.

And yes the 90w or the 130w the power system throttles the power to 90w +-2w regardless of what i do to it. theres evidence

 

Conceivable, but not likely. To wit, were it a thermal issue, we would not be seeing what the people who have measured wattage with their Kill-A-Watts have found, which is when an intense load is placed on the computer it spikes up to above 90 watts and then immediately throttles down, despite the lack of any decreased load. Were it a thermal issue, and it took only the instantaneous spike to cause temps to reach their predesignated throttle point, that would imply that idle temps were hovering just below that point, and a consistent light/medium should have the same effect (it doesn't).

Super OT, but Eclipse is used for a hell of a lot more than just Java.

 

Compiling is mostly CPU based, not disk-based. Packaging is disk-limited. The actual compile process (code into machine code) is very CPU driven.

 

Okay, without giving away my complete lack of understanding as well as unwillingness to read 87 pages of super-tech mumbo-jumbo, how will I know if I have this problem? Is it a problem or is it just something that computer people (those who dwell in the Land of Computers) notice and it bothers them because the 1645 is an expensive piece of thundersex that should be perfect and with this throttling issue it's not?

 

I'm assuming this is Java (yes I know eclipse can technically support more languages, but it's a guess) which doesn't do an actual compilation process, it just generates some intermediate byte-code and it gets JIT compiled at runtime so most of the time its just waiting on I/O.

 

Games i think are mostly affected as they use both the gpu and cpu at the same time. Looking at the same scene in tf2, i can have 50fps - 100fps by changing settings and switching to battery etc.

The degree of being affected seems to depend on your system specs, rgbled or 820qm processor etc that take more power. Everyday operation though the laptop seems to be great. But yes its a lot of money and the laptop has great capabilities to play games and other intensive programs and one would expect it too.

to know if your affected, run the programs you do like normal and use realtemp program to monitor your cpu multiplier. If it throttles down under intensive tasks to 7x then ya have issues most likely.

 

Okay will do. So, are you saying games run better on battery then when plugged in? I have the RGB and 7200qm. Also, when I game it's full screen, do I run that program in the background while playing (without being able to blame slow down on that) or after my session (which is usually 30-45 mins)? Again, I know these are answered in here, but the search function on this site isn't amazing and this thread is super long. I've also never had a machine that I cared about taking care of or that could play any game that didn't come pre-installed.

 

I run the game in windowed mode and look at the program on the side.

but the rgbled seems to be a power sucker. So i would enable fps in your game to view. and try lowering the brightness to see if fps increase. Also try running on battery vs adapter see if you notice any difference. (restart the game before and after, and make sure powerplay etc is set to max performance on battery)

The system seems to be able to pull a lot more power from battery, so as long as your settings are set to max on battery then you probably will see full performance on battery.

millage definitely varies so it would be nice to see what you come up with, under which game as well.

 

Called to cancel and they refused saying it was in production already. So I guess it depends on who picks up the phone and the reasons you give maybe?

 

Alright, I'll check tonight with Modern Warfare 2 and will post up what I find. I probably won't understand any of it, but I'm sure you guys here will decipher it for me. Cheers.

 

Hey atlstang,

I have ordered pretty much the same config system that you have. I just wanted to know that are you facing any kind of issues with the system??
Thanx

 

Yes but not as severe it seems. As others with higher end systems the problem seem much bigger. everything ive played so far is playable, where unplayable slowdowns are very few and far between. Even though its all playable it could be better, as in more fps etc. I think i have noticed some slow down in cad programs like solidworks due to throttling but havnt really put it through its paces and tested.

I do stress, just keep in mind though its mainly when the cpu + gpu is stressed the problems arise. other than the other 95% of the time your using the laptop its just fantastic. It feels faster then my c2d desktop, and the wled screen alone is amazing. That and the frame is rock solid, and the heat is pretty good, in fact as long as ya dont cover the bottom vent under the fan ya can put it on the couch, bed etc.

 

It IS power related, and there has been more than enough evidence given to that fact. I should have my wattage reading device today and I will post an itemized listing of how many watts each part in this laptop consumes just to prove the point.

I have already replaced the heat sync paste in this system and it very rarely goes into the 70c's anymore. Mostly when under full load in games such as CoD2 MW & borderlands it’s around 68-69c. Before it would always be around 78c after an hour of its current full load capabilities, and the system STILL stutters when in games. It’s not a throttling issue due to temps. (Although if they do fix this power issue it may well turn into a throttling issue due to temps because this laptop is currently not able to reach 100% complete system load with its power subsystem. If it could, the 1 single fan in the system would have a harder time keeping this system cool.)

 

Could you please tell me what WLED screen you have (LG, AUO, etc)?

The 1557 I just returned had the nicest 1080 LED screen I have ever seen and it was an LG.

 

Oh thats great news. Thats all i wanted to know. Thanx alot frnd. My wait isnt bad then

 

Hey guys, you may want to run a test similar to the one I ran for the M15X throttle investigation. Here is what I did:


With turbo disabled in the bios, a bit more information is shed on the throttling issue. Notice that each time the 4 cores drop below 100% load, the GPU throttles with them. At peak CPU load (100%), the GPU stays at 100% until the CPU is throttled below 100% and with it the GPU also scales back it's clocks. Normally the GPU cycles far more frequently when Turbo mode is enabled but with it disabled, the rate of the GPU throttling is much slower.

Normally when Intel Burn Test is used by itself (not in conjunction with furmark), all 4 CPU cores stay pegged at 100% but when Furmark + Intel Burn Test are run together, the CPU clocks begin throttling together with the GPU clocks. This seems to further reinforce the idea that this throttling issue is power related. Click on the HD button to view the movie in 720p (in full screen) so you can see the clocks more clearly.

<param name="movie" value="http://www.youtube.com/v/0cWy8rieIJI&hl=en_US&fs=1&ap=%2526fmt%3D22"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="http://www.youtube.com/v/0cWy8rieIJI&hl=en_US&fs=1&ap=%2526fmt%3D22" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width='560' height="340"></embed></object>

 

The hardware id is MONITOR\LGD0215 i assume so?

siphen when ya get it in let me know, and ill re run the same test with you, so we can compare the two side by side, and add more information to the mix. since we both have different systems etc.

edit joker, a lot of us been using prime95 + furmark and came to the same conclusion. but interesting on the disable turbo thing

 

Thanks, and +rep

Yup, that is the same screen, for sure!

 

@5150Joker
Thanks for the video.
It is worse for the XPS1645 because once it throttles under GPU+CPU stress, the CPU hits 7x at stays there, at least this was the case for me.

 

Java intermediate byte code conversion is still pretty CPU based. I was on the IBM Websphere development team (using Eclipse) a few years ago.

 

Dell is reluctant to cancel an order because they think you might like it after its in your hands to begin with. if this isnt enough.. you paid for the shipping right? if your answer is yes then that's money down the drain so all the more incentive for the user to keep it.

From my understanding when the GPU is not used heavily the throttling is reduced and when the screen is dimmed it also helps. this clearly indicates based on the data also by the submitted video (THX) that the computer is lacking of A) power resources or has B) the worst management of power ever known and the motherboard uses rubber to conduct electricity.

so yes in both A and B they are power issues not dead bunnies.

 

Try again, and don't take no for an answer. Ask to talk to a supervisor (and don't believe them if they say their supervisor isn't in).
PM me if you try and you still can't get it canceled. I know a friendly rep.

 

Since this is a power issue do you guys think that an i7 720 qm vs 820 qm or having an RGBLED vs WLED(1600X900)would get rid of or dramatically reduce the throttle issue.

I ordered a xps 1645 with the i7 720 qm and just the regular HD+ Widescreen 15.6 inch WLED LCD (1600x900). Will there still be throttling when trying to run CPU and GPU taxing programs?

 

I've been following this thread for a LONG time and a lot of the groundwork you guys have done has been immensely helpful. Dell still has no official fix which is frustrating but as for people who have tried to get 130 watt adapters and failed I'd recommend trying again, I was turned down on my first attempt but the second time around I was a lot more aggressive and threatened to return the machine and go with another manufacturer (which I think everyone can do since the 1645 is fairly new and no one can have had their's all too long) and lo and behold I'm being shipped a 130 watt adapter which "they are not authorized to ship".

 

I talked with the dell "engineers" apparently and they where pretty insistent on me sending in my current xps 16 so they can test. Makes me believe they either have no idea whats going on, or they have no systems to test and would rather have ones they know have problems, or both.

I finally agreed as long as i didnt go without a laptop, and its a brand new one theyre sending. Im just kind whatever right now, hoping maybe itll fix my cd/dvd burning issues at the same time. Its wierd though i even told them the problem isnt as severe on my system and an easy way to replicated it, but they were still insistent. Oh well, being cooperative i think is better then nothing, maybe well get a fix out of it in the long run.

 

130W doesn't even help so why do you want it?

 

Don't know if it's related, but while typing a paper (Word isn't stressful) my SXPS 1645 just shut down on me.

It had 85% battery, was plugged in [blue light was on], IDK....thank goodness, paper was saved, only missed like the last 2 sentences.

~Ibrahim~

 

Well its my understanding that it helps to some degree and I'm being optimistic in the sense that hopefully with a BIOS update a 130W will actually end up being the fix. Besides that I'm betting my laptop will charge faster (its pitifully slow now) so worst case scenario I just get a faster charger. What's the harm?