S-XPS 1645 AC Power Throttle Issue Investigation

Oh great!!! It seems like this throttling practice is rampant in dell.

It seems that it is not just an isolated mistake but rather a policy.


I'm just wondering the reason behind this throttling?

Users will need another laptop sooner?
Reduced hardware costs?

Its defiantly has a shady side to it.

 

Well thanks to unclewebb these laptops can reach their full potential now

 

So if its a policy... then we wont be seeing any BIOS revisions that will address these problems.

 

Well, it seems the throttling is common to Dell's i7 systems in general (M15x, 1645, 1557, 1757).

Not sure how bad it is in the C2D lineup, though if present could mean shoddy BIOS engineers.

 

Can anyone please send me the ThrottleStop application ?
My email address is seshankshekarATgmail.com

 

email or pm unclewebb.....

 

Does anyone know what's the 3dmark06 score when throttle stop is running?

Thanks

 

I emailed him, didn't get a reply

 

I'll run it later and post the results, at work right now.

 

3dmark 06 should be above 7000 when all is working well.

Is it just me or the dell's guy's explanation isnt really an explanation. He said a lot of words, but that was just words. Dells problem is their excessive use of clock modulation, so it is indeed the same between alienware and xps16. Reason theyre modulating may be different though. For the xps 16 they are doing it to limit power for the 90w adapter, Which no one at dell can seem to admit yet, in fact seems the opposite. Giving me false hope they arnt identifying the problem yet, or hoping this will go away. Dont be fooled its a deliberate design flaw they put into these laptop on purpose, the rep's may not understand or care to dig in farther but an engineer somewhere in the back sure as hell knows all this, and what we know.

 

I'm experiencing the trouble discussed in the thread. I live in sweden and bought a SXPS16 with a i7 720, RGBLED..

Is there anything I can do while I still own a 90W adapter?
Lowering the screen brightness, will it help as a temporary fix? Or am I forced to run it on battery whenever I play games?
As far as I've understood the throttleprogram is only good if I already got the 130W adapter, is there any other options or updates for BIOS that could help

The only thing that gave me any results is running my applications on an external monitor only (projector only settings). Which won't be an alternative next week.


I live in Sweden and have been in contact with the support. They have actually called me and given me the latest update on the problem these latest days (talking about dell in Ireland, trying to fix the problem). Though I don't believe that them sending me an adapter before christmas holidays is a possibility, which is why I'm looking for 'a quick fix'.

 

Sorry but with throttle stop you do need a power adapter 130watts or greater at the current moment. Battery is about the only option that gets you playing or running at or close to the way its supposed to be.

Doing everything you can to lower power consumption can only help a bit, but not solve. As when the clock modulation kicks in, its brutal. dim the screen if ya can, disable wifi if you can, and use without the battery in.

 

i haven't tried it cuz i don't have mine yet but i think undervolting it should help

 

READ CAREFULLY: An i7-720 running 8 threads at full load will most likely be limited to a 12X multiplier. Don't blame Dell for this. That is how an i7-720 was designed by Intel. ThrottleStop or an upgraded power supply is not going to change this limit. It's a limitation built into these CPUs by Intel.

The default multiplier of an i7-720 is only 12X. To get anything beyond 12 you have to be using the turbo boost feature of the CPU. Turbo boost is always available unless you are exceeding one or more of these conditions:

1.2 Dependencies / Algorithm
Intel® Turbo Boost technology core frequency upside availability is ultimately constrained by power delivery limits, but within those constraints, it is limited by the following factors:

• The estimated current consumption of the processor
• The estimated power consumption of the processor
• The temperature of the processor

These are the limits set by Intel and your CPU will always be forced to run within these limits.

The purpose of ThrottleStop is so that you can at least use the Intel limits with an appropriate 130 watt or better power supply. With ThrottleStop, you get to live by Intel's limits and without, you have to live with Dell's very restrictive and poorly implemented limits.
-------------------------------------------------------------------------
I'm just working on an option for owners that are presently stuck with a 90 watt adapter. It will let you disable Clock Modulation but you'll be able to fine tune your multiplier so you should be able to keep your power draw under that 90 watt number. It might help you survive with your pathetic 90 watt PSU until you can afford to get something decent. Phone Dell and let them know you're not impressed and maybe they'll send you a free one if you complain enough. A slightly lower multiplier is a hell of a lot better option than Clock Modulation that can bring a laptop to its knees.

Based on feedback, I've also decided to try and add a couple of more features for Core 2 owners. Direct access to Super Low Frequency Mode, half multipliers like 8.5 and adjustable voltage ID. I've already got a lot accomplished on this updated version. All laptop owners are going to have a lot more control over their investment with the help of ThrottleStop.



By the way, I'm not a machine. It might take me some time to reply to your emails. I think users will be very happy with the finished product so try to be patient.

 

Should I skip buying this laptop because of these problems?

 

If you buy this laptop just make sure you get a 130 watt power adapter to plug it in with, as that coupled with throttlestop cures the issues.

 

Im kind of confused, how can I enable turbo boost?? By using less than 8 threads? if yes how can I go about doing that ?

 

The 130W psu is the PA4E isn't it?

 

I'm starting to think this forum could use a FAQ for this issue

 

That's the way it works on my desktop and that is indeed the way it should work on the laptop as well. For normal usage the 12x should be enough. And when needed a little extra power can be unleashed.

That wold also provide a way to measure the true maximum performance possible with the standard 90W PSU, and therefore allows us to better argue the need for a bigger PSU.

Hehe, youre quicker than Dell anyway, so we should all just wait patiently.

It depends on what you intend to use it for. If your main purpose is to do hardcore gaming 12 hours a day, then either wait or get a bigger PSU.

If you're mainly going to use it for office tasks, videos and some minor gaming, then go for it.

I orderd one, being fully aware of the current situation. To me it seems resolvable, and I'm not a full-time gamer (need to earn some money too ).

It has been said before, and I think it still stands. Under normal usage the throttling shouldn't be (much of) a problem.

 

PA4E = 130W Adapter
Dell p/n = JU012

 

daraj: The fewer cores that are in the active state working, the more turbo boost will be available to the remaining core or cores that are active.

The Core i7-720QM can use a maximum multiplier of 13, 13, 18 and 21 when 4, 3, 2 and 1 core are active.

In all Core i7 CPUs, the multiplier is completely dynamic and can be constantly changing. It can transition between the above 4 values hundreds of times a second based on load and how many cores are in the active state.

If you are running a single threaded activity and only a single core is in the active state, the multiplier can go as high as 21. If any background activity kicks in and wakes up a second core, the maximum multiplier will instantly drop to 18. If a third or fourth core becomes active then the maximum multiplier drops down to 13. As soon as the background tasks are completed, cores can go back to sleep which allows the maximum multiplier to increase. This is happening constantly.

Those are the maximum multipliers. You will not reach that maximum if your CPU is being limited by temperature or power consumption which I posted about earlier. I have seen even at an overall load of 58%, the turbo boost feature on these CPUs can start to cycle on and off. I'm assuming that this is the Intel design spec but more testing still needs to be done.

 

Edit: unclewebb was quicker...

Turbo boost kicks in when the processor thinks there is some headroom left. So when it (the i7) is not fully active, when it is not too hot, and when it doesn't consume too much power.

For example. If your machine is completely idle, and you will start a single threaded program (WinRar some huge folder for example), it will quickly jump to the highest allowed multiplier. If a second program (thread) becomes active the multiplier might be lowered, but still be above the default one.

If howerver the single thread heats up the cpu too much, it might also lower the multiplier.

In real life it is quite a clever system that works very well (for me, on my desktop).

 

I am definitely no hardware engineer, but why would it be the case that the 130w adapter is supposedly too much for the SXPS when the Latitudes use it?

"PA-4E - New Original Dell Latitude E5400, E6400, E6400 ATG, E6500, Precision M2400, M4400 Series 130W Laptop AC Adapter"
​

Hell, they even have a 130w adapter for "XPS Gen 2 laptops":

130 Watt 3 Prong AC Adapter with 6-ft Power Cord for Dell XPS M1710/ M170 / Inspiron 5150/ 5160/ XPS Gen 2 Laptops / Precision Mobile WorkStation M90​

 

Ok, First of all that is the funnyest thing i have ever read. at first... then i went back and read all my older posts... i can officially say that i need help. Is it normal to lmao at my older posts. i dont even remember them and why did bunnies come up in my older posts?

Well i just wiped my laptop clean of fingerprints but now they are filled with my prints again .

After a whole session with a poor woman on the other side of the phone listening to me yell at her for more than an hour, this laptop will be sent in soon and i will get the same config for 500 dollars less if i decide to purchase it after they get a fix. I am still contemplating wheither i should get this or not because the pink monkey left again and i think i want to chase him.

This will show you how well the screen performs in full screen. You might want to put your browser to fullscreen as well by pressing f11 and turn your brightness up a little.
www.youtube.com/watch?.php/12QG3A

 

WHAT THE????

I almost got a coma-induced migrane watching that. I think a disclaimer is necessary - DON'T watch if you have a seizure risk! LOL

 

Do a volume check before hand. Ready the mute button.

 

JESUS CHRIST MY EYES! Damn man warn us before you do that, please lol.

 

My contacts dried up in my eye instantly. geez

 

All right....I have agreed to send in my SXPS 1645 for "testing" purposes. Here's the dealio I got from:

Update: Brad is actually my backfill for my old role as Dell Community Liaison for laptops. I took him down to the engineering campus in Austin this afternoon and hooked him up with "my guys." We have a solid plan to raise awareness internally. What we did so far is check out a 1645 from the hardware library to Brad. He's replicating the issue and the NBR community's testing and workaround methodolgy. This is mostly for our anecdotal knowledge, and it's not enough to move the needle inside Dell. So, step 2 was we sat down with one of "my guys" and he volunteered to open a PSE (that's like an engineering investigation request). What we need to submit a PSE is customer hardware that replicates the problem. Not hardware from the factory. Don't ask. Long story. I never got a clear answer as to why we can't just pull something off the line. So, that presents a new problem. When we "capture" a customer's system from the field, we have to first replace what they have with a new box. If you Google around, you'll find there are significant delays getting SXPS 16's out the door right now. The problem now becomes waiting for the customer who agrees to a capture to receive the new unit before we can ask them to ship the old one in to "my guy." But once he does get the customer's box in, he can get the testing done and get the PSE submitted. Once we have a PSE# to reference, some engineer somewhere MUST take ownership and fix it. If they don't, we can escalate all the way up to Michael if we have to and have that PSE# in hand.

So, here's the pitch: You want a new box (that will admittedly still have the problem until the PSE is worked) in exchange for cross shipping on our nickel your current box to "my guy"? If so, hit Brad and me in email ASAP at ......................................., and we'll take it from there.

BB

I blotted the email addresses for privacy.

~Ibrahim~

 

Wow, sounds good, and you get a new laptop too. Lots of interesting insights into internal Dell processes, they certainly have a lot of boilerplate stuff they have to go through to fix things.

 

Theyre "capturing" mine as well. i would agree a lot of hoops

 

wait so you just send them your system and they will send you a new one. and they pay for shipping? they must really not like dell.

 

The information provided by the technical support guy in the following chat is NOT ACCURATE. Please see sleey0 post 1437.

Not to let you down guys, but if you got your unit replaced, you loose the option to returning it later according to their customer support guys when they offered to capture mine a couple of weeks back.
Hope they take action.
"
me: "If it does not get resolved, will i have the option to return the machine by then?"
Agent (Harsha _205295): "No"
Agent (Harsha _205295): "You can return the system only within 21 days from the invoice date."
me: "so basically, you are telling me that if the problem is not resolved, then i need to live with the problem. Am i getting it right?"
Agent (Harsha _205295): "Do not worry, the issue will be resolved."
me: "hypothetically if not"
Agent (Harsha _205295): "We will do the needful."
"
This conversation was just after they told me to capture my unit.

 

Yup. Thankfully, I get to keep the laptop until the new one comes in.

I'm glad the ball is rolling at least somewhere now!

Good luck and here's to hoping they "run out" of 500GB hard drives and just have to ship atlstang and I 256GB SSDs...lol.

~Ibrahim~

@razrkiller: yeah, they pay for shipping.

@khaled: well, I am not planning to ever return my XPS 1645 for a refund. They seem like good guys and are willing to fix any problems I have/will have. Only downside I have is making a 110GB image of my Windows 7 installation, LOL.

 

Good Stuff Man, and you're right it's good to get the barrel rolling .

 

Not true, khaledsief.

It is 21 days from invoice date - your new system will have a new invoice date when it ships.

I have, in fact, returned a replacement that was much older than the 21 days from the original

 

Thanks for the information.
This was the factor i considered mostly for returning it. I thought i would share it.

 

No prob.

The rep that told you that is, by my accounts, retarded

 

DAMN that's good to hear. Honestly, that's all this issue needed from the beginning was someone to be honest with us about how Dell plans to fix it. Whoever this BB guys is, they need more people like him to liaise with the technical community. Dell would have a LOT better reputation if they did. Just reading that e-mail made me feel a lot better since I know someone is listening.

 

I agree someone that makes sense. Good they are looking into it. But it will probably be a long while yet i think to get a fix. They really need to get their act together though, the statements by different reps through this whole thing have been ridiculous. It is good news though they havnt wrote us off just yet.

Hopefully they can learn a lot from our fix's and findings.

 

"The Core i7-720QM can use a maximum multiplier of 13, 13, 18 and 21 when 4, 3, 2 and 1 core are active."

Would someone please fill in the sentence above with the 820 specs.
And what multiplier would i run in the new throttlestop for a 90w PS? (820qm) or would i need to get a kill-a-watt to experiment watching draw on ps?
I am waiting for a 130w to test on.
Really cool of uncleweb to do this for us.
Uncleweb do you also own a 1645?
Thanks.

 

I got my 130W today after Fedex kept yanking me around all week. >_< Hopefully the new realtemp throttlestop thing will work for me as well, I'd like to finally get the most out of this thing.

 

I wanted a 1645 for Christmas but this mess kind of scared me off. With ThrottleStop, there's no reason to be scared any more.
Now I just need some donations so I can afford one.

This nice chart compliments of Intel should answer your questions.
Divide the Maximum frequency by the standard BCLK of 133 MHz and you will have the value of the maximum multiplier.

http://download.intel.com/design/processor/specupdt/320767.pdf



No worries Fenikkusu. ThrottleStop and a 130 watt adapter will let your laptop run the way Intel intended a Core i7 mobile processor to run.

 

I have a 150w power supply on the way, and I'll actually have a 130w power supply on hand too (I remembered the Inspiron 9100 brick was massive and I believe was 130w). I look forward to testing with both of these units.

Did someone around here mention flickering of the LCD? I notice this. I only notice it when the backlight is dimmed. It's very subtle--most times I don't even notice unless I'm really watching for it. I'm not sure if it only happens on battery, but that's only when I've seen it (since I'm usually on battery anyway. Then again, brightness is turned down when I'm on battery).

I'm glad to hear some of our machines are headed in for inspection. At least something is happening.


atlslang, does this mean I've finally gotten my laptop before you?

 

Wow, we've been going through lots of emails back in forth in just a few hours!

Things are moving along swimmingly.

~Ibrahim~

 

Oh, as an addendum...

I think we owe a lot to Uncle Webb. I know he seems to be a late-comer to this thread, but he definitely helped me in this issue initially, probably as one of the first few people to see the issue and without his help that would have been impossible.

So, here I go on my soap box. It's Christmas time (I don't care if you celebrate it or not....I mean, look at my name, lol!) and I think it would be most beneficial for anyone who uses ThrottleStop or Real Temp or i7 Turbo to donate to him for all the time and effort he has put in to help get this issue out there. You heard him a while back when he said he was "low" on cash. If people here spent the $1500 to buy this laptop, I certainly think most of us can chip in some bucks, too.

Here, here!

BTW, Uncle Webb, where is that magical Donate button? I can't find it...lol.

~Ibrahim~

P.S. I am doing this because I feel guilty of the upgrades Dell has offered me? Maybe. Does Uncle Webb deserve a little something back? Hell yes.

 

I might be able to find something in the coffers for UncleWebb

Ibrahim, how did you get Dell to offer you upgrades? Just call and complain?

 

ikjadoon: Thanks for the suggestions.

If you go up to the title bar of RealTemp and click on the About... menu item, a window will open up with a Donate button hiding in there.

I'm not looking to get rich but if this app improves your laptop experience 100% then I think a $5 donation is a very reasonable price to pay for that.

If a lot of users throw some pocket change my way then I can continue working hard for the community.
I really enjoy the job but developing quality freeware software means the pay kind of sucks.

 

wow i was under the impression that the XM model of the i7 family had an unlocked multiplier count which was why it was more pricy as well as a higher base clock. anyway this doesnt consern the QM series where the 720 is dumbed down... alot.

if i had money i would so give it to the guy... not just donate i would ask him to share his knowledge and pass it on.