S-XPS 1645 Throttling Info. and Updates

BTW I'd like to have the original email forwarded to me. [email protected] Maybe someone can check the BIOS out.

Also you wouldn't get in trouble if you didn't sign a NDA or similar.

 

Well isn't that nice of them? The plot thickens...

I look forward to the "does anyone have a floppy of DOS 6.22 I can use to flash the A3 bios back" thread....

 

Hmm, that's a fix isn't it?

 

I have noticed that the sound issue that I had seems to have gone away while playing Sims 3 (the crackling sounds). I also haven't had my system lock up yet so it seems to be some improvement. It may take longer to see if my system still randomly hangs.

 

Here is a link for the A03 Bios, just scroll to the 1645.

ftp://ftp.dell.com/bios/

 

What if I want to go back to A3 after trying this one and (maybe) finding out that not only is the problem not solved but the machine now wont allow TS to run or it won't boot when a 130W adapter is connected?

I'm possibly overreacting because I own a router (DIR 655) where they did a firmware update that added software to the router that redirected web browser requests to a site selling software AND added a "feature" that disallowed you to go back to the old firmware once you realized what was happening.

 

But it sounds like if I don't like this new beta bios, I cant go back to A3 using the winphlash software Dell provided. Maybe I am misunderstanding something here. Maybe the official release of the new bios won't have this "feature"

 

UncleWebb, can you read the release notes in the X29 version? Just curious your thoughts on the fixes they made. I am a technical person, but it is above my level of knowledge.

 

Beta BIOS installed and working here; I got a 130-watt adapter and I'll test if it boots or not.

 

BTW all I had to do was run 1645_X29.EXE right? Or do I have to run 1645X29.EXE as well?

 

I would wait as I also am not sure, better to be safe than sorry, I am sure they will release it soon enough through there website.

 

Boots fine with 130-watt adapter, and as before it properly detects you're using a 130-watt adapter in the BIOS.

 

Mirror: http://dl.strags.com/R254303.zip

 

I was told one is for Windows 7 and the other is for vista. The underscore one is the one I was told was for Windows 7.

 

Does it still throttle with 130 watt PS without the use of TS?

 

I flashed with 1645_X29.exe

So for the test :

TS OFF - AC 130w - Furmark + Prime95 -> THROTTLING coeff 7x
TS ON - AC 130w - Furmark + Prime95 -> NO THROTTLING coeff 12x but the laptop SHUTDOWN SUDENLY AFTER 30SECOND dues to high temperature or protect comsuption I think...

Now the draw with Furmark + Prime95 is ~145w (with X29 bios). Before the flash (A03) the draw was ~135w. Very strange.

I look now that the temperature of the GPU jump to 95° before the shutdown... And 90° to the core after 30 seconds !!! I remplaced the thermal past by AS 5 and before update the max temp of the GPU was 85-88° and the CPU 83-86° with Furmark + Prime95 after 5-10 minutes.

BTW, Dell will release a 1645 with 57XX according to the readme.txt :

 

fr0x: Your testing seems to show that this bios should be avoided. It also shows that a 150 watt adapter might not be such a bad idea.

Stuff like that in the documentation sounds good but I don't have enough knowledge to fully understand if that is a good or a bad thing.

 

I don't mind a bit of throttling while run p95 and furmark, it's a bit of an abnormally extreme load. I care about throttling while gaming or transcoding a DVD.
Could someone that was brave enough to try the new bios see if you can game without Throttlestop and still get good FPS?

 

You guys running into 90C+ temps - seriously get some Artic Silver 5! I barely go into the 70's with maximum load. It really will help your system out, run a lot cooler, and run more stabily. I am able to run Furmark and Prime95 together without the system rebooting. The system performance drops really low though, I think the load might just be too great and causes inaccurate response in these tests.

 

Here's my theory on Furmark and people's expectations of the Mobility Radeon 4670 - it's not that powerful of a card and this is a laptop. There might be some major congestion/bottleneck other than cause by throttling that is triggered when both Prime95 and Furmark are ran together in their maximum settings. I'm able to run Furmark 640x480 8xAA @ 240 fps stable, while running Prime95 with 4 threads working. The multiplier stays locked at 13.2 for me and my 820QM. That's pretty decent IMO. Going up in resolution though in Furmark drops the multiplier and rendering of even the Windows UI drops to 10-20 fps. I think this is the best we can expect from this system, especially with such a low-end graphics card.

 

I tested 3Dmark Vantage with the new X29 and TS ON :



Same score with A03

Very strange my temperature issue with this bios. And yes, with Furmark and Prime95, the performances drops

EDIT:

I tested with Crysis :

WITHOUT TS - AC 130w : THROTTLING and it's realy unplayeable. My kill-a-watt jump to 91w and after drops to 65w when the cpu throttling.
WITH TS - AC 130w : NO THROTTLING. My kill-a-watt jumps between 110 and 120w during the game.

 

BTW Windows still warns me that the hardware is limiting processor performance. Doesn't look like a fix then:

Code:

Log Name:      System
Source:        Microsoft-Windows-Kernel-Processor-Power
Date:          1/12/2010 3:48:31 PM
Event ID:      37
Task Category: (7)
Level:         Warning
Keywords:      
User:          SYSTEM
Computer:      chipmunk
Description:
The speed of processor 0 in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
  <System>
    <Provider Name="Microsoft-Windows-Kernel-Processor-Power" Guid='{0F67E49F-FE51-4E9F-B490-6F2948CC6027}' />
    <EventID>37</EventID>
    <Version>0</Version>
    <Level>3</Level>
    <Task>7</Task>
    <Opcode>0</Opcode>
    <Keywords>0x8000000000000000</Keywords>
    <TimeCreated SystemTime="2010-01-12T23:48:31.397787100Z" />
    <EventRecordID>16181</EventRecordID>
    <Correlation />
    <Execution Processid='4' Threadid='52' />
    <Channel>System</Channel>
    <Computer>chipmunk</Computer>
    <Security Userid='S-1-5-18' />
  </System>
  <EventData>
    <Data Name="Group">0</Data>
    <Data Name="Number">0</Data>
    <Data Name="CapDurationInSeconds">71</Data>
    <Data Name="PpcChanges">38</Data>
    <Data Name="TpcChanges">0</Data>
    <Data Name="PccChanges">0</Data>
  </EventData>
</Event>

 

fr0x: An increase in core temperature after a bios change, CPU or GPU, is usually a sign that at least one of them is getting more core voltage with the bios update. Maybe an engineer had a good idea to boost the core voltage a little to make things more stable but that only results in more heat, more power consumption, more throttling and more problems.

Even if just the GPU got a boost in voltage, the extra heat would heat up the CPU as well. Does any software report the core voltage of either the CPU or GPU?

 

Meh....CPU-Z tries with CPU voltage, but I've never seen anyone quote that as "accurate". GPU-Z reads voltage only on nVidia desktop cards, IIRC.

But it'd be worth a comparison. And make sure people are using the same version of CPU-Z...that's been known to cause problems as sometimes it reports the VID and sometimes it reports the actual voltage.

Wait...I think OCCT can check the voltage (12V+ and CPU) during its stress test.

~Ibrahim~

 

Got my PA-15 in the mail today. This thing is a massive brick compared to the 90W unit that came with the 1645 to begin with. Its kind of sad and funny actually. I wouldnt recommedt it unless you can hide it under a desk

bios recognizes it as 150W and the battery charges happily.

 

So i guess so far, even this newer bios update doesn't help anything
in fact it looks like it makes it worse?
i dunno but i should be getting my sxps 1645 tomorrow or something this week
hopefully something special haha fingers crossed!

 

Yeah it sounds like things are now materially worse.

I appreciate the brave people offering up their 1645s so Dell can run test code on them, but if this is what Dell has up their sleeve as "guaranteed to solve the problem" (someone here quoted a Dell rep as telling them that) then I think we have a pretty good indication both of what Dell will ultimately do and of what Dell thinks of us.

Looks like A3 bios and Throttlestop run at startup or send back your laptop are your options. I know this wasn't official firmware, but it proves Dell isn't even trying. I expect to be ok with TS because I only want to run my rig about "10%" over what it wants to run at anyway, but gamers aren't so fortunate. What are the long term effects of pulling so much juice through your motherboard?

Maybe we can get $1000 together between us and get Unclewebb to see about writing us some custom firmware! I once added a wireless network card to the bios whitelist on an HP laptop, but this power management stuff is a little beyond my abilities....

 

How is releasing a BIOS with a handful of changes, most directly related to throttling, seen as "is not even trying"?

I mean, yeah, the new unreleased BIOS doesn't fix the problem, but we are moving somewhere and there IS evidence Dell is trying: capturing the laptops, responses to these threads here, this new BIOS.

I think "is not even trying" is an unsubstantiated qualification.

I know end-results are what matter and this problem shouldn't even be happening, but let the COMMUNITY at least be truthful, even if Dell isn't. I also know you may think they're not trying hard enough (hell, we all are), but there are still a few people working hard (Bill included).

All that aside....no updates from the Dell engineers about my captured system...last email was four days ago...the last three I've sent in those four days have not been replied to. Though, only one was really a question, the other two were just more info on the issue and didn't require a response.

~Ibrahim~

 

So let me get this straight:

-Still throttles
-Runs hotter
-Can't go back to A3

SOUNDS LIKE DELL REALLY FIXED IT TO ME

 

.....if this was an official release, I'd be right there with you....but that's like telling the doctor half-way through the surgery that he forget to staple your ribs back. He's getting there...just like Dell.

I think, and maybe this is just me, that Dell is making progress. And that BIOS is still nearly two weeks old. They didn't even have my system yet and for all we know they just gave this to Viper so he would STFU and stop calling them (no offense, Viper, I for one appreciate your persistence).

~Ibrahim~

 

yeah but if you send it to some one that isn't part of your testing group (i.e. dell employed) then its GOING to get spread and leaked out.

It really makes you wonder if Dell has any idea what the problem really is.

 

I think it all looks very "unpromising". While it's possible that this is yet another minor misstep by a company that just isn't on top of the whole "public relations" thing, I tend to think that not much is going to happen now in the absence of some kind of major media escalation (Tomshardware?).

Dell reps have been quite insistent that the throttling issue would be resolved without an upgrade to the 90W adapter even when confronted with the "physically impossible" argument. This was their chance to prove they could do it by at least making things a *little better*.

Funny thing is, assuming TS + 130W is a viable long term solution, this machine (with RBGLED and blu-ray burner) is still a giant killer at it's price point. Maybe we should all donate to UncleWebb and get on with our lives?

 

The entire situation is very weird isn't it? How many potentially bricked laptops are they going to get back if there really is a mechanism to stop you from going to back to A3 and people try to do it anyway because the laptop runs hotter now?

There are undoubtedly a few really good people at Dell working very hard to help us, but the "command and control" part of their tech support arm just isn't functional.

Very sad, because this product is otherwise sooooooo close to perfection in my opinion.

 

Dell's not trying, theyre either covering up or shooting in the dark. There is no reason to give someone a placebo bios to shut them up, its risky enough as is. Their engineers are either very retarded, or they are covering it up. Its been said before they think the 90w adapter is plenty enough, by more than just the brainless reps. This test bios is and indication of what we predicted earlier based off their history with the latitudes, a tweak to some profiles and a mission accomplished marketing campaign.

Your not the only one that has their computer captured. Soon as my guy sent the waybill, that was the last i heard of. When i persist to try and talk, I get passed on to reps that know nothing now.

the 90w adapter just isnt enough, if the tests one of the users did was correct earlier. the computer probably is only pulling 70w from the adapter.

 

LOL. I don't think they just gave it to me to get me off their back. They are really trying to help and I guess some of their customers are having success with the new update(probably the less technical customers). The guy at Dell I am working with seems to be very sincere and apologetic for problems I have with my system. He took screen shots of the Prime 95+ Furmark after I installed the update to bring back to the engineering team. he promised me an update/reply from engineering by the end of the week.

Also, the reason I am being so persistent with Dell is that I want to get visibility with upper level management at Dell. For all I know, they know nothing about this problem and how severe it is. My hope is that if I call them enough, it will get to the right person who can actually make this happen. At the same time, I question Dell's ability to run their company effectively based on the fact that all the management is in the US but everything is done elsewhere in the world.

 

I strongly suspect that it will still be possible to downgrade -- just not through the normal (simple) method. There's usually a workaround for this sort of thing.

For example, here's how someone got around a "can't downgrade BIOS" message on his Dell Mini 9. (Note that I have no idea if that specific method will work for a X29-to-A03 downgrade -- it was just the best example I found in two minutes of Google searching. Wouldn't surprise me if it did, though.)

 

@unclewebb, can you explain the difference between the following three log lines for throttlestop, especially the first two compared to the last?:

DATE TIME MULTI C0% CKMOD CHIPM DTS

01/12/10 21:25:50 7.00 44.0 37.5 100.0 46
01/12/10 21:25:50 7.00 42.5 50.0 100.0 45
01/12/10 21:39:22 7.00 91.1 87.5 100.0 42

Thanks.

 

renz that shows some heavy throttling, not only is it locked to a 7x multiplier, but looks to be throttled even more by modulation. Its slowed down majorily, imagine usain Bolt trying to run down the track gagged and gasping for air.

Other news i updated the main page and hopefully improved it a lot. Give me some more FPS etc type info on this beta bios and ill include it. Anything else you think we should have on there let me know as well.

 

renzska: All three lines show various levels of throttling. In each line, the multiplier is reporting a reading of 7.0. A Core i7-720QM should be using a multiplier of 12 or higher when all 4 cores are fully loaded. 7 divided by 12 means the Dell throttling scheme has your CPU running at 58.3% of its rated 1600 MHz speed. That's just from the multiplier reduction.

Your CPU is also suffering from clock modulation which is another scheme that Dell is using to reduce the power consumption and performance of your laptop. In the first line, clock modulation is being reported as 37.5%. To put this another way, for every 8 clock pulses going to your CPU telling it to do some work, the bios is telling the CPU to take a break so the CPU is ignoring 5 out of 8 clock pulses. It is only doing some work every 3 out of a possible 8 pulses.

133 MHz X 7.0 multiplier = 931 MHz
931 MHz X 37.5% = 349 MHz

At that moment in time your CPU is crunching numbers equivalent to a Core i7 running at approximately 350 MHz. That's pretty pathetic considering what you paid for your laptop.

In the second line the clock modulation is up to 50% so you're running at about 465 MHz. The final line has clock modulation up to 87.5%. In a normal desktop computer that doesn't have any of these throttling schemes, clock modulation should always report 100.0%.

The Chipset Modulation isn't being used on the XPS 1645 so your log file should always read 100.0%. So far that has only been seen in use on the Alienware M15x but I wouldn't be surprised if it has been used on some Core 2 laptops as well.

The throttling Latitude E6400/E6500 that started "ThrottleGate" used basic clock modulation combined with multiplier reductions just like the XPS 1645 is using. This isn't some sort of freak occurrence that Dell doesn't understand. They know exactly what's going on. I assume they were just hoping that no one else would notice this. ThrottleStop monitoring put an end to that.

There was no need to inconvenience laptop owners by "capturing" their laptop. Grab any laptop off the assembly line, plug in an under powered 90 watt adapter and throttling is the result.

The C0% is a measure of how hard your CPU is working internally. When there is no clock modulation going on and all 8 threads are fully loaded, this should read 100.0% on all 8 threads.

As clock modulation increases, the Task Manager will continue to report that your CPU is running at 100.0% but the C0% will drop lower and lower indicating that it really isn't accomplishing as much as the Task Manager is trying to lead you to believe.

The DTS column is a direct reading from the digital thermal sensor on core 0. It tends to be the most accurate. In line 1, Dell is throttling your 1600 MHz laptop down to the equivalent of 349 MHz even though it is still 46 degrees away from the thermal throttling point that Intel has set for your CPU.

I included this last column to clearly show that the throttling of this CPU has absolutely nothing to do with the core temperature of the CPU. It's not Intel slowing your Core i7 down. Dell is 100% responsible for the slow down. I'm surprised that Intel hasn't come forward and said something yet. This slow down is giving the Core i7 mobile chips a bad name and some users are probably going to point the finger at Windows 7 and start comparing it to the early days of Vista.

 

A formal review of the studio 1747 on the credible notebookcheck mentions the power problem (not in details though) and this will for sure get their attention as all the reviewing websites will target this in the future.
link

 

I am receiving my 720qm + RGB LED 1645 tomorrow. Reading the above CPU throttling info by ubclewebb leaves me thoroughly un-impressed with this laptop. Also, I have some low hopes from dell on fixing it. Those who have it, do you think it is still a keeper?

 

khaledseif: Can you do the readers there a favor and post a link to ThrottleStop so they can properly review this new laptop? The tools of yesterday like SpeedFan are useless to monitor for throttling.

http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip

Edit: Thanks khaledseif.

 

Why not wait till you get it, open it up....install whatever...do whatever it is you are going to do with it....and tomorrow nite or so, you let us know how its going so far and if its a keeper or not.

 

yes, it's a good idea. I will give it a try and post my comments here. I got a good deal on it, so it would be great if it works out for me.

 

Still waiting on Dell to call me back on my RMA.... not a keeper IMO, as Dell could easily invalidate our warranties for using TS and 130watt adapters if they wanted to.

 

Recieved my laptop yesterday.

For anyone that wants to see whether their laptop is throttling or not before installing throttlestop or any other monitoring program, simply install Crysis Warhead and fire it up on fairly high settings. Watch in amazement as the system throttles/doesn't throttle while you remove the power supply and plug it back in again (change from battery to AC power). It is quite remarkable. I will try and take some footage and put it on youtube since it makes the problem so back and white. Maybe use the youtube link as a point of reference when contacting dell?

I fortunately purchased a 150W adapter before my laptop arrived and with TS this machine runs like a dream at this price point. I'll be donating to unclewebb and keeping the system because a) I don't game that often and b) when I do i will use TS and c) it looks and feels the business.

I'll be contacting Dell anyway to get themn to send me 130W adapter to replace the one that came with the system so I have two viable supplies should one go walkies... I'll be doing so under the false pretence that I will return the system if they don't do as I say of course!!....

Thanks again unclewebb.

M

 

I got my laptop yesterday as well. Its a beautiful solid laptop i have to say apart from the smudges you get when you use it, which is kind of annoying.
My BIOS is A01 and im not planning to upgrade it either way, why meddle with something that is working correctly. The screen is amazing and i dont see any dead pixels or colour contrast issues.

Plus im not a gamer like i said that's PS3' job.

Let me know if you manage to get the 130w PSU

 

Same here, I have a PS3 and 360 sat next to each other so its probably only going to be a bit of CS Source every now and again

As fas as CS Source is concerned, it runs superbly on 1080p/everything maxed out on video settings. Solid 100fps and above! Looks the business too! Crysis warfare at 720p/everything on gamer is a solid 30fps i would guess which I think is pretty awesome. Wasn't the crysis engine designed to take advantage of multi-core cpu's? Maybe thats why it performs so well on these i7's....

All of the above with TS of course...

I'm on A01 as well and will stay there as long as no problems arise or a fix for throttling magically appears!

The screen is indeed amazing and no dead pixels here either. I would reccomend downloading the hubble 3D imax trailer off apple movies in 1080p to really show what it's made of

It's early days but overall I'm really happy with my new piece of kit

 

I don't understand Dell's testing procedures very much either: why ask someone to test it when you have at least two captured laptops, captured for THIS EXACT REASON.

@unclewebb

No one, including the PR guys who captured the laptop, knows why you can't grab one off the line.

~Ibrahim~

 

If I were you, I would call Dell to get the RMA. I shipped mine two days ago after getting the RMA last Thursday, and I received the refund today in my Paypal account. There were just too many issues with this machine IMO, I have a Sony F11 on the way.