S-XPS 1645 Throttling Info. and Updates

Now you just need to mount that light bulb in a thick plastic case and see if the surface temperature gets up to 60C or 70C like some users are experiencing with their PA-3E adapters.

 

This is the fallacy of invalid inference.

Even if the conclusion is true, the premise is not valid.

 

Hello Bill;


I placed and order a week ago for a SXPS1645, and it haven't shipped yet.

Do you know if I will get my laptop with the 130w adapter?, as it is for estimated delivery in february 17.

Thanks in advance.


Iván.

 

Almost like sitting on the fence.. that comment. Don't take it personally, I just find it funny how people can ignore facts. The closest answer I've seen on this is that the techs showed Bill the new numbers and they looked much better - what does this mean?

To me personally it doesn't sound like 'Dell has taken it upon themselves to remove all Dell applied artificial throttling on Intel CPU, because we now trust that Intel actually got it right in their specs'.

The fix should be really simple: Battery does not throttle -> apply this to the Adapter profile. Done. IPCC value may need to exist but they can simply raise it to 125W and if it keeps throttling, then we know 130W adapter is not good enough.

At this stage they've had to raise IPCC (I think that's what it was called) by 50W! That's pretty clear evidence that the requirement for artifical throttling was known and implemented so they can use 90W adapters.

Feel free to correct.

BTW: I've got 1645, so no I'm not a disgruntled 1640 owner who feels they've been shafted. 1640 guys have done the same amount of testing as the 1647 owners. Thankfully their thread is growing so their voices will soon also be heard.

 

I posted this further back in the thread but am still yet to come across a cause or solution.

I recently purchased a 130W PA-4E Dell Adapter of eBay for use with my 1645 and Throttlestop. However when I plug in the new adaptor, my external monitor, a Dell 21.5" model, seems to get interference on it, with 'ripples' flowing up and down the screen (not sure of the technical term). This is to the extent that the monitor is not usable and considering it is my main display, I currently don't use the 130W/Throttlestop solution.

My question is, will the A07 BIOS address this issue and stop the interference at the VGA port. And also, has anyone else experienced similar issues/know of a quick fix?

 

Since VGA interference does not equal processor throttling, perhaps you are posting in the incorrect thread?


So my post isn't completely useless: no, I do not experience this issue.

 

Power consumption comes up on a regular basis so I went back and found an excellent post from page 5 of some real world testing. The measured efficiency of Just Mark's PA-3E wasn't that great.

http://forum.notebookreview.com/showpost.php?p=5677699&postcount=48

 

Sounds more like an issue with the refresh rate... update all relevent drivers and make sure the output settings to the vga port are compatable with the modes the monitor can support.

 

Before you start updating any drivers of any sort as other user suggested, use your old power adapter and see if you are faced with the same issue.

 

Actually, if you read the forum front to back, Bill said word for word that "The system should throttle no more with the next BIOS update in conjunction with a 130W ac adapter.."

I'm not sitting on the fence either, that was pretty straight forward. You cannot draw a conclusion from what he read. It certainly doesn't say it explicitly, and I don't know how you can draw that implicitly.

As far as taking a side, I say this eliminates throttling 100%. A 130W adapter at 130 watts is going to draw around 160 watts of power from the wall if its floored and with a good efficiency of 75% ( normally efficiency drops off a lot after 90% of capacity).

Also, remember you can plug any wattage adapter and the system will scale to that. As far as we know, they figured 90 was enough and 130 wasn't tested with the system yet. Who knows. Lets just wait and see if this fixes it 100%.

 

I can't wait for A07 to be released. Do you guys think it will be posted by the end of today?

 

The update will only release if it passes all the Tests. I really hope it does . Desperately waiting to use this notebook to its power

 

There's no need to wait. Bios A03 + ThrottleStop + a 130 watt adapter will let you use your laptop in most situations at very, very close to its full potential.

Bios A07 may equal that level of performance but there's not much room for it to improve upon that in any significant way. Give the bios guys time to test this properly. Three months have already been wasted so what's another week to wait for them to get this right?

 

I've seen Dell post downloads on the weekend so don't assume that because it's due before Chinese New Year that it can't come on a weekend.

 

Got a call from an Escelations Specialaist today. They said that the Bios should be released by end of day today. Not sure how reliable that info is but it might be something to hope for.

 

I'm not sure what part of Asia the BIOS engineers are based but its passed midnight is most of Asia I think. Unless they work the graveyard shift I don't see the bios coming out today
Hopefully I am wrong.

 

Pleased to get a reply from you and I am a fan of ur software. The only problem is I dont have 130W AC Adapter.
I had a question for you. I have A06 BIOS and all the bios I have used. My system seems to throttle more on battery than the adapter. How can I test and give you the log? Please check

Ok I got it and I am testing. I will post the Log soon.

 

I just played NFS Shift with High settings at 720p. Here are the results.
On adapter its in 10~12 and on battery its worse 7~8 . !!

 

I just had a concern. There was a post someways back saying that the Dell engineers said the XPS 1647 would be fine on the 90w adapter and new bios. I am concerned that this is false due to the fact that the only real difference between the 1645 and 1647 is the processor, and the difference in power between an i7 dual core and quad core is only about 10w. Hopefully the BIOS will work in a manner that allows the system to pull all the power from whatever adapter is hooked up to it and it wont be an issue. If I have to buy my own 130w adapter, no big deal. If the system is not changed...we have a problem. Thoughts?

 

The BIOS that's just got released dint address Throttling issue. It would be after the release of the BIOS for 1645, 1647 would be looked at. So just wait for A04 BIOS and check the performance then

 

In the CCC make sure power play is turned to performence for both battery and AC.

 

Want to address one part of this post: "...very, very close to its full potential..." From what the engineers have told me in conversations and data presented, you should be able to say the same with no TS, with A07, and with a 130W adapter. We do run tests that max the CPU and GPU concurrently, like Furmark and Prime95 do, but the purpose of that testing is to prove that throttling induced to address thermal issues is working as designed. We did not intend for this platform to be able concurrently run max CPU and GPU. To acheive that goal would compromise other design requirements such as fan noise and chasis volume. Our testing of A07 is such to the extent that a user should expect to be able to run a game such as Cyrsis at medium settings and acheive reasonable frame rates.

Generally speaking, if an escalations agent "guesses" at a release date, they will eventually get it right if they make enough guesses. They do not typically have the lead engineer in IM or on the phone when they make those types of statements. I checked with the lead engineer by phone and IM today. As of 1100 CST today, we are not prepared to say "today is the day" just yet.

 

Oh My bad . I just turned on Max performance for Battery. Lemme check now.

 

Gr8. It did run on 10~14. Not impressive but atleast better than 7~8

Thanx man.

 

um.. "not intend for this platform to be able concurrently run max CPU and GPU" then how come this be a XPS system ? If am correct XPS does mean Xtreme Performance System. . IMO it should live up to it's name at least when on AC and high performance mode.

 

Now that is something that I am fairly dissappointed about. I am kind of surprised that the XPS name is no longer associated with performance but for looks and sound. XPS used to be the Gaming/Media king for Dell. Sadly what you are indicating is that somewhere along the line XPS became the "pretty" brand for Dell.

When I purchased this system, it was intended to be a media/gaming system and not a "looks pretty, but dont use becuase of thermal limitations" type of computer. If I wanted looks I would have gone with a HP or other system.

Perhaps its time to retire the XPS name rather than re-invent what it stands for.

 

I would hope that at the minimum A07 allows the computer to run equivalently or better on A/C than it does on battery. I tried looking back for specifics comparing battery benchmarks to throttlestop but the threads are too large to sift through.

 

I think this might be the best expression of a "reasonable standard" for what we want.

I'm prepared to accept that neither this nor any other (slim) notebook is meant to be taken as "a desktop, but smaller". I think it should be able to run "full-out" on a demonstration basis, but you should not imagine that you can leave the thing running Prime95 + furmark for hours on end and have it last for any length of time.

I hope Bill's latest comments don't set off another grand controversy before we can get our mitts on A07 and test it for ourselves.

 

Got the new adapter via the Dell forums very quickly, thanks DELL!
Its a little bigger of course. The cord from the power outlet connects a little awkwardly. Details:- Family PA-4E, Model DA130PE1-00, DELL P/N JU012. If someone wants photo's let me know.

 

It is a media/gaming system, but it's not a Furmark+Prime95 system. I'm saying just install the new BIOS, use a 130W adapter, play some games while running CPUz, and let us know if and to what degree it throttles.

 

Thanks for being honest Bill. That needed to be said.

Bios A07 for the Studio 1747 was just released and some users are already extremely disappointed that they can not run Prime95 + Furmark without severe throttling.

The new 130 watt adapter for the XPS 1645 and new bios should greatly reduce throttling in most applications even if it doesn't cure it 100% while running extreme synthetic tests like Prime95 + Furmark.

 

What's the point of owning a quad core CPU if all the CPU's cannot be used at the same time. What's the point of paying for a video card, that is totally reliant on those cores, if they cannot be used concurrently.

Right now games may only employ 1-4 threads. But games will use more as time goes on, While we will never see a game that pushes this that hard ( Furmark+Prime95), and I strongly urge people not to use furmark. I think Prime 95 plus Crysis/ whatever game should run at the rated speeds. Future use of compute shaders could very well push this that hard combining the use of the GPU for general purpose as well as the CPU. I didn't buy this for boasting purposes. I bought it for usability.

I don't want Dell telling my how fast my 4670 mobility / CoreI7 720 should perform, AMD and Intel has already said how it should perform, I don't care if the frame rate is crappy, just as long as its running at its advertised specifications. Just as long as its running in safe conditions (not 100 degree day in the sun) that is what I expect.

Bill I really do thank you for all your help. But any throttling beyond what Intel and AMD put in to there processors I find unacceptable.

 

Let the games begin.

 

Intel and AMD post their specs with no consideration for thermal concerns for a given chasis. I am saying pretty much what I understand unclewebb to be getting at when he points out Pime95+Furmark is an extreme usage model. This is not the system for exreme synthetic benchmark testing, and as far as I know was never advertised that way, regardless of the name XPS or not. I think if you run any games on the market now or when the system went to market, under A07+130W adapter, we should have nothing to be ashamed of, if all goes according to Hoyle. We did our testing for A07 using Crysis to measure FPS. We used programs similar to Prime95 and Furmark to validate the thermal throttling implementations perform their designed tasks of keeping component temperatures within our spec.

 

Does that mean the throttling scheme will be thermal only? IE would it be possible to run "extreme synthetic benchmarks" at the maximum if the laptop was cooled sufficiently?

 

studio xps does makes it media/ xtream performance(gaming) and Furmark+Prime95 would be classified unrealistic. So, i won't blame dell if my system were to start throttling but it should only be after, HDD max temp is reached as it seems to the component with least temp tolerance.

Considering your above statement, there's a possiblity that it might still throttle. Throttling the systems performace is not bad, but throttling when the performace is need is indeed bad. Hopefully, the next bios revison throttles the system only on thermal limit basis.

Waiting for dell's ingenious solution the bios. It seems there won't be an ETA for this one. It would just pop up out of the blue, will it ?

 

I totally agree with you Bill, let's wait and see what the BIOS + 130w does (Thank You for sending out my adapter). However, on another note, one user recently posted his Furmark+Prime95 results on an HP Envy 15 and there was absolutely no throttling (Uncleweeb can probably back me up, as he commented on those results). I'm not saying that the HP Envy 15 is the same laptop as the XPS 16's and neither am I saying that the BIOS+130w fix might not reproduce these same results on the XPS 1645 when released. But I believe that the HP Envy 15 is HP's direct competition against Dell's XPS 16. I'm guessing HP, as well, didn't build the Envy to run Furmark+Prime95 tests all day, but results show that it does so without throttling (not all day but at least throughout the test period). These are all within the parameters of performance set by Intel and AMD when they built their products, and consumers should at least expect the performance to be perform maxed if they so chose to max it. Now I'm not a synthetic tests person myself, but they do serve some purpose. Dell should be out to best HP every step of the way, especially with a name like "XPS." Now correct me if I'm wrong... Again I'm not ignoring the "fact" (hopefully ) that some fix is coming.

Add: I just saw your post above, but I believe my point (if it exists ) still stands...

 

Well, I'm still confident and think we just need to wait and see. Getting bummed out over something that, probably needs to be said by Bill isn't going to help anything. I'm being educated in the medical field and I understand CYA .

However I don't agree that AMD and Intel have no consideration for thermal concerns for a given chassis. AMD and Intel post the TDP numbers so Dell knows how much heat they have to dispense for the CPU/GPU to run at its advertised speeds. Dell's Chassis and HSF should be manufactured accordingly. They also have there own throttling mechanisms in place to protect the system from overheating.

I agree however Furmark is too extreme, I may have contributed in closing a thread arguing that point. But keep in mind, as a media system this may be asked to do heavy transcoding, which in the future will use both the CPU and GPU to trasncode video on the fly.

I'd like to add, I've seen other games cause throttling much worst then Crysis. Dirt 2 seems to be the worse offender I've spotted so far.

I'd say if it runs with Prime95 ( not technically a synthetic benchmark, as the Mersenne Prime Search maybe important to someone, not to me though). Anyways Prime 95 8 threads + Dirt 2 with out throttling I'd be pretty happy.

 

I doubt 4670 will support anything in the future. Only 5xxx card supports direct computing and opencl. ATI's own converter is low quality and slow.

 

@Bill

I have a question.
Would the new A07 bios perform at full potential even when the battery is at persay 0% and charging.
What I would wish for it is that I would be able to use my computer at max performance while it is still charging the battery when plugged in. Since the current status it barely charges when you use the computer.

 

I think we all just need to see what this new A07 bios will do before we all get worked up about something that's still just talk.

Something to think about, if we could all suddenly use our systems at max capacity with 0 throttling then I bet we'd see lots of complaints about the fan being on all the time and the system getting too hot. There's already some of that going around in it's current state. To get extreme high performance computers they will generate massive amounts of heat requiring massive amounts of cooling. It's just physics. Dell is in a very tough position of trying to make all the extreme power users happy as well as those that don't need it to run at full power all the time while also trying to keep the system thin, small, and light weight.

If A07 comes out and I can do my day to day activities as well as play games with no perceived throttling I'd be happy. I won't worry so much about running several applications together simultaneous to try and peak the system to see if it throttles as that's not how I would normally run it. Now if it significantly throttles when I just run one game with little else going on, then I won't be a very happy camper.

 

Not true, DX 10 cards are fully capable of Open CL. Just will be limited compared to 5xxx series. As time moves on these compilers may or may not get support for the 4670.

 

you said it , "extreme high performance" and that's what the brand name XPS is all about. Think about it.

 

Question is, do your daily tasks require you to use a Core I7 w/ a 4670 to notice the difference. Or would a CD2 w/4670 give you the same performance. I bought a Core I7 for a reason. If I wanted something just good enough to do a given task, I would have bought something much slower.

And no, With no throttling the system should not over heat in normal operating environments, that's what TDP numbers are for man. If these CPU's reach this state then they themselves will throttle. These CPU's GPU's do not produce an unmanageable amount of heat at max load (not including FurMark), the Envy 15, in a smaller chassis handles it just fine.

I'd also like to add, with Throttle stop, and the GPU running max w/ furmark+prime 95, the CPU and GPU have not reached their max thermal tolerances. So I'm not really worried about it. Bring on the new bios .

 

The problem is that Studio XPS 16xx weren't designed for a i7 processor. Having such a demanding and powerful processor under the same C2D chassis will need some restrictions. I know that its wrong that Dell did this, but thats just the way it is for now...

 

Throttling is also present on 1640 and 1647 - both with similar TDP. In which case I cannot agree that this would be valid argument. Besides, customers were never officialy informed about such restriction.

@ Dell-Bill_B
Although we all really appreciate your time and effort put into this case, unless you(Dell) are planning to get throttling fully eliminated - you are just wasting your time. If this remains unsolved to our satisfaction, you will get bad press, disappointed customers and a lot of returned/replaced laptops in the end anyways.

If thermal design is a worry - you should start redesigning the rear vent. If that doesn't help, go a step further and redesign heat sink, etc.

Don't think that we are going to accept semi-baked solution, which would only reduce throttling to some extend.

 

I hear ya. I'm just the messenger, but I will take every reasonable input you guys have summarized and verbatim back to the engineers.

 

One of the main weakness of the XPS lineups is the venting. It has a design flaw that hinders adequate ventilation. Not to mention the LCD screen blocking 25% of the vents opening. Its just crazy.

 

I think if the throttling is reduced to a negligable point its a win... Obviously 100% performance under full load running furmark+prime 95 is not possible. I mean what do you expect from a laptop... if you need more than 95% from 8 threads @ 1.73ghz and keeping in mind the GPU (4670) is only a mid level chip at best, you may want to be using a more suitable desktop.. expectations need to be reasonable.

In comparison to my HP DV8T (which is somewhat comparable) my xps is only 15-20% behind benchmarks for for my HP. If you want i can re-run and post screenshots of the comparison.

I will be satisfied with equal or slightly better results with A07+130w psu.