S-XPS 1645 Throttling Info. and Updates

Yea, But I'd gladly take a performance reduction on battery. But this is arguable I guess, they "can" get max performance without hurting anything. The only way we can is by going on battery and potentially hurting our battery's.

At least, I believe it was HP that cuts the multiplier down on battery.

 

I doubt these have been validated to last 3 years. But the CPU will throttle regardless of what dell asks it to do if its in danger. So THIS type of throttling might not be necessary.

However for you guys using Furmark, your using a power virus. The GPU may throttle to protect itself.

In a perfect world it would be able to perform at max in any given situation, but in the real world there might be things they didn't account for and the throttling is to protect it from such situations.

The way that Dell is using it is very very disgraceful.

 

*shrugs* I suppose that could be true SlyNine. I don't think the Dell XPS 1645 isn't performing to what they advertised. They said it could do some gaming, and over-all multimedia. This guy sure seems to be professing what he got what he paid for WITHOUT TS and using a 90watt:

http://www.youtube.com/user/fred2028#p/u/1/V1BSSQaoyvA

Maybe simply are expectations are higher then they should be? The Dell Studio XPS 1640/45/47 was never truly expressed as high-end for gaming. Perception is a fickle thing.

I think throttling was introduced, because of the pressure of being more 'energy efficient'. Notably if you limit the power consumption, it makes it more energy efficient and get those 'ENERGY SCORE RATINGS'. Anyway, that's my take on it. I don't profess to know the intricacies of throttling, just what I've mainly read/heard here.

On another note, I guess I invited stickler interpretations with that comment.

 

All of this spin is exhausting...

 

I think its the G51J that has some throttling issue.

 

My problem with Dell is that ok you throttle the laptops to keep the temps low, or energy consumption, whatever. If a user wants a cool laptop they typically don't go with a quad-core system and if you're going to try to force "power saving" and lower thermal readings on end users at least give them the choice. Essentially just like there's power profiles for how you run your laptop maybe there should be a choice in BIOS, like "I want my laptop to run cooler and be more energy efficient" there'd be a BIOS for them, and for "I want my laptop to be able to hijack a satellite and run Devil May Cry at the same time" there'd be a BIOS for them because odds are they don't care too much about heat or power. I just think Dell shouldn't be making these choices. Besides that I don't think that was the issue with the 1645, maybe more with the latitudes. With the 1645 it seems like the aforementioned bean counters were trying to save money and avoid the splurging a fortune </sarcasm> on higher rated PSUs.

 

Shipping a high end processor with a power supply which is not able to support enough watt to run the system at 100% has nothing to do with energy saving! Thats just making money in a short term horizon.

Also the xps is still branded as a high end gaming notebook(even on the german homepage) and is still shipped with the 90W adapter.

 

This issue is really annoying learning the fact a proper adaptor would solve this. I'm currently on the market, was planning to buy a sxps 1340, than changed my mind when the 1645 hit me with it's power,after reading this - and other - threads this power fade away in a blink.

Just as the others, I too hope DELL is interested in solving the problem (and not just leave it's powerhouse considered as an expensive brick) - though really don't know why the possibility of a "mass system returning scenario" does not convince DELL of providing better, proper adaptors.

p.s: sorry for my wording, posting on mobile with foreign spelling

 

Don't get me wrong, I don't like what HP is doing either.

I Definitely think that the 1645 should run at rated specifications in virtually all cases. That means that Turbo mode should also be working. But artificial conditions like FurMark make it necessary to implement throttling. But I DO NOT agree with how dell is doing it.

But we need to stop looking at tests conducted with furmark as any indication. Also keep in mind some system may not be performing adequately and using more power then they should.

This is why when people act like the simple solution is Throttle stop and 130watt PSU need to step back. Realize that this bypasses the normal operation of the system. I couldn't thank Uncleweb enough for bringing us the program and exposing the problem in a clearer light. But I don't think its the answer to our problem.

 

Once again , to many tests have been taken with Furmark, and the assumption that 90watts should be at the wall.

The fact is a 90watt PSU might use 130 at the wall. Furmark makes these tests done irrelevant. It is considered to be a powervirus by AMD/ATI and Nvidia also recommends not using it.

 

SlyNine, this is why i havnt included anything with furmark or prime95 etc in the main page. Besides trying to deduce how much each component is capbable of pulling, everything else should be a 100% real world application.

Under verify you have a problem, it can be seen that multiplier throttling along can cause a near 50% decrease in performance, and this is at low resolution etc. Tf2 has a near 78% decrease due to its high dependency on the cpu. This is for my system with the wled and 720qm. Only throttlestop makes only Tf2 play possible at 1080p and smooth framerates, other wise i would have to settle for 900p etc. I have hit clock modulation in games like l4d2 causing a stuttering issue here and there.

There is also a list of games, watt numbers, fps comparisons above the faq section of the main page. from multiple users.

I didnt know if you were aware or not, but i thought id mention it for anyone else viewing as well.

 

If you use the laptop selection tool on Dell's website and tell it that you are interested in "Gaming - superior speed for hi-def mobile gaming," the only options that come up are the Alienware M17x and the Studio XPS 1645.

They're clearly advertising for gaming - that's why I bought one. I'm pretty disappointed to learn that gaming on the system is capped by this whole throttling issue. I don't know what I'm going to do about it... and I have a while to figure it out since I only ordered yesterday.

 

@Perdire: no offense, but you're repeating the same stuff I've read in these previous pages (both posts, not just this one). I don't want to keep re-reading the same arguments we were hearing two or four weeks ago.

It's great that your optimistic and all, but here's the facts: my laptop throttles. It makes me unhappy. Your optimism does not make my laptop stop throttling. You can start a new thread where everything is just gravy, and I'll jump right in when Dell fixes the issue.

@Slynine: I agree: TS is not the end-all solution, running the laptop as advertised is. For those who bought this laptop to actually *use* it, we have TS for now.

Since my last correspondence with Dell, I haven't heard back from anyone, despite several emails. I love the laptop, but I'm already giving serious consideration to returning the lappy if communication (and progress) remains stale.

(tl;dr - let's get back on topic. )

 

No I wasn't. Thanks for informing me. Any Idea what the usage would be if say using prime 95 with 2-4 threads while playing these games. This could be useful and realistic information to indicate what future games, especially RTS games might have in store for us 1645 users.

I'm thinking that the 90Watt adapter might be able to handle 110watts from the wall, but if adding 2-4 threads jacks that number up much.

My 90Watt adapter hasn't got too hot honestly, but I am concerned what might happen if the throttling were removed. Any tests done on a relatively cool (not burning threw your floor) 90 watt adapter and throttle stop?

 

The reason, IMO, that we use Furmark is that is a reliable and easy-to-use GPU stressing tool. Anybody else got one?

3DMark '06 = takes too darn long and you have to judge on that highly-variable score it gives you.

 

Maybe I'm naive, but I'll bet that a 130W adapter doesn't cost Dell $5.00 more than a 90W one. I would hope they had a better reason than that.

 

Yea but when the 2 video card makers warn against using it, one of them calling it a power virus. It definitely doesn't sound like something I'd use on a laptop.

I think the only way is to find the game that makes the GPU run the hottest and use that. But then considerations like, is the GPU not being used because the CPU is taking more time.

In any case, furmark is not the answers.

 

In my testing, actually GPU seems using full power even not in 100% load.
My returned laptop showed 30% GPU load while being CPU throttled.

 

Im alittle confused at how you tested that on an individual component level. GPUs will use more power or less depending on the situation and usage. Just measure the heat from the voltage regulators on the 5870s.

I know turning the power play feature to maximize battery power, even when plugged in the CPU doesn't throttle much (15-17 mulitplier), even when running a game like supreme commander which is vastly CPU limited.

But perhaps your Laptop was worse then the average laptop.

 

They will give you one for free if you complain enough. Or just file a Better Business Bureau complaint (US/Canada) and they will respond the next day.

 

While I agree that Furmark might not be the ideal program to use for testing, don't believe everything nVidia and ATI/AMD say completely either. I should know, I worked for ATI/AMD for a couple of years, not too long ago either.

Regardless though, I'd be afraid of little to no throttling on the 90W adapter. It gets far too hot even with massive throttling, upwards of 70C. I don't think I'd want to see 100+W sustained for a few hours on that thing, to be quite frank.

 

Mine gets no where near that. But what AMD/ATI is saying is quantifiable. GPUs can push massive amounts of threads at the same time, obviously designing a benchmark to push something that dynamic in every way at once is extreme.

I don't believe everything I hear. From you ( no offense intended) or AMD/ATI or anyone. But if it has supporting evidence then you have to consider it. occam's razor comes to mind.

 

When CCC set to maximize battery power, GPU will run at minimal speed. Any 3D FPS or action game will not be be playable.

 

This thread turned into the first throttling thread, speculation and dell bashing. (Which is cool lol) However the quality of information in the thread is on a steady decline. Maybe someone should dedicate a thread to discussion of throttling?

 

lol...we wish. The first post has/should have the most relevant information and that's the best we can hope for.

As soon as we find a viable replacement for Furmark, bam, we're on it. Games are fine, but in no way are they as reproducible as Furmark.

Games/relevant applications will and always will be the last word on this throttling issue, bar none. But Dell wanted proof, so we gave it to them.

In all seriousness now, there is NO point to any further testing including Furmark as our point is our there, Dell knows, and is working on it. Did we have to use an over-the-top program to get them to listen? Probably not, but it definitely made it easier.

~Ibrahim~

 

lol Fen it is true, but there isnt much to say really anymore. The important stuff i try to keep updated on the main page, and discussion is just whatever in this thread. Educating more people on this is most important, mainly presenting the facts and helping others understand them to come to a their own conclusion.

slynine. Discharge your battery, and then plug it in. and use it like normal for a while. It should get pretty hot. The hottest i ever seen though was gaming for a few hours in l4d2 i think it was. Which would pull a constant 90w ish from the wall.

Also the 90w if it could run at its rated capacity. It might run laptops like mine just fine for "real world", but others with rgbled screens and 820qm etc have a bit more overhead. My experience i would be very careful where you put it, it left a square imprint on my cheap couch. Perhaps theres some pretty wacky tolerances with these adapters and experiences vary. The throttle point is pretty deliberate at 90w +- an ish from the wall, obviously thats not what the laptop is getting but its most likely that way for a very good safety reason.

I would do the test with extra cores on prime95, but i just mailed my laptop back today. As it is getting returned, i cant deal with all this right now, as i need one to be able to handle my engineering simulations without issues(yes they use all 8 threads). If you would like to test and send back clear results i will add it to the main page.

 

That is a very blanketed statment, it is true that the games will not run as fast and you may need to find settings that they can run at. But the CPU will not throttle as much which can make for a more consistent gaming experience.

I've got UT3 running at a solid 40-60 FPS with that set.

I agree it is not the way I want to play it. But before the CPU was dropping and even on the lowest settings it was totally unplayable. until a fix by Dell is released it is an option in some games.

 

UT3/Source-based games are all CPU-dependent: they run great on basically any GPU. They're, however, far far far in the minority compared to other games. Just throwing that out there....

~Ibrahim~

 

Im using a RGB and a 720QM. It was charging the battery because I just ran tests on battery to check performance. The Battery was about half full. I played through supreme commander and the adapter didn't get that hot.

So I don't know if I got really lucky, some other variable, or you guys just got very unlucky with your adapter.

I think there is just too much speculation.

 

That game might not be intense enough on any of the system parts to increase power demand. This is the problem with using games, there are too many variables. There is a general consensus that the 90W adapter gets extremely hot when you work the 1645, so I think you just haven't seen it (or felt rather) yet.

 

UT3 is much much more demanding on the GPU then source based games. I've ran assassins creed, Grid, Dirt 2, Mirrors Edge. I'd like to know what games he's thinking of being so demanding.

 

Baha. x) So they do have it listed under gaming. I found in interesting under the 'Power - Ultimate speed for complex applications' they listed their Adamo Laptop as well as their Inspiron 15. Erhem. And I'm also sorry that some of my information seems 'repetitive'. I would note although I've read most of the threads and posts its a lot to keep in mind. ::wacko

I have to wonder why Fred (from the Youtube link video) didn't experience any (apparent) throttling issues playing Crysis? Could it be it wasn't because it wasn't very extended gameplay? (I noticed you commented there SlyNine.)

 

You can use Prime 95+ a game to give you a pretty good worst case scenario, if you try to artificially use every single component at max you are making a much bigger mistake. As you are begging for a failure. yes a laptop in this design there may be some give and take. If you want that kinda of performance you better go with a desktop, and even then you could run in to some problems.

It seems like you are asking the impossible. If you run a benchmark you can get repeatable results. Try the Crysis GPU benchmark, GTA4, or show results using different games.

Using Furmark is truly pointless. You are just showing you can reproduce pointless results over and over again.

 

GTA 4 is a bad one lol that game is terribly coded for the PC. Crysis is alright I guess.

 

Relatively speaking, though, UT3 is still more CPU-dependent than many other games.

I can't comment on any of those games....but (my inner skeptic), can you provide the settings and FPS for those games? Being able to run a game: can mean basically anything. I'm just saying that to illustrate my point, no need to actually do it.

Many people who game on this laptop expect to get the performance of a ATI 4670 Mobility GPU. The "maximize battery life" option is there and will work, but it's going to reduce the playability of most games. I think it runs at 200MHz/350MHz (core/memory) when using "maximize battery life"?

Sure, CPU-dependent games will run fine, but this is a very weak temporary solution.

It's a workaround, though, and for people who don't care about the performance/quality of games, it's definitely the easiest option.

Another thing you can do that works exactly the same way: turn down all the settings in the games. It works, but it's not that great of a solution.

I can't remember the Crysis GPU benchmark too well any more, but right: it stresses the GPU, but what about the CPU? This laptop doesn't throttle unless BOTH are being heavily stressed. I mean.....

~Ibrahim~

 

I have tried assassin's creed and dirt 2. They are playable on lower setting, but my 2 years old thinkpad can run them better.

 

I was able to play Assassins' creed on max settings with no throttling...before I knew what throttling was lol. Which is weird.

 

It runs, but random input freeze and audio problem in that game. It run very smoothly on battery.

 

I didn't experience any of that, but I played AC 90% of the time on an external controller.

 

And the plot thickens.

(Sorry, I know that sounds really cheesy.)

 

Then again...AC was fugly. So it was probably light on the system.

 

theres no plot thickening. Its been said over and over that it will vary with system configuration, and application. Even application settings, which is why i started the game throttlestop database on the main thread to record this info, now if only ya guys would send results this way i could add a lot more lol.

 

What kind of results are you looking for? And how would I go about recording them?

 

Look at what others have done. A lot of before and afters between using throttlestop and not. watt measurements if available, and system config. I add just about anything if it shows solid evidence of throttling, The layout is not set in stone.

so far this is what i have, ill repeat it here for clarity. Think of anything more to add, or adjust let me know. Keep in mind i dont get too thorough info on applications setting etc when ppl send me these.

edit: if not running throttlestop, log multipliers/modulation and i can post that information as well.

 

I don't have a kill-a-watt but could provide TS logs with it enabled and not enabled to compare + info like FPS and the settings used. Would that be helpful?

 

Lol. I realize what I said probably had no relevance, but I wanted to say it anyway. (Which is why I put in parentheses it sounded rather cheesy!) I don't actually own a Studio XPS 1645 yet, but it is either en route or still being configured. I don't actually own a lot of intensive games. Mainly just things like AoEIII, BfMII, RC3, etc. That's primarily because I've been using a Dell Inspiron 1521 and it couldn't do much else! Which is the reason why I wanted/am getting a Dell XPS 1645. I'll be able to do more along the lines of Photo Editing/Video Editing/Games/HD Video/etc. In fact, it was my complete infatuation with this system to dig up any dirt that I found this thread! I'm praying though results will show soon enough, and if I am thoroughly displeased with my system I can always return it. Maybe I'll mysteriously get lucky and it won't have any problems at all. (Although I of course doubt it.)

@ Fenikkusu, not that I'm really the one who should be replying to this, but I believe the answer would be ' yes'.

 

of course, the bottom line is to show how severe throttling affects applications. FPS is by far the best measure, as thats the main issue at hand.

Extra info like watts etc, is just that extra, and helps confirm all our previous statements even more.

 

But Dell already admitted all of that. They emphasized that it is multimedia laptop. Poor performance in game is expected.

 

We all know that's BS, it's on their site advertised as a gaming machine.

 

Wait, wait, wait. I'm getting confused.

As I understand Dell has been advertising it for gaming performance? (As pointed out via their configuration page under 'How fast does it need to be?' and 'Gaming'.) I think my initial confusion has been from reviewers (who've stated it isn't a reliable gaming machine). I mean, I do think it is wrong to call it more than a true multimedia performer, but I'm trying going by what things have been portrayed or perceived.

Heh. . .

(Although I did point out there entire configuration system under 'How fast does it need to be?' seems a bit wacky.)