S-XPS 1645 Throttling Info. and Updates

I played mass effect 2 for 2 hours yesterday on full settings at 1900x1080 with no visible throttling.

Anyhoo, my new 130W adaptor arrived today, so I'm considering this issue (and the iphone sync issue!) closed for me now, thanks Dell!

 

How hard can it be to set the throttle limit to 95c instead of 84c if (according to unclewebb) it's able to handle those temps..?

 

How did you explain your issue to get them to just send the ac adapter?

 

I told him my 1645 slows down on high stress and the 90W adaptor is getting extremely hot. And I told him about the official Dell solution, which he knew about. So it was no problem.

 

no throttling for me... you sure you've got a dell 1645?

 

The 1645 experiences CPU throttling if running Prime95 (8 threads) and FurMark simultaneously. Run either separately and you won't experience CPU throttling.

Some folks see this throttling as unacceptable given that the 1647 (and some i7-quad systems from other vendors) can run the combination at full bore. Others aren't as concerned. Dell seems to consider the case on 1645 throttling closed, however, so it's up to each user to decide what to do.

Personally, I'm content with the system as is. That said, if I was a gamer, I'm certain that I'd feel otherwise.

 

Vent location and throttling are two separate issues.
A07 has huge impact on throttling, i can not get this laptop to throttle with real world applications.
130W adapter was put in place to eliminate throttle issue (with A07) by allowing more power for the laptop, it had nothing to do with overheating 90w adapter.
Please try to make an intelligent comment backed up by data next.

 

Yep. Famx is once again confusing issues where the throttling we were seeing previously was related to the system being underpowered. What we have now is throttling of the GPU when it reaches temps of 82C. This of course wouldn't really be an issue though if the ventilation was better set-up, and better application of thermal pastes installed.

 

Kudos quietCat. I think "environmental tax/incentive scams" are going to be the economic scourge of our times.

 

Laptop with previous bios was not only throttling but also its 90w power adapter was overheating because it was not sufficient enough for XPS 1645. DELL major concern was safety issue and by swamping with 130w power adapter they certainly eliminated power adapter overheating. At the same by providing more adequate power to all laptop components they slightly eliminated throttling but throttling still exists especially with RGB panels -still not enough power. Now with that being said by providing more power laptop generates more heat and overheats much faster throttling kicks in even faster due poor vent design. On the top of that all that heat gets blown right into laptop screen which gets hot pretty fast and will cause for the screen to malfunction.

 

Here's a few things the Dell engineers could learn about laptop design.

Asus G73JH
http://usa.asus.com/product.aspx?P_ID=8je5Ot4HBnKOdT81

If you plan to use a Core i7 Quad and a separate ATI GPU then include two large air vents at the rear so heat can escape away from the laptop without being blocked or directed at the screen.

Use a 120 watt base adapter for your Dual Core CPUs and if a person upgrades to a Quad Core then use a 150 watt adapter like Asus is using.

A simple button to overclock the CPU is a nice touch and so is a 2-year limited global hardware warranty as standard.

Laptops that are designed to be barely adequate aren't good enough in today's market.

Published: January 30, 2010
Dell Inc. acknowledged Thursday that it has cut jobs this week at a number of locations, including central Texas.

 

Yes, that looks like a great design for cooling.. but for my needs, 2lbs heavier, it's too big (I already have a 8.5 pounder).. my Studio 16 is already pushing the max weight of 6lbs I wanted.

Dell will most likely have to add weight in the XPS 16 redesign that we all know will come (or have to).

what's up with the G73JH's 4GB limit for the i5's and i7-620m though

Great performance gains can be had at 6/8GB

 

Hmm, its a monster. I looked at the specs and found something curious, the rating on the adapters are as follows:
120W : Output :16V DC 4.7A = 75.2 W = 62.5% efficient
150W : Output :19V DC 4.7A = 90 W = 60% efficient

Something smells fishy, maybe asus does not know its own specs or it will throttle like hell? and its ugly as hell.

 

I found this the most interesting. Been trying to understand what it is and it seems to only really helps the 920 while the 720/820 do get a boost from 133 to 142 their intel turbo multi doesn't appear to kick on anymore...

The guy who seems to be recording most of the benches reported 168W system draw while running twin-turbo and intel turbo kicked on (920 CPU). Seems their Power-supplies allow more than advertised.

edit: clarity.

 

Hi Guys,

sorry to be a little bit off topic. Todd from Dell is asking questions regarding the vent issue in the other thread. Therefore, if you feel you have anything to share, please comment there.

http://forum.notebookreview.com/showthread.php?t=460540&page=9








Hi everyone,

I just got done talking with Bill about this issue and wanted to provide an update. We're talking with our engineers and trying to get a resolution.

I do have a question (and forgive me since I havne't read the entire thread yet) but the people seeing discoloration on the systems, how long have you had them?

I ask for a few reasons. 1. to get an idea how long it takes for the problem to show up and 2. to see if any of the changes they did to the manufacturing line a while ago helped with this issue.

As I get more information I'll try to get it posted here for you.

Todd

 

Hi,
I want to know if dell has begin shipping 130w adapter with the system
itself.If not when will it begin shipping it....

 

hello ppl...
sorry to bother but i still feel my system throttles
while playing assassins creed even with the A07 BIOS and 130W (BRICK) adapter.. i hardly get these no's:
RESOLUTION: 1920x1080
DETAILS/GRAPHICS: 3/3(FULL)
FPS(i get during playing) : 30-35 (Hardly!! max-38-40)

whilst my batchmates STUDIO 1558 (With ATI HD 4330 512MB card and core i3 330M processor posts almost the same performance:
RESOLUTION: 1920x1080
DETAILS/GRAPHICS: 3/3(FULL)
FPS(he gets during playing) : 28-32 (max-35-37)

and the game also stutters while fighting or running from soldiers in the game.. i correlated this problem with overheating of GPU (by GPU-Z sensors and logs)

so why did i have to pay so much for this system when my batchmates system performs almost equally good for just half the bucks????

and can anybody else playing assassins creed post their benchmarks or fps??? please???


i just have the complaint regarding this game... and every other games performance has increased very much with A07& 130W... just this game which is alas my favourite....

 

First I have to say it sounds VERY unlikely that a 4330 would perform anywhere near that. The 4330 has 80 shaders running at 450mhz. It also only has a 64bit memory bus running at 600mhz and probably just DDR2 memory at that. Even a throttling 4670 would vastly out perform that. In fact, the 4670 running at 250mhz is still 2x as fast. The 4330 cannot even play fear at high settings. It can't play Crysis in med settings.

Your 4670 should be running at 675mhz and 800mhz memory (128bit GDDR3). Check GPU-Z while playing AC.

Second 1980x1080 is pretty dependent on the video memory bandwidth so 30FPS sounds about right. What you need to do is objectively test it with GPU-Z and throttle stop to see if it's throttling. Given the nature of the CoreI7 turbo boost anything above 13x multiplier isn't throttling.

Also make sure powerplay is turned on, I noticed with mine turned off it would drop the core speeds down to 400 mhz and memory down to 600 if I remember correctly.

 

It sounds like it would be a good idea to move over to the other thread for now to help push the issues that the heat needs to be taken off the GPU so it doesn't throttle as much. . .

 

+1
Definitely. Let Todd hear about GPU clocking/temps on that thread.

 

http://forum.notebookreview.com/showthread.php?t=460540&page=11

There you go fellows. I have now posted what I think explains our needs/proposed solutions in regards to the venting/cooling of the machine for Todd to pass onto the engineers. If I have of course missed anything that is why I have provided the link above for you to review and correct/add anything I may have missed or spoke unto incorrectly.

 

I think I will be able to answer this question with specific factory cut in date Friday.

 

Hi,

I just received my 1645 from Dell, and it arrived with the 90 watt power supply.

Can you please share with me what I ought to do to receive the 130 watt slim supply, and what I can do to test the system?

So far, at first impressions, this machine is beautiful. I would love to learn where it's edges of performance are.

Tonight I am going to delete the OS and reformat a fresh Win 7 Ultimate, and make sure the latest drivers are in here.

Many thanks everyone,
Mike

 

I guess the vent is not affecting my screen as there is no discoloration at all and I checked it with different color backgrounds as well. I also have had my laptop for 4 months now and I have used it every day for about 12 hours a day, it would be nice if the vent was able to blow out the hot air and not hit the screen, as that would help in venting the air properly and help with keeping the Laptop cooler.

 

"Greengate"=hogwash. Try proving the inverse, and I'll bet you will be able to shoot holes in that theory all day long.

 

AtlasApallo in order to get your 130W adapter you need to call DELL and request 130W adapter to replace your 90W adapter. (Unfortunately you can't keep it.) If you have any problems though, either ask for the manager, or try calling again to get a different rep. (This is because not all the reps read the notices informing them to hand out 130W adapters.)

 

Thank you! I called them, and after some run around, found my way to tech support who were very helpful. The 130W slim adapter is on it's way!

When it arrives, can you please share with me what I need to do to test it all?

Many thanks!
Mike

 

The components in the adapter may be rated just fine, its the heat that tends to open things up and kills performance. What i remember was the form factor of the 90w was extremely small, not much surface area for it to give off heat. If i remember right it took a while to heat up, probably meaning it wasnt wasting that much energy, just couldnt get rid of it fast enough so it got hot. Heat rejection is huge in power management. There is a plus to bigger powersupplies with more surface area. Just like ice on water pipes, actually makes the water warmer.

no surprise the dell 130w and hp 120w are much bigger.

 

The thing is, before anything could be proven someone would need a approved measuring tool that matches the energy label requirement.

Approved meters will include the following attributes (Characteristics of approved meters taken from IEC 62301 Ed 1.0: Measurement of Standby Power):
• Power resolution of 1 mW or better;
• An available current crest factor of 3 or more at its rated;
• Lower bound on the current range of 10mA or less.

Anyway, all that's beyond me. All I want is 125W (or higher if possible) out of the 130W adapter straight into my system.

 

hi,
Sorry for being off topic. I posted this on owers thread, but no one seem to have taken notice to this. I hope someone in this tread will

1> Does 4670 support openCL, n if yes why doesn't gpuz show this?

2> Windows show a total graphic memory of 2810MB with 1786MB being shared. Could some one tell me how to not share any memory with the gpu?
i don't see any option to alter the shared ram value in the BIOS


Thanks !

 

As I understand this, shared memory in this case is dynamic; the shared memory is actually only made available to the graphics card if it's currently not needed by the host operating system. Whenever the system memory is actually needed by an application (or Windows itself), the shared graphic memory is restored to the system.

TL;DR version: You won't notice that it's missing, because it isn't.

 

Sent this post to the design engineers working on the next series.

 

Factory cut in for 130W adapters on new 1645's is 3/8. Anything that ships after that date will have the 130W.

 

Quietcat the latitudes were due to thermal throttling.

Theres 2 things wrong with your spiel. For this to be a reason, for starters ya have to show that it is impossible to get energy start compliance with the 130w. Theres a good bit a doubt that ya couldnt get compliance, as other manufacturers etc can do this. That and removing the rgbled option would probably take out any inefficiencies in the adapters if needed be.

secondly since the 130w is now available, and the energy star ppl dont read max power. There is still no energy star reason to limit the computer to around 110w. Its obvious theyre doing this to keep heat down, as the components drawing more power are now sources of more heat, with increased internal temps, it most likely at risk of failure over the long term if they dont limit. With the chassis that doesnt vent and only cools cpu+gpu, the increase in power is apparently detrimental to those components

While energy star compliance could provide motivation, it certainly isnt the main issue. As im sure its not as simple as that. I dont know if it was a copy and paste job gone wrong, or if the engineers knew the heat limitations of the platform and decided to ignore it. Either way you cut it, its bad design, bad engineering

 

Looks like i just missed the cut in. My 1645 is on the truck to be delivered today. Based on the service tag it came shipped with the 90W adapter. Bill, i sent you a couple PM's but i don't know if thats the proper channel anymore to get a 130W slim. If not could you point me in the right direction. Thanks a lot for all your help.

 

What about the current machine Bill?? Any news for the vent issue and GPU Throttling?

 

All you need to do is contact tech support by phone, email or chat. I recommend email so you don't have to wait on hold and you get everything in writing.

Todd should be posting an update on the "vent thread" shortly.

 

Oh great. Thanx alot Bill

 

E-Star facts:
1. AC adapters and PC's are tested independtly for compliance.
2. Both the 90W and 130W are ES-compliant. To be compliant, the adapter has to show 87% efficiency. It is independent of PC compliance testing, and which adapter is used to test the PC or is shipped with the PC has zero to do with the rating.
3. PC's are tested for compliance in 3 states: Off, Sleep, and Idle
4. The gov't has a formula based on % consumption in all 3 PC states. If you hit the spec in all 3, you are ES-compliant. To be clear it is (% of time in idle state) X (% of power consumption in idle state), and the product of those 2 variables has to be under the spec, whatever that is. Same formula applies to Off and Sleep.
5. Let's say the spec for the PC is 15W (arbitrary number). As long as the PC does not draw more than 15W at idle, sleep, or off, it is compliant. So, the rated power of either 90 or 130W adapter (which are, again, tested seperately for compliance) has zero to do with PC compliance.
6. Throttling also has zero to do with compliance, since it obviously happens at way over the power draw at idle, sleep, or off. Duh.

*Tin foil hat back on. Back to the drawing board.

 

I'm sure he'll think of something.

 

Of course. He's a man on a mission.

 

Bill, get any feedback regarding the USB power/disconnect problem?

 

Since you dare to say that ES wasn't Dell's reason to throttle, maybe you could enlighten us as to what was the real reason?!

 

Good pickup, was wondering why it wasn't listed.

Until Dell correct this, my guess is to separate the Alienware and XPS series. When 1645 came out the 5xxx card wasn't available (and well it still isn't when the 1647 could have included it).

The only big difference between the 2 systems is the GPU and 2 fans on the AW. Other than that, XPS series is much cheaper for the same exact hardware, therefore a better system overall (when RGB was available), so unless you're only basing your purchase on the 920XM CPU and the best GPU Dell provides, the XPS wins. That there is a pretty small group of people not making it very viable, so they had to do more.

If Dell remove throttling from future XPS systems, better cooling and includes 5xxx card then it starts looking questionable what they can do between these 2 systems. Just a personal opinion but why would anyone buy AW then.

 

I'm afraid I disagree with you hands down. The problem is they don't need to throttle the system to distinguish it from the Alienware and Studio lines. Each Dell model is made to fit in a certain category of 'use'. With some systems its cost, portability, power, gaming, work, expression (of wealth like the Adamo), etc. The list goes on and on. For Dell all they are trying to do is offer a portable to best meet that 'use'. The SXPS is there to meet a certain type of 'use'.

Anyway. I do agree that the gap that makes the Alienware, Studio XPS, and Studio lines is getting less distinguishable. Really, yeah, the biggest difference is the GPU they offer + build + cost for the SXPS in comparison to Studio or Alienware models.

 

I "dare" to say. LOL. Like I'm Evel Knievel because I see throttling happens at anything but idle, off and sleep, then "jump" (pun intended) to the conclusion that we did not throttle for ES compliance.

I could not enlighten you because I am not a design engineer, and I was not in the design meetings.

 

Bill,
I'm still waiting for your contact here.
Hope you didn't forget me (brazilian xps 1645 owner who'd reaally appreciate to receive an 130W adapter!).
PM you back in 02/18!
Thanks

 

I. . . ? Just decided not get into pointless speculation? Erm. . . wait. . .

Anyway, all this speculation as to why SXPS 16 throttled to begin with is pointless why? Because we don't have all the facts. We are not privy to all the reasonings, mechanics, or other mechanisms to know why there was throttling instigated in the BIOS and then an underpowered supply. For all we know it could have happened on accident. (The inverse argument that it all points to 'intent' is about as good as reasoning the universe was created by "intelligent design".)

The other side of this explanation is also: What does it gain us in knowing? While it could certainly shed some light as to why these methods may have been accidentally or intentionally implemented it does not change the fact that it was obviously inadequate or not worth the cost. What then should be focused on is the here and now. Which is getting what we/you (at least thought) were entitled to out of your purchase.

I made my case. AGAIN.

(I like going 'Anyway'. . . )

 

Sorry, but I don't have a PM from you. Can you try again, please? Apologies.

 

Bill,

I just got my 1645 today.

Unfortunately, it came loaded with A06 and the 90w adapter. Are you still sending out the 130W adapters?