*OFFICIAL* Alienware 17 R2/R3 Owner's Lounge

Wait. Wait. LOL WAIT. He's got TDP throttle lighting up... at like 35W? The BIOS locks TDP for the CPU to ~36W when CPU and GPU are stressed? What the hell is the point of the 240W PSU?
That doesn't even draw 180W of power!

 

My CPU showed 57w a little while ago. I'm not even sure if this is good or not, I'm so freaking lost with this TDP crap that Dell and Intel are pulling.

I'm using throttlestop right now. Temps aren't great even with a cooling pad, room doesn't have AC yet so that may contribute.

I need Dell to get off their butts and put out a honest, no BS, working fan profile, BIOS. We need to bombard them daily about these issues. This is all a BIOS update away. What's taking them so long.

 

One can not have a high end laptop with a processor that throttles and not throttles sometimes after that this suit themselves whether it is in gaming or heavy processor usage. This is totally unacceptable regardless of brand... But Dell has had 6 months now to fix the problem but nothing happens. Aw Laptop is the brand now who have the most problems with just this trottle problem... (whatever is not a processor that manage max 47W much to have).

 

Can you run this same test and when your CPU is throttled back to 35 Watts and the 28 multiplier, can you run my MSR Finder program and take a SnapShot of the CPU registers?

MSR Finder
https://www.sendspace.com/file/luzete

I am trying to find out if I can come up with a solution for this new type of throttling. There is a register in the CPU that I recently found out about but I need to see if Alienware is using this new register for throttling purposes or perhaps they have another trick up their sleeve. A 4980HQ should definitely not be throttled to 35 Watts.

You can copy and paste the data that MSR Finder produces to www.pastebin.com or send it to me directly in a PM message. This program should create a new folder called Logs and in that folder should be the info that I am looking for. Thanks for your help.

@ blackbalt89 - Can you upgrade to ThrottleStop 8.00. It has a few more features for the newer CPUs.

 

I could try 8.0 but will it really benefit me? I don't want to change to 8.0 if no one gets over 47w on it but somehow I managed 57w with 6.0.

I remember before TS I would only see 47w. Something was different tonight. Not sure what it was.

 

The processor uses 47w as soon as I start the test but drops to 35 watts after 30 seconds.

If I use the CPU alone it will run at 47w indefinitely . but as soon as the GPU is used in conjunction that's when it drops.

 

Had to paste it here because I couldn't PM you for some reason. Hope it helps!

 

Paste your test in the Trottlestop guide in this forum...

Your drop in processor power to 35w after 30 sec is made by Dell. This is similar an very cheap Acer laptop. Dell engineers can't have tested the new modern "thin" Aw laptops properly during the development... Dell to blame. It's also Dell's engineers who want an 180w weak psu in those laptops. They should not have this job. Maybe they have a engineering job at Acer/Packard Bell earlier. Who knows. Give them back to Acer...

 

Unfortunately, engineers don't have that much decision power these days....
If some genius in the marketing or whatever decision units decide that the laptop must be thin to make some pinny-pinching economies on the DVD-blueray reader and on the PSU power what can you do....

The questions are can this kind of throttle be reverted?
and if yes, can the laptop handle full power OC CPU/GPU without burning?

If the answers are no, then they missed to advertise automatic throttling/downclocking as a extra-feature of the laptop in the webstore page. Because what they write under "full throttle gaming" in their website seems to be misleading advertisement according to what we got in real world.....

I just hope they can fix that because if it is OK right now with current games in the future it might be another story with more demanding games...

 

I never keep my laptops long anyway. So I will probably have moved onto something else this time next year.

If we can get to the bottom of the throttle issue on this machine then who knows? I might actually keep it. We will see.....

 

If you have temperature problems running idle, you can change your power profile to balanced if not done already. This setting lets the processor downclock itself on idle situations.
It worked for me by changing it to balanced.

 

Nobody should need to put power profile to balanced on a so called high end performance gaming laptop due to throttling or heat issues... If you must do this then the product is not good enough. Then it is a crippled product. Pay $ 2-3300 for a gaming laptop that can not run hardware manufacturer's own maximum specifications with ok temp is a useless product. This should all understand... Not acceptable.

 

On High Performance you could have it set to clock down when utilization is low (set Processor Power Management --> Minimum Processor State to 5%). I don't think it hurts performance any because it'll clock up very quickly when it's loaded.

UncleWebb/Throttlestop! You always post in a thread at the perfect time! This must be a fun and rewarding career or hobby for you.

Throttling to 35W, CPU temperature is below 90F so why do they force it to throttle? Can't be to protect the CPU/GPU unless it's to protect voltage regulators or other power delivery hardware. But, it doesn't actually throttle that much when at 35W because the iGPU is idle (dGPU active) and holding 2.8GHZ on all cores isn't far from the max 4 core of what, 3.1GHZ? (EDIT: 3.8GHZ! Wow that would be fun). But still no excuse, along with resume from sleep CPU TDP limit.
EDIT: It's interesting the GPU runs cooler than the CPU even though it takes more than twice the power of the CPU.

Same result with A00 BIOS? Thanks Mickbt26 for helping to find out this laptop's behaviors.

180W causes a lot of CPU throttling (down to 800MHZ) from what I saw when GPU is running Furmark, CPU Prime95, and GPU throttles to ~850MHZ via Furmark sensor. 240W solves the CPU throttling so now it's at 2.8GHZ all the time. 180W may cause GPU throttling?

So when using the 240W AC adapter, it's just Furmark in which the GPU doesn't stay at peak core?
If so then 240W seems to take care of the power problems indeed if we can figure out the 35W CPU TDP issue.

EDIT 6/27; Looks like Mickbt26 is the only one with the 2.8GHZ throttling when GPU is used, issue, never mind. Thought it was another widespread issue.

 

BIOS A00 is exactly the same. Ive tried both. The only thing A00 does for my machine is keep the IDLE temperature down a bit more. But load temps are the same. The 35w Throttling still occurs on both Bioses in the same test.

Funny thing is my GPU never ever throttles with the 240w adapter. It always stays at 1126mhz, be it gaming or testing.
So at least my games framerates remain constant even if the CPU is throttling.

 

I can confirm that the i7-4710 is also TDP limited to 36W in my AW 15.

When GPU is on , the CPU never goes above 36W..

On another note, I've just been sent out a 240W adapter to see if it solves my thermal throttling issue and I have the 970M!!

 

On A00 240watt, while playing sc2 (max graphics) for 10 mins, framerate drops from 100+ to 4. happens everytime.

 

Wow, That's interesting. So even with a 970 the 4710 cannot even keep its TDP at full load. This just get better and better.

Can any one else with the 2015 15 or 17R2 do a similar test to mine? (run Furmark in windowed mode whilst simultaneously running 8 threads in Throttlestop and keeping an eye on the CPU clocks and TDP) and check for this 35w TDP throttling?

Might have to run it for a while as it can happen after 30 seconds or after a few minutes.

 

Got the same problem with my 17r2, also while playing wither 3 the cpu throttles after some time down to 2.8ghz. At the moment I try to give it back to dell but it's not that easy because I own it about 2 months. Also the technicians told me it is normal if the cpu drops down to 2.8ghz and they won't accepted the throttling under 2.8ghz I had while running prime95 and heaven benchmark because I shouldn't run 2 benchmarks at the same time.... I spoke to 3 technicians and all told the same, now I requested another instance but I'm waiting for response since 2 days. Goin to write them again later this day if I haven't heared from them. It's really frustrating at the moment and I just want my money back.

 

People, don't run Furmark, NVidia automatically & artificially throttle their GPUs when the Furmark.exe is detected - it's not a valid test because it's a power virus of a program. Instead run something like Unigine Heaven Benchmark (because it runs on a loop forever until you quit it, and is a good representative of gaming GPU load), and run Prime 95 just on 1 or 2 threads perhaps to increase CPU load a little while running Heaven benchmark (Heaven benchmark doesn't use much CPU) - then you monitor your CPU & GPU clocks from there. (Because most CPU's will throttle anyway when running Prime95 on 8 threads - again not representative of a high CPU gaming load).

 

My GPU does not throttle in Furmark at all. I get constant GPU clocks and 99% utilization, (I monitor it with MSI afterburner) Whereas in Heaven, even on extreme I'm only getting 80-95% utilization in windowed mode. So furmark is better in this case. (I'm using an old version of Furmark, if it matters)

 

Well that's very surprising, I look at the reviews on notebookcheck.net, and there hasn't ever been a single high end card (980M included) which didn't throttle during their Furmark tests, so I'm very surprised by your example indeed. And if it's not throttling at all running Furmark I'd be very concerned for the health of your GPU - it will stress it like no other, even if the core temperatures are ok you don't know what the temperatures are of all the power delivery chips on the board for instance. If you're only getting 80-95% utilisation in Heaven benchmark, then your system probably has other issues (should be a steady 97-99% ish GPU load) - like maybe your CPU is throttling & not able to supply enough frames to your GPU, or maybe the NVidia driver is not installed properly or some other weird problem with your system. (Also try running Heaven in Full Screen rather than windowed - maybe that will make a difference to GPU load. Also make sure Heaven is running at maxed out at Ultra & Extreme settings, although you might not need to max out the AA to get 100% GPU load.).

 

Robbo99999, a machine that is well engineered ought to handle Prime95 on all cores 24/7. It is stressful but if the laptop/desktop/x86 tablet is built well it's fine.
Running CUDA on a GPU 24/7, same thing. Furmark does increase power consumption more.
Running a video game looper with a bit of CPU load isn't really putting the computer through it's paces. It is when everything is taxed that you don't want it to wimp out.
That being said, the 35w TDP on the CPU whenever the GPU is being used (even being used just a little?) might not be a big deal because when I was running Prime95 on my sister's 17 R2 4980HQ/980m with the dGPU not being used, it held not anymore than 3.1GHZ on all cores. Didn't look in to see if it was TDP throttling or temp throttling or nothing at all. So 2.8GHZ vs 3.1GHZ full utilization on 4 cores, eh not a big deal.
Could it be the difference between 35w and 47w TDP is 2.8 vs 3.1GHZ?
Mickbt26 post#1356 "If I use the CPU alone it will run at 47w indefinitely . but as soon as the GPU is used in conjunction that's when it drops." - Even if just used a little like 5% GPU?
EDIT: Could be another Dell ideastorm!

Does anyone see higher 4 core speeds?
So with the 240W adapter Mickbt26 and others have experimented with, now there is no GPU throttling and pretty much no CPU throttling besides the 35W TDP issue and resume from sleep 25w(or so) tdp issue? And with 180w PSU there was some GPU throttling or no? I know there was CPU throttling with it cycling down to 800MHZ (would be fun to see what TDP that was) when the GPU was fully taxed and then back up to the 2.#GHZ. I forget if you guys had GPU throttling with the stock 180w PSU. Maybe it's because some did some didn't!
A02/A04's battery dipper algorithm doesn't pull enough power to really change the throttling I guess. Saw without a load the battery, 180w PSU [the battery] was at about 17V, with a full load of furmark+Prime95, went down to 16.7V so not much draw but throttling of everything.

So it seems like the machine is fairly unthrottled then if my assumptions are correct. A bit of CPU throttling and no GPU throttling (except Furmark maybe) with 240W PSU regardless of BIOS.

 

I don't think I've seen any laptops with an i7 Mobile CPU able to run Prime95 without throttling, at least not without tweaks to the cooling system and other tweaks - even if temperatures are ok then full 4 core turbo (my definition of not throttling) can't be maintained due to exceeding the hard coded TDP limits of the CPUs. (Some of the previous gen slower CPUs (and XM CPUs) could be an exception to this rule, like my 2630QM because it only runs at 2.6GHz, and I was able to apply an XTU/Throttlestop tweak to allow for 56W constant full turbo, and Prime95 pulls about 53W on it).

Running Heaven benchmark on a loop with 2 Threads of Prime95 running is a very good test of gaming performance where there is both high GPU and CPU load. If I run 2 threads of Prime95 and run Heaven benchmark I see a global wattage consumption of 180W, which is more than any game I have played by about 15W, and the CPU load under this condition sits at about 85% while GPU load is at 100%. You're probably not gonna get any games that push the CPU that hard, not that I've played anyway. So, I think this test that I outlined is pretty much a worst case scenario gaming load. That's why I think it's worth running that scheme to monitor your CPU/GPU behaviour, because it's more applicable than running Prime95 on 8 Threads while running Furmark which was suggested by someone a few posts back.

The 35W throttling of the CPU when the GPU is under load - that's not a good sign, if you run my "Heaven & 2 Thread Prime95" test I outlined above I'd bet your CPU would be wanting to use way more than 35W, so I think that would be a good test; I would guesstimate that it would be wanting to use in the region of 45W at least in that test scenario. (If it doesn't load up the CPU enough, just up the Prime95 Thread count from 2 to 3 for instance - you can use Throttlestop or other monitoring programs to see how many Watts your CPU is pulling). I also think that CPU throttling at 35W is a big deal, that's gotta be hampering performance in CPU demanding games at high framerates.

 

This is mine running wprime95 and heaven at same time, Not sure how to see if its throttling, Anyone can give some insight on this

 

Btw im using 240 PSU + A00 Bios

 

I just ran the Prime95 (2 threads) and Heaven benchmark. 240W PSU and A00 bios. hwinfo64 fans max, and TS 8.0.

My CPU stayed pretty much at 46-47W the whole time and didn't throttle much. But what scares me is my CPU got up to 98°C during the test. Ouch!

 

my max CPU on 88 but averaging in 80
wprime on 8thread, no throttlestop
i was getting 47W does that mean its ok ?

 

Forget XTU for testing. use throttlestop in monitor mode instead. Click TS bench and run 8 threads. then at the same time run furmark or heaven (extreme/windowed)
Keep looking at the core clocks in throttlestop. Sooner or later your CPU will drop its max frequency to 2.8 or 2.5ghz depending on your CPU's non turbo limits

The picture below is my test using only 4 threads in throttlestop and heaven extreme bench as a typical gaming scenario. Still throttling even with throttlestop multiplier set to 40x

Only pulling 176w at the wall during this test . much less than with it did with furmark.

 

Yes, it is a bit hard to tell what's happening to the CPU from that information. You could maybe try running Heaven in Windowed mode so you can view what the actual CPU frequency is at the same time. You could also use something like HWInfo (a better option) to create graphs of the CPU frequency, although I don't use that function much so can't describe how to do that in detail - have a play with it & you'll probably work out how to graph the CPU frequency. You just need to know what your 4 core max turbo frequency is for your CPU and compare your graphs against what the rated 4 core max turbo is supposed to be (3.4GHz for 4720HQ).

That's pretty good that your CPU stayed at 46-47W, looks like it's not throttling much then, maybe monitor the frequencies so you can see what it's doing. 98degC is hot. It's either really hot in your room, or your CPU might need repasting. I think the R2 should be able to successfully cool a CPU running at 47W.

Enough with the mad advice for testing Prime95 (or TS bench on 8 threads) while at the same time running Furmark. Furmark is not relevant for today's GPUs, it's just too much of a load that you'll never see outside of running Furmark. If you want to learn nothing about the performance abilities of your laptop while at the same time stressing your GPU for no reason then please continue to use Furmark.
EDIT: oh, you edited your post since then, yes the reduced Thread Throttlestop Bench & Heaven Test you did is more sensible, that's good, although yours is throttling the CPU. Maybe you need to try some of the settings sirbevo is using (he said he's using A00 BIOS and 240W adapter).

 

Looking at yours @thunderbirds It doesn't throttle at all. So i am thinking I may have a lemon after all.

Any one elses throttle like mine?

 

Do you have A00 BIOS and a 240W adapter? sirbevo was using those, and it looks like he was able to pull a consistent 47W on the CPU.

 

Im using A04 at present. I have used A00 in the past and it did the same then too. I will try it again and yes I am using a 240w PSU

 

Cool, that 240W adapter should sort out some of the problems. As far as I know a lot of the later BIOS versions were to enable a hybrid type power, where battery power was used to supplement the weak 180W adapter power to prevent throttling - however if you've got the 240W adapter it could be that you'd get better performance with the BIOS version that doesn't include the 'hybrid power' - A00 for example. (I'm not sure what each & every BIOS revision enabled, so you might want to check you're not losing something by reverting to A00).

 

Back on A00. It still throttles to 2.8ghz after 11% of throttlestops 4 thread test.
In the limit reasons box [POWER] lights up in the core and uncore when it throttles

 

during throttle does it go back up to 47W or stay the same?

 

OK, so doesn't look like it's related to the BIOS version then. What happens if you Turn Off Throttlestop (just having it in monitoring mode)? What happens if you select a different multiplier for Throttlestop - on my CPU and the version of Throttlestop I'm using I run it on the highest Multiplier, which is called 'turbo', maybe try the 'turbo' mode if that version shows it.

EDIT: do you have BIOS settings that allow you to determine the length of the Turbo boost? On my Alienware I have those settings. Have you changed anything in the BIOS settings related to that?

 

It will only return to 47w after I stop the GPU test. otherwise it remains at 2.8ghz.

I see the throttle in games too when I use MSI afterburners OSD tools to show the clocks, temps and usage

I know the drop to 2.8ghz from 4ghz isn't exactly drastic and in GPU dependant games the frame loss is virtually zero anyway. but I would still like to get to the bottom of this weird throttling issue.

 

I have ran games on start up and saw the CPU clocks drop to 2.8ghz without any tools ever being opened (throttlestop etc)
So I know throttlestop has nothing to do with it. (Most of the time I use it in monitoring mode anyway.)

As for BIOS the only thing I see is for turning off its Turbo which limits the CPU to a constant 800mhz.

 

How about ticking the 2 boxes for modulation (above the multiplier box) in Throttlestop? Oh, and the Windows Power Plan settings: have you set it to High Performance, and then go into the Advanced Options and select for Minimum Processor State 100%, and for Maximum Processor State 100%?

 

Put high-performance in Windows power plan and not balanced performance. Easier to see if the processor throttles when you monitor with Hwinfo64 and Rtss during tests. Intel processors can use turbo boost short power max to 1.2 X max TDP of processor. Your processor does not go higher than 47w. Dell may have made a maximum limit of 47W for your processor in bios. Who knows.

 

Will anybody's rig post 57w when playing games, not benchmarks?

With my TS settings its been routinely going above 47w lately. I don't know why its doing so if these BIOSes are so locked into 47w.

 

1.2 x 47W = 56W
But, it will only stay at 56W for a short time. There may be a way to get round that and have it above 47W the whole time, there is with some previous Alienwares by using a combination of XTU and Throttlestop.

EDIT: you need an XM CPU to go past 56W. (Again though, I have seen some Throttlestop hacks on MQ CPUs that can surpass the 56W limit - think it's MQs, could be HQ's too, not sure - it's in the Throttlestop thread, and might not have been Alienwares)

 

Anbody know if it will be expensive to replace the motherboard &/or CPU in case of hardware failure when the warranty has run out? I heard the CPU is soldered onto the motherboard.

I feel like I made the wrong decision in buying this machine & stepped into a trap like a rat looking for cheese. That's not how someone should feel when they bought an expensive, new laptop...

 

How much the system board parts cost? Just wondering

 

Have you looked to see if you can buy replacement parts from ebay or dell maybe, then you could see the replacement cost. I doubt the CPU will ever fail in the lifetime that you have the laptop. The motherboard is more likely to fail I think, and I know sometimes motherboards fail due the power connector part coming loose where your adapter plugs in, but I think I read that with your R2 they've made it modular so that if that part fails that you can just buy the replacement part that clips into your motherboard (the bit that accepts the adapter from your power supply). Just go ahead and be happy using your laptop if you're happy with the performance from the CPU/GPU, I think it's unlikely to fail.

 

A nice guide that everyone should read...

If the motherboard, processor or GPU fall apart; This will be very expensive because all the hardware is soldered on the motherboard. Earlier with socket hardware could you buy only the motherboard, processor or GPU if one of these were destroyed. It is also easier and cheaper to find parts when you needed it because there was a larger variety of parts. Cheaper to replace a processor or gpu than a motherboard with all hardware soldered.

 

I don't like to stress my Alienware too much, it is just nonsense to use your computer on high performance, because of the core clock is being maxed out all the time. If you put it on balanced, it stays idle at 800mhz instead of a constant ~3.3ghz.
But, I can see your point. I think Dell needs to change its Bios for the Alienware 17 and 15, using better fan profiles.

 

I have always used high performance plan. Run my i7-4930mx at 4.3ghz on all cores with idle temp of 37-39 degrees, light use 40-45grader and heavy use up to a maximum of 75 degrees. Can not say that this stress the processor . Using Trottlestop and Unpark core because I want all the cores are in use all the time at maximum. Adaptive function is rubbish for both the graphics card and processors in a high end system. This is more suitable for a tablet or mini computer used for web browsing and who has a too Inadequate cooling...

 

So I just put my Alienware 18 on ebay and purchased a 17 R2 from microcenter. I really got tired of carrying the thing around and I didn't game on it enough to justify having it. Let someone else put it to good use.

All crap aside from how everything is soldered into this thing it's form factor is dead sexy!

Now.. should I read the forum forwards.. or backwards.. I can never decide..