*** Official Clevo P65xSA/SE/SG / Sager NP8650/51/52 Owner´s Lounge ***

hm, wondering if the psu itself or maybe the adapter cable is the culprit here... have u checked if the cable gets hot during load lonesyndal? might be that it just isnt rated for such high amps.

 

Here are the results. Settings kept the same with only the PSU swapped. As I expected pretty much identical. Remember I am stock with no overclock.

Stock 180W PSU:

Ice Storm Extreme - 105708
Cloud Gate - 21754
Sky Diver - 21556
Fire Strike - 8315

Dell 240W PSU and Adapter:

Ice Storm Extreme - 105401 / 105774 (ran twice as first was a bit lower)
Cloud Gate - 21788
Sky Diver - 21576
Fire Strike - 8298

Let me know if you need any other tests done. I am not wanting to flash new firmware so willing to test anything but tests relying on unlocked bios/vbios.

LoneSyndal: Didn't fry my adapter. I have one from Amazon that is right angled instead of the BIC one that is straight. Sounds like you might have a crappy crimp on the barrel connector on yours. Adapter I use: http://www.amazon.com/gp/product/B00G700QQ0/ref=oh_aui_detailpage_o09_s00?ie=UTF8&psc=1

Hope that helps ya fine folks!

 

That scores are impressing! Seems that nothing changes between the 2 PSUs, seeing those numbers.

Just to be sure, can you run 3D Mark 2011? The fact that it's got three distinct stress tests for cpu, gpu and cpu+gpu could definitely clarify if there's a limit or not...

 

@weav: Edited. Try 135/360.

 

Don't have 2011 version installed. Will get it installed and run. I assume you want runs on both PSU's?

 

What tool you using to up the clocks on the GPU? MSI Afterburner?

 

+135 on core clock wont require it. Thus my edit. Memory can be clocked without vBIOS mod. I have done this with the Gigabyte P35X CF1 before I returned it for thermal throttling and power throttling.

 

Yes... if it does not bother you too much.

As I mentioned before, the combined test of 3D Mark 2011 is definitely the most interesting. Since it stresses both CPU+GPU to their limit, running it should give a definitive answer to our problem.

 

hmm.. looks like the free version only has 1280x720. Have to pay to get the 1920x1080 setting. Did you need 720p or 1080p? Not wanting to pay for an old benchmark. I do have 3dmark Vantage. Let me know if want me to run the 720p benchies.

 

Most people use the 720p runs with 3dmark 11

 

I wonder why Alienware machines see a big difference with the 240w PSU while so far it seems Clevo does not

 

720p is fine, that's the test that's commonly used as a reference.

 

Well... 17 inch screens require more power... I think. Or maybe Clevos are just enginereed way better...

 

Well the ASUS ships with a 240w supply, which means ASUS engineers determined it was needed. Companies won't spend money on something unless they have to, and a 240w supply is a bit more expensive than a 180w. Which leads me to believe that the ASUS requires the 240w supply to properly run the 980.

ASUS also comes with a blu-ray drive and a bigger screen. Maybe that adds to the power draw.

 

Yes, I was just kidding. Dunno, I think that the 180W is really at its limit on the P650SG, so probably your guesses are right. I think we have to wait the results of our laptop in 3D Mark 2011 with the 240W PSU first... so we can say if a 180W PSU it's enough for us or not.

 

Here ya go:

Dell 240W PSU w/adapter - GPU clocked +135/360

Ice Storm Extreme - 108030 (Graphics 184878 Physics 44007)
Cloud Gate - 22224 (Graphics 66454 Physics 6675)
Sky Diver - 22508 (Graphics 33973 Physics 8875 Combined 18348)
Fire Strike - 9022 (Graphics 10771 Physics 9391 Combined 3964)

First OC and first time I hit over 9k in Fire Strike! FYI that I never see Thermal Trottling only Current Limit Throttling. My ICD repaste is keeping things nice and cool thankfully.

 

Now you have to do a run with 3D Mark 11 with both adapters to see if redbytes is right or not during the CPU+GPU stress test.

 

Pretty much. Set everything to stock/default (no OC), and the run 3dmark11 @ 720p, once with the 180w psu then a second time with the 240w power supply.

 

Here are the benchies for 3D Mark 11:

Dell 240W PSU w/adapter - Stock Clocks
3D Mark 11:
P11109 (Graphics 12822 Physics 8076 Combined 7722)

Stock 180W PSU - Stock Clocks
3D Mark 11:
P11080 (Graphics 12826 Physics 8167 Combined 7459)

 

hm, 3.5% difference could be either a fluke or a real difference...

 

Noob question here, I haven't really used 3dmark 11 very much, is the physics test CPU based?


Sent from my iPhone using Tapatalk

 

Still think PSU isn't a player at stock clocks. Will run the benchmark three more times on each PSU to get a good average to be sure and put this to bed. Will post in a bit. Gotta say I do like how the Dell PSU blue LED on the power cord matches my power button on the laptop!

 

Um... the combined shows that there's a difference, even if subtle (around 3% as jay pointed out). But we should repeat the test on several laptops to achieve a significant statistical sample well, don't know about you, but I'm quite happy with this result, since it shows that IF there's a limitation, it's minimal.

This score also shows that my Kingston modules are crap... oh well. What tool did you use for overlcock in the prev bench?

Yes.

 

I'll be able to run tests with a 240W psu when I get my p650sg on Tuesday, the g73 that I have has a 240W dell psu that my brother modded to 5.5/2.5mm plug when he was going to put a Radeon 7970m in it


Sent from my iPhone using Tapatalk

 

Here are 4 runs on each PSU:

Dell 240W PSU w/adapter - Stock Clocks
3D Mark 11:
P11109 (Graphics 12822 Physics 8076 Combined 7722)
P11042 (Graphics 12785 Physics 8101 Combined 7472)
P11033 (Graphics 12778 Physics 8166 Combined 7371)
P11110 (Graphics 12788 Physics 8132 Combined 7745)
--------------------------------------------------
P11073 (Graphics 12793 Physics 8119 Combined 7577) AVG

Stock 180W PSU - Stock Clocks
3D Mark 11:
P11080 (Graphics 12826 Physics 8167 Combined 7459)
P11154 (Graphics 12827 Physics 8112 Combined 7881)
P11081 (Graphics 12825 Physics 8159 Combined 7476)
P11161 (Graphics 12801 Physics 8136 Combined 7959)
--------------------------------------------------
P11119 (Graphics 12820 Physics 8143 Combined 7694) AVG


In the previous OC'd test I used Nvidia Inspector to up the clocks to what LoneSyndal wanted.

 

At least now we know the results dont change too much or the 180W is just better.

The converter you used might be better than the bix, so I'll have to wait once again before I do some OCing again above 200/380.

 

Thank you. These results confirm the awesome work made by Clevo on this laptop, since the 240W PSU is really needed only for overclock!

Regarding Nvidia inspector... I'd love to use it too, but I have the sliders locked, and the unlock button is locked too maybe I need some modded drivers..

 

make sure to use 344.x drivers, since 347.x are locked for ocing

and of course prema mod

 

The adapter is OK but wish it had heavier gauge wire. It's about half the thickness of the PSU cords. I don't think it's an issue but heavier gauge is always better . Not sure if the Bix is any better on wire thickness as I don't own one. I like how the stock PSU has a barrel connector that locks in solid and has the detent ring. But who can gripe too much as the adapter was like 7 or 8 bux!

 

Mine would give out if I OC any higher than 120 core clock with the dell charger+converter. I'm hoping a different one would fix issue.

 

Here's the edited INF's for Mr. Fox's Desktop Drivers for those who don't know what to do. Just place them into the Display.Driver folder after extracting the modded drivers. (For P65XSX with 980M)

INF's

 

I think too many people are making too many assumptions here.

Don't forget that Clevo's 180W PSU has roughly 85% efficiency which means it can draw up to roughly 211W from the wall socket and still be within limits.
That leaves an unknown margin to exceed as well so there's headroom on top of that.

My P670SG with a 17" screen runs fine on the 180W PSU it comes with and OC's just fine too.
My Dell 240W PA-9e works like a champ with that BiX converter as well and is not even needed.
If other vendors choose to upgrade their PSU's because performance is hindered so much by the 180W PSU then it's just unfortunate, but Clevo's are fine with their 180W PSU's. Better design? Better PSU's? Who knows, it's sufficient.

Don't let yourself be influenced by one, only one, person that had it's Dell PSU fail on him and don't reach for your pitchforks and torches just yet.

 

I don't think the Dell PSU that failed really caused any concern. For me, it's really a curiosity as to why only Clevo's seem to run the GTX 980M at their full potential on a 180w PSU. The new Alienware machines have been proven to need a 240w PSU, MSI changed the PSU to 240w with their new machines, and ASUS ships all their new G751s with a standard 240w PSU.

Whatever Clevo is doing, it is doing it well. I just want to know what it is. Maybe it is just a high-quality PSU. When buying a desktop PSU, you need to pay attention to the efficiency rating. A high-efficiency 750W PSU is preferable to a low-qualtiy 1000W PSU. Maybe the same is true here.

Perhaps Dell just cheaped out on the PSU, and as such, they're paying the price now, by having to ship a more expensive 240w PSU in order to get the 980M to perform at its full (stock) potential.

At any rate, doesn't this just make you feel even better about buying this laptop?

 

Not all Dell PSU's are created equally either.
My PA-9e has a Dell sticker, but many PSU's are sold as Dell but are merely Dell compatible and have the same form factor without the quality.
And even then, Dell PSU's die too...

In my own tests the P670SG drew around 207W at most with Furmark AND some prime counter running, an impossible situation in real life.
Clevo really has been getting better over time and maybe now it's exceeded the standard of shinier names...

 

It's possible that Clevo PSUs are rated for the wattage delivered to your system, and the others are rated for what they pull from the wall. In this case, a 180W "pull from wall" PSU may deliver under 150W of power, which is barely enough for a 100W GPU and a ~50W CPU to run. But a 240W pull-from-wall PSU would be delivering ~180-190W of power. To be honest, 180W of power (with the brick being able to deliver above that for short bursts of time) is enough for stock operations of a non-OCable chip and a stock 980M. It's not IDEAL, but it is sufficient. The older AW PSUs were rated as the Clevos are, for power delivered, so they'd pull well above their "rated" wattage from the wall.

This is just my speculation however. But ~30W are enough for a system's average components to run, as far as I can see, and 150W for CPU + GPU appears to be covered with 180W of power. Unfortunately I'm not able to test the bricks and machines myself. But I *DO* know there are two ways to "rate" a power brick. I just can't remember the terminology for them. One is "RMS" rated, but I can't remember which one that is nor what the other one is.

 

Well, based on the tests done both here and in the Alienware forums, it isn't a stretch to assume the possibility that the AW PSUs are rated "at the wall", while the Clevo's are rated "delivered". Because we've clearly seen that the 180w Clevo PSU runs the GTX 980M just fine, while the AW17 requires the Dell 240w PSU to function properly.

 

To discover if we're right we just need to read the labels. My Clevo PSU is rated 19 V - 9.5A = 180,5 W output. A Dell-Alienware PSU found on Amazon is rated 19.5V 9.23A = 179,985 W output. So... maybe that's not the problem. Or maybe the Dell PSU can't actually output 9.23 A as it claims.

 

Hi all,
Can anyone tell me if the Prema BIOS updates the EC and BIOS, or is it just the BIOS.

Sorry if the question is a dumb one, have looked back through the pages and can't find an answer.

Thanks to LoneSyndal for the INF files!

 

It updates both

 

Hey guys! I'm planning to order the P650SE barebones next week and was just had a few questions before I went ahead with the purchase of it and also the other components(separately RAM HDD SSD)

Just finished reading the thread ....all 442 pages of it ), and saw some people saying that the i7 4720/4710 can't reach it's turbo clocks due to TDP limitations.

I know in the past Prema's Bios unlocked the TDP setting on some CPU's so that you could modify them.

Is it still the case in this laptop or is the TDP locked in perpetuity?

Also I'm trying to choose between 2x8GB HyperX Impact 1866Mhz/CL10 RAM and 1600/CL9 RAM. The difference in price is practically inexistent in my country so I would like to know which one you guys think would perform better for day to day tasks such as gaming/browsing/torrents. Does the higher latency offset the increase in speed? I know that the differences would be minimal anyway but why not go with the better option if the difference is only 5 bucks?

Last but not least for the moment I can't find any satisfying M2 SSD options where I live so was wondering if using a 7mm 5400RPM HDD for torrenting would affect my downloads in comparison with a 7200RPM 9.5mm drive.

If i where to stick with the 7200 HDD I would need to wait for a while for Samsung to release the 850EVO in M2 format but after that I would be satisfied knowing that I have the Hitachi Travelstar 7k1000(one of the best laptop HDD's on the market). If i went with the 7mm HDD i could use a normal sata III 850EVO and not have to wait anymore but be stuck with a 5400 RPM HDD.

Any opinions are much appreciated

P.S. My intention is to use the HDD only for torrenting and storing my music/movie collection so I believe I would get no benefit from a SSHD. Am i right or am I mistaken?

Also any better alternatives to the Hitachi Travelstar 7K1000 would be much appreciated!

 

The first thing I did when my NP8651 (P650SE) arrived was to change the stock thermal paste with IC Diamond. The first time around the temps weren't too great so I redid it and now my temps are very stable and never reach 90c in real world usage. I must note that I also undervolted and am stable at -60mv. When gaming my 4720 sits at 3.4Ghz, I've seen it sit at 3.5Ghz while windows was installing updates and I've never seen it sit stable at 3.6Ghz one core max.

Get the CL9 memory, lower latency is better, even if by 1.

I'm running 2 Samsung 850 Pro 256GB in RAID 0 for 512GB total space. I'm not the biggest fan of the RAID 0 setup, as it has benefits and drawbacks to a single drive. Another option is to have two separate drives and use them separately.

The difference in HDD RPM will be slightly noticeable, especially on load times and installation times.

 

Do also note that the test was done on Prime95 in which you will never run at. In 90% of general use purposes, you will never hit max clocks at any point in time.

 

That PSU isn't the same as what I used to test.

Here is the label on my PA-9E:



19.5V @ 12.3A = 239.85W

 

I already ordered some Gelid Extreme GC paste for repasting. but still I would like to increase the TDP limit if Prema's Bios allows it. That's why I was asking ....does it unlock the TDP?

I will have a single 256GB SSD for windows software and games and a HDD for movies/video/storage.

I don't plan to use the HDD for booting or running any programs on it...just simple storage of big files.

My question was if a 7200RPM would bring any benefits in torrenting over a 5400RPM HDD because i can not find any 1 TB 7200RPM drives in 7,mm format and the 850EVO is not available in M2 format yet. Since the 650SE can only hold 2x7mm or 1x9.5 mm drive I'm stuck deciding wether to get the 7200RPM drive and wait for the M2 version of the 850 EVO or to just get the 5400rpm 7mm drive together with a 2.5" 850EVO

Would also apreciate some other peoples opinions on the 1866MHz CL10 vs 1600MHz CL9 debate

I know that but I don't like it when my hardware is limited by artificial barriers. Also I'll be running some CPU dependant software as I'm just starting in the video/audio editing department . I have a band and want to make some video\s with our music....wich I will later distribute via torrents

I also play some CPU dependent games.

 

TDP is hardcapped around 60 for these chips. I have tested all I could on it. Any more and you would hit Current or Thermal limit (97C; and I used CLU).

 

Mines the same PSU as Weav's.

 

AnandTech | Memory Scaling on Haswell CPU, IGP and dGPU: DDR3-1333 to DDR3-3000 Tested with G.Skill

1866 CL10 is faster

when in doubt, then go for the frequency and THEN for the latency that article is for desktop chips & ram, but since were talking about the same CPU arch and RAM gen, the performance numbers are totally comparable

cheers

 

I read that article aswell but wasn't sure if the same applied to laptops. Thanks for the clarification

Now only need to decide on the HDD+SSD conundrum.

P.S for wireless I went with the Killer 1525AC.... i believe I made a good choice choosing it over 7265AC?

On the other hand I saw atleast 3 people complaining about the bluetooth not working on the 1525AC amd Cakefish proved it was a hardware problem by simply replacing with another new 1525 AC card.

Oh well... i can always swap it under warranty if that's the case or even live with it since I almost never use bluetooth.

 

Updating BIOS has shown to fix bluetooth issue.

 

For Cakefish it didn't...his was definetly a hardware issue but as I said...it's not that much of a biggie for me.

I'll just see when I get it. However I'm worried about the fact that qualcomm now withdrew special driver support for the Killer cards and now they are only covered by general drivers.

Also with the last killer suite available the bluetooth is not working under Windows 10 preview