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

How did you mod the dell psu?

 

Tried using a simple bix converter. It does work normally with the machine when not OCing.

 

I'm surprised you killed the psu though.

 

Technically it still works. When you start pulling more power, you can hear the noise from the PSU at the connector to laptop. PSU then shuts itself off. Works if you pull it out from source and plug it back in.

 

Sorry to hear that LoneSyndal. Very disappointing. Oh well, perhaps another reason to start focusing on the P75xZM.

 

I am a bit surprised by that as the gs60 only has a 150w adapter and I know that on their boards people complain about that sometimes not being enough for the 970m.

 

Depends on your needs. Since I am mobile, the utility of the longer battery life outweights the P750ZM's.

 

*couldnt edit on phone*

Its due to using 3D Mark from Steam from what I know. Someone earlier managed to get the CPU to be properly utilized and heavily outclassed all other users' results in the CPU based tests.

 

Understandable. I am always plugged in with my laptop so battery life really isn't much of a consideration.

 

Don't expect performing more than 60fps with ALL settings maxed out as ir has some demanding specs which should kill a desktop GTX 980m too.

 

I am in fact already using the standalone version. Running out of ideas.

Perhaps the low scores are because I only have one RAM stick? I doubt dual channel memory would make a huge difference to the icestorm score as it's half of what it should be.

 

Max CPU temp of 77° in XTU bench, clocked with stock values. i7 4710 HQ.

With +100 MHz it reaches about 10° more, 87°. Doesn't worth the difference...

 

I can confirm that having dual channel memory extremely bumps up 3dmark11 and intel XTU scores.

Obviously I'm not talking about 50% performance difference but maybe about 25% could be more realistic.

 

That is interesting, how a 100MHz bump translates into such an increase in temps ? Are you talking about an increase on one core or all of them?

 

I haven't checked it.

Anyways, it has been the second run, so it surely was way hotter.

 

i did, got the same problem. Any more ideas?

 

I think not all versions of Nvflash work from within windows. If nothing works for you that was listed here, then something has been done differently or your issue is very unique. Try making a help thread on it.

 

Out of curiosity I ran 3D Mark benchmarks too. My results are:

Fire Strike - 8384
Sky Diver - 21514
Cloud Gate - 21776
Ice Storm - 86823

Stock CPU&GPU, balanced power plan, my config is in the sig.

 

Dual Channel RAM?

 

Yes. A couple of basic Kingston modules.

Edit: oh boy, now I notice you didn't quote me. Well, anyways...

 

Postisup.I tried the two different methods, but i noticed something strange. When i write "nvflash --protectoff" (or on) i get the same exact error, any thoughts?

 

Ha, no worries.

I guess, for now, I can attribute my poorer scores to not having dual channel memory - wishful thinking but can there really be anything else - should I even be worried?

Would be interested in 3dmark scores from notebooks with a single RAM stick

 

Okay thanks, maybe I'll try that. But is it risky updating/flashing bios? I've never done it before. Does it come with any more benefits, like better temperatures or anything? Thanks

 

Um, I sincerely don't know what to say. Yeah, it's definitely possible that's a RAM thing. Other things... driver version? I'm running 347.09. Are your scores with stock clock or OC? Did you check temps to see if there's any throttling during the bench? Do you have latest BIOS/EC from XMG website?

 

Well, several tests over on the Alienware forums showed conclusively that the 180w PSU was holding back the performance of the 980m. Granted this is on Alienware systems, but a 980 is a 980.

Additionally, there are some tests out there showing the 980M draws as much as 210w while under load (in combination with the CPU). In other words, in games where both the CPU + GPU are stressed, 180w is not enough.

MSI got some flack for shipping out their 980 systems with weak PSUs, and I believe they've made changes.

Also, ASUS ships 240w PSUs with their new ROG G751 systems.

I'm just relaying information, not making this up myself.

 

Yes.. 2x8GB Gskill cas9

 

I've been on 344.75 for the OC and now on 347.09 for testing purposes - scores remain the same. I'm running stock and the temps are in the 70s, the GPU clock remains constant and zero thermal and power 'throttling' detected by XTU (although the CPU clock hovers between 3 and 3.4GHz.

It seems my physics scores aren't up to suff - In icestorm Iaverage around 130 fps in the physics test.

 

Sounds like a faulty/incompatibly PSU.

EVERYBODY with the new AW17 who switched to the 240w PSU instead of the standard 180w significantly increased the STOCK performance of their 980. Everybody.

Just trying to point out that pretty much all the other companies shipping machines with the GTX 980M are either providing 240w PSUs, or have taken flack for not doing so, after users conclusively showed that 180w PSUs were hampering the stock performance of their machines.

Maybe Clevo is more efficient, and doesn't need that much power to satisfy the stock requirements of the 980M.

But don't be so quick to dismiss.Again, I'm just trying to help people here obtain maximum performance from their machines, and I'm not looking to start fights or anything.

 

Can you link us those proofs? I'm quite curious.

You didn't answer about BIOS/EC. Do you have i7 4710 or 4720?

AFAIK I get around 130 fps too in icestorm physics, anyways.

 

At stock, there's no additional benefits and works the same as 180W. My test was during OC which caused problem.

 

Is this compatible with our laptop? The connector seems identical to me.

(if it's bad to link shops I'll edit the post...)

 

Sorry, forgot about those things - CPU is a 4720hq and I was getting stuttering and throttling with the stock BIOS, so flashed the latest premamod BIOS/EC (.03/.04) and that fixed it. Also using prema's vBIOS.

 

So... our configs are almost the same . Only diff is the RAM and stock BIOS/EC vs Prema's... but that should not affect the score. Well... you should buy a 2nd module then

 

You'll have to ask if it's a 2.5mm/5.5mm barrel connector.

 

I'd really like to come to that conclusion to put my mind at ease. But damn, seems I've missed a trick with the RAM - I've not been keeping up with the news this past decade - all i remember is my desktop days when the Barton Athlon XP's came out with dual channel support and it had absolutely no impact on performance.

Ramzay, does your setup have one RAM stick? We're the only ones in the world with such low icestorm scores, and we may also be the only ones with one RAM stick.

 

Might be a possibility. I also only have 1x8GB of RAM, and I was comparing my scores to the ASUS with 2x8GB. Maybe dual-channel does help.

 

Here's one thread with several posts saying they get better performance

http://forum.notebookreview.com/ali...nware-17-3dmark11-results-16.html#post9903262
http://forum.notebookreview.com/ali...nware-17-3dmark11-results-10.html#post9901507

I'll see if I can find the other posts, they're strewn all over the place.

Again, this is for the Alienware systems, I admit. But AW, ASUS and MSI have all shown (through either actions or tech support) that the 240w PSU is what should be shipping with the GTX 980M

Maybe Clevo is different, but that's what I've seen so far around the forums.

 

I think it may be different. Cause with the 180W, I can still force CPU into 3.4 GHz for 4 active cores while running GPU overclock at 190/360MHz (stable overclock without artifacting or hiccups or crashes). Not to mention, comparing the scores of those Alienwares, it seems that they score below 8K in Firestrike with either adapter. Could possibly be a motherboard thing. I may do further testing with 240W PSU later at stock instead of OCing.

 

Have anyone here flashed back to the stock vBios after using premas?

 

Please do, I'm interested in knowing what the issue with those new Alienware machines is.

 

Download NVFlash 5.190.0.1 for Windows | techPowerUp

 

Downloading 3D Mark 2011 so I can compare my score with theirs. I'll post the results here.

edit: to be fair, I see the problems reported are with the i7-4980HQ, which is a very different CPU from the i7-4720 (or 4710). If you remember the first reviews of our P65xSx with the Iris Pro GPU the temps were waay higher... i7s with Iris Pro are renowned for being power hungry and very hot

 

Thank you,but when i use "nvflash -4 -5 -6 (mybios).rom" it says "blabla OEM only" but when i just write "nvflash (mybios).rom" it works. what is the main differents?

Update: "WARNIG: Firmware image PCI subsystem ID <10DE.000>
does not match adapter PCI subsystem ID <1558.6556>

Help? :S

 

should be fine. check in nvidia inspector that it no longer says modified under the firmware/bios section

That would explain a bit. They may have been throttled by that chip alone. Hopefully no thermal throttling as well.

 

My score is:

Score: P10927
Graphics Score: 12876
Physics Score: 8075
Combined Score: 6809

Alienware does:

240 watt psu: P10824, g 12482, p 7851, c 7580
180 watt psu: P10497, g 12476, p 7504, c 6599

Edit: with 4710mq:

P10102, graphics 12226, physics 7642, and combined score 5553

Soo... given that they have a much more powerful CPU (+400 MHz in turbo boost), to me the score of my P650SG is impressive. Better overall, but must importantly better graphics ( 12800 vs 12400 ) which shows that we should not be limited by the 180W PSU at all at stock settings.

edit: naturally, to be completely sure we should test a P65xSG with a 240W PSU... But AFAIK the only one in this thread is dead

edit/2: ASUS ROG G751 makes P11304 with the 240W PSU. It's a bit higher, but with 24 GB RAM. Mine has only 16 GB CAS11, so my config is not the optimal one.

 

Technically still works, just cant keep up

 

Well... it would be nice to have a test if you manage to get it working properly somehow. As I wrote in the Alienware thread, it would be very nice to have a proof that our system is limited/not limited by the PSU. I'm quite concerned, since the combined it's a full 10% lower than the Alienware with the 240W PSU... but we're still comparing a 4980 to a 4720, so I really don't know what to think.

 

I have a 240w Dell PSU and adapter for use on my machine. I can run the benchmark again with that psu and report back.

 

Good luck and hope your adapter doesnt die during the run.

 

Great! Please post both results, with 180 and 240W PSU