Sager NP9377 / Clevo P377SM-A Review by HTWingNut

I'm in Albuquerque, so my environment is pretty much spot on the same as yours (i.e. high altitude, low humidity, 72F inside).

And like I said, I am running a Cooler Master Notepal U3 under mine.

 

That's about halfway through a Furmark stress test. The fan isn't on its highest yet, it was low til 87 then it kicked up a bit and temps are staying at or below 85C now.

Furmark says it's throttling the entire time you run it but that's just a Furmark being outdated issue.

Sent from my HTC One_M8 using Tapatalk

 

Sometimes those coolers make things worse. Have you tried without the cooler?

Sent from my HTC One_M8 using Tapatalk

 

In the end, it hit 87C max and the fans didn't kick on max.

It's not ideal but it's acceptable. I should note that same test on my 260M will cause them to pass 90 and throttle.

Sent from my HTC One_M8 using Tapatalk

 

I haven't tried it without the cooler, so I will give that a shot tonight and report back.

I wonder how much affect the thinner air is having on our ability to cool these laptops. Maybe I'll take mine to lower altitude tonight to test. haha

 

Honestly just elevating the back end will probably shave 2-3C off the top by itself lol

 

I've been thinking some more about my temperature issues, and I have a couple more hypothesis I'll test tonight.

1. I'm using a metal laptop cooler. After I've brought the GPUs up to temperature (~85C) and then let the system idle, the temperature comes down and idles around 50-55C (quite hot for idle compared to others). I wonder if the cooler's metal structure is heating up and then the laptop fans are pulling air over this heated surface, essentially pre-heating the air. yuck

2. The back of a laptop is basically up against my desk cabinet, and is exhausting the hot air into an open faced cube about 10" square. The air in this cube is quite warm after running the laptop for a while. This could entirely be a "duhhhhh" situation.

Additionally, I did some heat transfer calculations to compare the CFM requirements to remove 120W TDP from the GPUs at sea level, and at 5000 ft altitude, and the difference is negligible. So I don't think altitude is playing a part.

More testing tonight, and I will report back tomorrow. Wish me luck!

Kevin

 

Yeah my laptop has open air all around it. I have no doubt if it was in the kind of constraints you're putting it in that the cooling would not be adequate.

 

I was using the Notepal U3 with my earlier Sager np9380 and it helped a bit on that machine if I remember correctly it was able to take out 2-3c but nothing more.
The 3 fans can be placed exactly under the laptop fans for extra flow.
In your case I stongly believe that your temps are not caused by the laptop cooler... Maybe positioning or it's just how the system is or the thermal paste.
U still should be able to get the same temps as ethrem.
The extra heat sinks applied to ethrem machine will help delay the rising temp for some time but end result should all be the same.
Still he gets max 87 , work on achieving that result.

On a side note I managed to reduce like 3-4c on my np9380 just by adding some force on the heatsink of my 7970m... Dnt ask how it worked but it did.
I just used a small screw driver n forced it a bit with care on sides n center and
It worked for me.

 

Well just for giggles I forced the fans to max

Forced max

Normal fans

There's a definite difference.

The fan on maximum is unbearable though.

I'm going to have to pop the cover off and take a look. My temps are much higher than HTWingNut's were.

 

Afraid not. That is in fact your gpus throttling.

 

Really? Because here I was under the impression that all Kepler GPUs will show that due to their boost clock functionality not being active 100% of the time. It is throttled from start to finish. Please show me a Kepler GPU that doesn't exhibit this behavior that has boost enabled?

 

I did that already, but everyone wanted boost disabled on the 780M
Also did it with the 880M and it boosted the whole time and then dropped when testings was done.

 

I supposeI should have said any stock Kepler, desktop or mobile. By default boost isnt supposed to run 24/7 which is why it always shows throttled on Kepler. That was my understanding of it. If that is not the case, well, learn something new every day.

 

Not trying to be the know it all or anything like that. Just letting you know from my testing point of view. That picture in my sig actually means something.

And boost will run all the time if everything is running optimal for it to do so.
Meaning: Temps and power draw. Room temps have a lot to do with how you system will perform. More so than that "each hardware is different" speech.

 

Yeah I understand the idea behind boost. Personally I think it causes more problems than it helps.

I apparently have discovered an incompatibility with GPU-Z and FurMark. Running GPU-Z logging during FurMark has frozen the system twice now. I'm going to give it a go with BioShock and see what happens. This is getting to be annoying though.

Its 22C room temp and there is nothing but open air all around the machine so its the machine generating the crazy heat. I was hoping to go without having to open this thing up for awhile but I don't like my GPU running at 85 when I'm gaming...

 

It's better to use on screen display and watch it in real time.

 

Here's BioShock, don't know why it crashes with FurMark

BioShock Ultra + DDOF

Clock is fixed around 1GHz.

EDIT: Same story with Catzilla

Catzilla

Guess I'm gonna go do a bit of BioShock actual gaming and see what's in the log. I didn't know about this log reader until I just did a Google search thinking there must be something out there haha.

 

Easy. Furmark takes as much power as you can give. And if that is not enough, it will throttle.
Games work a bit differently. Most aren't trying to drive crazy high watts like furmark. And fur mark has been known to kill more cards than any other software. It can kill cards on stock, but more so when fully unlocked cards run at near max clocks.

 

Yeah, I've killed a desktop card with Furmark before. I'm very careful with it and don't really like to use it at all to be honest. I just find it odd that GPU-Z won't log while Furmark is running, even when I put it on battery saving mode it still froze the machine.

Well it looks more than acceptable. Bioshock Infinite pushes the GPU pretty hard, especially when it has no frame rate limit and its set on 120hz refresh.

This is what some gameplay looked like

Bioshock Infinite

Its obviously not what the card is CAPABLE of but it is better than I thought. The system does an adequate job of keeping the card cool once it passes 85C and the fan system kicks in. That threshold is a bit high though if you ask me and it didn't go near the max (although my ears are thankful for that). You can clearly see in the graph where it hit the mid 80s and then quickly dropped off in response to the fan.

 

See how the memory is a straight line? Well that's how you core is suppose to look like.

 

Yeah I know, hence why I said I know its not where it could be. It works for now though, I thought it was worse than it actually is.

Still doesn't explain why I'm scoring less than HT in the Futuremark benches with the same drivers and a faster CPU. Since I actually am a touch faster in Catzilla, Unigine, and Bioshock, I'm going to assume that it is likely an incompatibility between 3DMark and the 4940MX right now.

 

So johnksss .
When you gonna make every one with an 880m life easier ....
When will the vbios be released ?

 

So johnksss .
When you gonna make every one with an 880m life easier ....
When will the vbios be released ?

 

Better to ask over at tech inferno really.

 

Yes, but I think it's still far slower than the 780M at the same speed.

Will still get the same answer since it's a "Team" effort.

Soon. When other situations work themselves out.

 

Just joined tech inferno website and upgraded to elite member .
Looks a very interesting website tbh ... Lots of reviews and forums are loaded.
Now we wait I guess....... Not in a rush yet systems comes on 3rd June finger crossed.

 

Did my testing last night, but it's not extensive due to time constraints (i.e. I have two young ones).

Just used the demo from 3dMark11 to have a consistent test. What I found was the cooler does help (as compared to simply propping up the back) by about 1-2C on the GPU and 3-4C on the CPU, and moving the laptop away from cabinet gained me about 1-2C on the GPU and 2-3C on the CPU. This test was just to get an idea if those changes would help, but it's not a real world temp soak. I will try soon to get back on The Witcher 2 (TW2) and see if my temps are lower. I'm not too optimistic about it though, as I'm really only a degree or two lower from my original conf. I think TW2 will still hit 90C.

Also, I haven't re-applied thermal past yet, so that's still to be done.

Kevin

 

Such issues can ruin the gaming experience ...been there on the np9380 , didn't enjoy it swapped to desktop gaming for a while ..
Seems will have the same thing with the 880ms ... This time I'll just close my eyes and enjoy the machine the way it is I'll try not to bother much...as I said I'll TRY.
Getting 85-90 isn't healthy for a GPU that's for sure....and in few weeks/month it will go higher due to dust , so guys keep your fan super clean.
Something I wanted to ask which I don't know much about.
So let's say I have a gtx780m at stock bios and it throttles , after Installing the a proper bios mod for it , will it stop throttling directly or there are other adjustments that are needed to be done to achieve to the point where throttling stops ?

 

Honestly 90 isn't all too high my desktop 780ti can get up to 87

Sent from my Nexus 5 using Tapatalk

 

My desktop with 670 sli gets max 74-78c after hours of metro last night with slight OC.
I call that proper cooled GPUs.
A desktop GPU should not run at 87c in my opinion.
No GPU should but guess all that power put in these things got us to where we are today.
Maybe that's why we have maxwell incoming.

 

To be fair a 670 doesn't really compare to a 780ti.

 

The 670 is a 170W card, the 780Ti is 250W. One has almost 1/3 less the TDP of the other. All else being equal, the 170W card will run much cooler.

 

Yeah considering that the fan isn't going nuts when its hitting 80s, I'm assuming that nVidia has determined that to be a safe operating temperature. Isn't the upper limit in the 90s to 100s for Kepler?

I'm not entirely happy with the temperatures but watching the graphs, it seems that it is the way the cooling system is designed is to kick in when the 80s hit. Short of modified fan tables and increased noise, I don't really think that will be resolved. I've had cards that run hotter than the 880M that didn't have 1/10th the speed.

Kevin, sorry to hear that your temps are still stupid high. I would guess in your situation, its a paste problem most likely. I don't understand how they can let so many machines go out with bad paste jobs... I mean supposedly these systems are put under rigorous testing... you would think someone would notice the temperatures......

 

780 is flagship of desktop GPU and 880m flagship of mobile GPU ...they have the highest power thus giving out the highest heat ... Thats what I really meant in my post .
Temperature SHOULD be around what my 670s are having....
But due to the excessive power these flagships GPUs will end up with these situations.
Maxwell will prob solve this.

 

I hope Maxwell does solve it but I'm not going to hold my breath. Heat issues have been promised to be resolved since the GTX 260Ms and its either stayed the same or gotten worse across each generation, with a few exceptions.

 

Maxwell will have 100W cards too. Sure the 100W Maxwell will likely be much faster (>50% hopefully) than the 880M, but it'll produce just as much if the TDP stays the same.

Also, you can't compare desktop temps to laptop temps, the cooling is much more powerful and efficient in a desktop. I'd say as long as you keep your cards in the mid-80s range you'll be fine. I had a crappy Toshiba with a GTS 350M that constantly ran into the high 80s while gaming, and it's still kicking after 3 years.

 

The desktop coolers are less efficient usually but much bigger.

 

Just realized that the 87 Celsius temperature I reached is exclusive to linux.. for some reason linux realllly stresses out my cards, makes me hesitant to use it to game at all.

I get about 78 Celsius under load in windows.

 

Anyone know which GPU is GPU1?

 

Should be slave GPU; master GPU is GPU0.

 

Sorry, let me rephrase my question. In MSI Afterburner, I show two GPUs. The first, which I believe handles most of the rendering, is labeled GPU1, and the second which handles Physx is labeled GPU2. Looking from the bottom of the laptop, is GPU1 on the left or right?

 

The master GPU is the one off by itself, the secondary heatsink loops round and sits next to the cpu heatsink.

 

So the master is the GPU on the left, looking from the bottom of the laptop. Correct?

 

Well, how about them apples!

So I re-applied my Turniq TX-4 thermal paste to both GPUs, making sure to only apply about 1/2 a pea size since the contact surface is pretty small compared to a desktop CPU (which a pea size is usually recommended). Ran my test with 3DMark11 demo and the secondary GPU's temp came down considerably, but the primary GPUs temp was still high (80C). So I immediately took the laptop back apart, unscrewed the primary GPU with it still hot, and wiggled and pushed down on it until I felt a metal-on-metal scratching feeling. Buttoned her back up and re-ran the test.

The main GPUs temp came down to about 75C, woohoo! Played The Witcher 2 for a bit, and the temperature never exceeded 83C, whereas before it would hit 90C and throttle pretty quickly. Hallelujah! I'm satisfied with the temps now.

Kevin

 

Glad to hear you sorted your temp issues Kevin.am hoping my thermal paste job would be as good
As the ones you did.
Ethrem so am still waiting your thoughts of the machine overall ... Did you play more games on it ?
How's the throttling and is it really affecting frame rates badly ?
Have you noticed any difference from 60hz to 120?
Feed us up mate.

 

I just got Watch Dogs and i'm going to update the nVidia drivers... if the performance is not fixed, I'm going to be disappointed.

A full review is going to take me some time - I'm still tinkering.

 

Depends which orientation you are talking about, but like I said the secondary's heatsink is right next to the cpu heatsink with the longer heatpipes.

 

Got it, thank you Meaker.

 

Ethrem,

Would you mind elaborating a little on what performance issues you are having? Problems with the game or your hardware? Framerate issue? Very interested as this is a game I plan on getting soon.

Thanks!

Kevin