*** Official Clevo P37xSM-A/Sager NP9377 owner's lounge ***

You want to undervolt the CPU voltage offset.

GPU-Z reports a lot of things wrong... These machines actually run PCIe 1.1 out of the box.

RJTech sells the video card and heatsink for 720, haven't seen a better price. Model is P377SM-A and even comes with the x bracket for our machines and the heat spreader http://rjtech.com/shop/index.php?dispatch=products.view&product_id=30303

 

Wow that is pretty convenient. They have everything I need well except for the cooling fan. But thanks guys

 

dynamic cpu voltage offset?

 

And my laptop is back to RMA... DAMN usb sticks... Went to flash a new Bios... The USB stick was f*cked up... -.-'
My bad luck... Almost as bad as bad luck Brian LOL

 

Yup that would be the one.

@joluke damn that sucks! I'm on the fence about RMAing mine again because of my GPUs being fireballs. Waiting for a kill-a+watt to get here so I can see if there is some abnormal power draw by the motherboard or something. So weird that two pairs of 880Ms and now my pair of 980Ms have gotten progressively hotter. My 980Ms now need max fans in order to not throttle (the master hits 87C and throttles if I have auto fans on when auto fans when I got the cards back in December never kicked up at all and maxed at 82C) so I'm starting to think it might be something electrical. I've got the thing plugged into a heavy duty surge protector so I guess I'll just have to see the power draw..... *sigh* Good luck with your RMA!

 

Maybe you have an issue with your heatsinks Ethrem?

 

According to Prema my EC failed to write when it was being updated... But i'm currently looking at the worst scenario, like a motherboard replacement... Something like the first RMA I had with my old P370SM, which after RMA, got updated to the P377SM-A with dual 880M's from Clevo themselves
Let's see what Clevo and my reseller, what they say about it

 

I thought about that but these are actually much better than the ones I had originally. When I went to take the 7921 off to put the IC Diamond on, I was amazed how perfect the coverage was, there was a perfect impression on the heatsink and so much of the paste got squeezed off on all four sides that it really was just filling in the gaps. If it was the heatsink IC Diamond would have improved the temps but it actually made them worse by about 2C so I don't think it's a heatsink issue. It's the same story as both pairs of 880Ms. Started off great and got progressively worse under the same conditions.

Man, Clevo sounds like they're easy to work with. I really had to push Sager on my issue to upgrade to the 980Ms (and I had to pay for the upgrade as well). I can't even begin to imagine what I will be in for if I start another RMA. I think in my case though that Sager was thinking I was just trying to get an upgrade at their expense, which is ridiculous considering all the documentation I sent them through Xotic.

Hopefully they get you back up and running quickly.

 

Yah I didn't think Clevo or Sager covered flashing new bios either so that's surprising.

 

I'll have to pay for my repair... The warranty doesn't cover bad flashes on bios/ec..

 

Yah so then if its the motherboard thats not gonna be cheap, good luck man

 

Had it sent to RMA before so it will be around 350usd to 400usd. At least it's fixable... And I'll have it functional in one month..

 

That doesn't sound bad at all, hey if you can ask them how I should connect my 120hz monitor while they are in there by the motherboard lol jk good luck man that sounds like a very reasonable price tho

 

That's why I bought a Clevo and not an MSI or an Asus. They do care for the customers.. Well at least they are not the worst.. ASUS.... Jesus... Worst customer support and rma support ever! Lol
But that is another discussion

 

Well just an update I am LOVING this laptop I can play GTA 5 all maxed settings with vsync on and it never moves from 60fps so constant 60 fps looks amazing and I just ordered an xbox controller to use while I play so can't wait to use that

 

Haha, yup me too . The controller is a must IMO.

I hope you can get that 120hz display fitted. I'd be very interested in your opinion once you see 80+fps in GTA V

 

Lol maybe a controller for driving but for on foot you can't touch the keyboard and mouse.

 

You're right, I just got lazy and find the x-box controls easier to remember in those moments of panic!

 

Ethrem,
Thanks for the reply....I guess when I downloaded the update for the Nvidia...it must have defaulted back to turning SLI off...
so anyways I reran the 3Dmark tests with SLI on....
I am going to post the results from everything shortly

Here are the results.

Hello Ethrem, Everyone,

I would love to know how I stack up against other people with the same system with 880's and how I compare to people with 980's.....

I also would love to know if based on all these results if there are tweaks I can make using the XTU software....

Thanks for any and all input....

 

I went ahead and merged those all into one post.

Looking at your benchmarks, they seem to be on par for stock 880Ms. Trust me, you don't want to compare Kepler to Maxwell, a 970M blows away the old Kepler cards and 980M even more. But your performance for the video cards is on point.

Looking at your CPU, however, it is a bit disappointing to see the chip throttle to 2.9GHz during the stability test but the XTU stress test seems to throttle a lot less. Max temp of 88C on the CPU is pretty much what I would expect from a stock paste job. You may want to consider redoing the thermal paste to shave off a few degrees but if that CPU is anything like my 4940MX, you won't get it under 85 with a turbo load.

In XTU, set the dynamic CPU voltage offset to -70mv and run the stability test again. Hopefully you won't throttle as much.

 

I thought that the 880ms were more closer to 970ms on 3dmark11-firestrike...but as far I can see your results they have differences

 

http://www.3dmark.com/3dm/6733622 - STOCK I really wish I knew how to put this 120hz display panel in, I can't seem to find any videos on youtube or anywhere else with this beast's display being replaced or disassembled..... @kolias would you have any interest in showing me what you did to your p377sm-a so that I could try doing it to mine and see what kinda results I could get, I am not much into benchmarking but I am a gamer so anything that can boost performance without putting the system at risk for harm I am down for, It would be awesome if you could educate me a little if you have time I would really really appreciate it thanks

 

Why the hell is @Brent R.'s CPU package "BGA1364"?
I checked my old FS runs and I get "FCPGA946".

I would really like to see the inside of your machine... that CPU should never report using BGA.

Have you seen anything like this @Mr. Fox @Prema @octiceps @n=1 ?

 

@D2 Ultima - I have seen numerous examples of Futuremark's database reporting bogus information. Not that one specifically, but others that are similar. It is because their hardware database has a typo and/or the Futuremark SystemInfo scanning utility is reporting information that has no matching hardware information in their database. Keeping that database updated is likely a never ending chore, and I suspect it is a massive database.

This happens every time a new GPU is released. The GPU will initially be identified as "Generic VGA" when you run benchmarks, but CPU information can also be messed up. In the case of a new CPU, it is usually identified incorrectly.

Some Alienware 18 benchmarks that I saw on machines with 4910MQ and 4940MX were initially mis-identified by Futuremark as a 4710HQ (which was not even an option). The information under Package that reads BGA1364 is clearly an error.

When I come across that kind of error on my own systems, I send an email to them ( [email protected]) so they can update their database. I provide detailed hardware ID, machine model detail, links to benchmark examples and a screen shot. I get a reply, usually in less than 3 days, once it is corrected. The fix is also normally retroactive, and any previously submitted benchmarks reflect accurate information based on what was corrected on their end.

Below is an example of an email I sent for 980M installed in an M18xR2 back in December.
[parsehtml] [hr][/hr]</hr> [/parsehtml]

From: Mr. Fox
Sent: Monday, December 22, 2014 3:32 PM
To: [email protected]
Subject: Request to Add to GPU Database

Hello. I need to have my hardware information added to your database for my GPU configuration. They are not recognized and show up as “Generic VGA” in error.

Thanks, and Merry Christmas!

Mr. Fox

Benchmark Links:
http://www.3dmark.com/fs/3528857
http://www.3dmark.com/3dm11/9165595

System:
Alienware M18xR2

GPU Hardware ID:
PCI\VEN_10DE&DEV_13D7&SUBSYS_05501028&REV_A1
PCI\VEN_10DE&DEV_13D7&SUBSYS_05501028
PCI\VEN_10DE&DEV_13D7&CC_030000
PCI\VEN_10DE&DEV_13D7&CC_0300


Screen Shot:
(attached)
​

[parsehtml] [hr][/hr]</hr> [/parsehtml]

 

What is BGA? what does all this mean sorry....I have never ran futuremark.....whats all this mean is it a good thing or bad thing? what is FCPGA? please explain whats up?

 

BGA = Ball Grid Array

It's a method of attaching a chip to a board using small balls of solder as different connection points.

As it uses solder the chip is not removable unless you use the right equipment.

FCPGA = Flip Chip Pin Grid Array

It uses and array of pins that mate with a socket thus allowing the chip to be inserted and removed without the use of special equipment. The downside being the socket adds vertical height to the product and you get less connections (The BGA and FCPGA are the same size package yet the BGA chip has an extra 400 connections).

 

but my chip is suppose to be removeable they were suppose to put IC Diamond on it....I am confused..I think it is just an error in 3dmark reporting it incorrectly...

 

Exactly, as he said it's a reporting error.

 

okay thanks I will send them the info with screenshots I guess I know HWINFO reports it correctly but I didn't even notice that @D2 Ultima good eye

 

They do sometimes break the detection as they go along and edit things.

 

The thermal paste they use goes on the die of the CPU. Whether or not it's in a socket does not matter. Either way, the die looks the same and can have paste applied to it just the same.

I always look for the little stuff. Sometimes I catch it.

I wonder if all of these inconsistencies (like CPU-Z reading a friend's 4810MQ as BOTH a 4900MQ and a 4810MQ, or FS reading this) are because they try reading the chip's data itself, and that silicon was initially planned to be something else, and they didn't write in all its data properly? That'd definitely be something weird, especially if not all machines do it.

 

They tend to make a best guess based on cache levels, frequencies etc, sometimes they get it wrong, especially more with mobile chips.

 

sometimes my 4810MQ also reads as a 4900MQ

 

Exactly, as it's pretty hard to tell the difference between the two, it comes down to cache and a couple of feature bits.

 

Well aren't you just so full of knowledge thanks so much, I wish there was a way for me to download all the info you guys know about computers overclocking and etc....would make things alot easier S Meaker and Thanks D2 for bringing this up

 

Then some of us would be out of the job so at least no such system exists just yet lol.

 

Yeah. We all have our little secrets. Meaker won't divulge all the info he has on GPUs and I won't divulge the secrets of streaming without some... persuasion... XD

 

I went ahead and set the dynamic cpu voltage to -70 mV....I'll run stability test later tonight.....

once you change something like that, and then you apply it, does is stay changed, or once you reboot, do you have to open xtu and change it again?

Also, I noticed that when you open XTU, to the right there is sort of a chart that shows "core, default, and proposed settings" I noticed that everything matches for default and proposed, except under default, for processor current limit, it shows 73.000 A, and for proposed, it shows 72.000 A.

I am assuming that I should change it to 72? What are your thoughts on this?

Also, I read that you said based on my temps, it was normal for a stock paste job...supposedly sager used ic diamond paste? should I just roll with what I got, or would it be worth it to repaste?

Thanks!!! I am still a rookie but learning thanks to everyone in the forum.

 

XTU settings apply until the watchdog is triggered by something like a crash or bad start up at which point they reset. Current limit can be set a bit higher to avoid throttle. Try 110 amps.

 

"watchdog" is I'm assuming already built in to the processor or bios? So you are recommending that I change the "processor current limit" from 73 to 110, even though XTU is saying 72?
is it recommended to decrease the dynamic voltage offset...then run a stability test, then increase the processor current limit?

or you can decrease the voltage offset in combination with a current limit increase and run a stability test after that?

 

Hello everyone,

I am looking into getting a P377SM-A for my first gaming PC (for work I have to travel between the UK and Japan) and have been doing some research. If you guys could answer some questions for me I'd be really grateful. I am looking at a company in the UK called PC specialist who have the P377SM-A under the name "Vortex IV Elite" I'm currently saving like a crazy person so I can buy this machine when I go back to the UK for a month in July.

http://www.pcspecialist.co.uk/notebooks/vortexIV-Elite-SE/

1) I understand heat is a major concern with enthusiast laptops. PC Specialist offers something called an 'Arctic MX-4 Extreme Thermal Conductivity Compound as their upgrade, but I've noticed other sites based in other countries use this IC Diamond compound as their top choice. Will this MX-4 be sufficient or should I be looking into getting something separate and applying it myself? Japan's summer is crazy hot so I'm really going to need to be careful about the heat thing.

2) Brent R, you mentioned how your laptop is giving really good performance for GTA5 at 60fps on ultra settings. Can you do a screen capture of the settings or just give me a general idea? Is it holding up at that high level of performance even in the countryside areas, which I've heard are an FPS hog? I've been hoping to get a laptop that can run this game at a really high level so that reading your post about it was a relief!

3) It seems that all the companies talking about this laptop are only factoring in the machine itself when they talk about weight. Does anyone here know how much the laptop in a carry case with the power brick weighs? I'm going to be carrying it on long haul flights; I'm not worried about carrying it as such, just wary of going over my weight allowance for my airlines.

lengthy post, I know, but I'd be really grateful if anyone could help me out with these queries as I'm new to all this...

 

Get something else and apply yourself. Without question. IC Diamond is a good thing. Gelid GC Extreme appears to be another good thing.

The countryside areas are not CPU-hogs, which make them GPU-bound. The city areas are where FPS is lowest. Most people who claim the countryside has the worst FPS have rather weak GPUs and strong CPUs, where the bottleneck is reversed.

Power brick is ~3 pounds maybe? The whole laptop with SLI and decked out with something in all the slots and the power brick as well is somewhere around 12-13 pounds. Accessories will add to that, of course.

 

Thanks, it looks like the weight will be just fine for my carry on luggage then.

My plan currently is to get this laptop with 980M SLI and a 4940MX CPU. Should this be enough to hold 60fps at ultra settings, at least on offline?

Also... Does RAM have any bearing on FPS in games? I've read a few articles where people are saying things like "Kingston Hyper X Impact" or whatever equivalent fast RAM from other brands, isn't a significant difference in terms of gaming and then people who say it is... As you might imagine with the internet being what it is.

Any opinions on whether type of RAM, brand or amount beyond 16gb has a noticeable effect on gaming? Specifically with something like GTA5 in mind?

 

I do not understand what you mean by "offline", but I suggest a 4910MQ chip instead of the 4940MX. Save yourself some money, unless you plan to use the P370EM CPU & slave GPU heatsinks, or CLU as thermal paste on the machine.

RAM does have an impact, and 1600MHz CL11 is terrible RAM. The 1866 CL10 kit should be fine for you however.

I would say 16GB is a great limit. Make sure you get 2 x 8GB sticks as it'll be easier to keep the 1866MHz speed etc with that. Only get 32GB if you're the multitasker of multitaskers and plan to do a lot of specifically RAM-heavy work.

 

Thanks again for the help! Sorry I wasn't clear, by "offline" I meant single player as opposed to GTA Online, the MP mode. I saw a vid on YT which seemed to indicate significant FPS drops when going into MP.

What makes RAM "terrible" or not? Is there something in the stats I should be looking for? I'm really new to the technical side of things, haven't really used a PC for gaming before, only for Graphic Design.

Thanks for the suggestions though as they can save me some money!

 

GTA is a CPU-hog, so it doesn't really matter what you do. I've actually noticed higher FPS online than offline for the most part. Take GTA 5 videos from people with a grain of salt.

RAM has a "speed" (1600MHz, 1866MHz, 2133MHz, etc) and a "cache latency" or "refresh timing" (CL11, CL10, CL12, etc). High speed ram with high latency is arguably worse than low speed RAM with much lower latency. Let's say... 1600MHz CL11 RAM vs 1600MHz CL8 RAM (I am *SEVERELY* dumbing this down for explanation sake):
CL11 ram will refresh every 11ms and within that 11ms time frame it will perform 1,600,000,000 actions.
CL8 ram will refresh every 8ms and within that 8ms time frame it will perform 1,600,000,000 actions.
So over a certain period of time, the ms will add up. By the time CL11 RAM has completed its 3rd refresh cycle (4,800,000,000 actions), the CL8 RAM will be completing its 4th cycle (6,400,000,000 actions).

The default RAM most of these machines are sold with is 1600MHz CL11 RAM. You can choose 1866MHz CL10 RAM from most vendors however, which will be significantly better for your usage purposes, especially with an SLI machine and such a strong CPU. It may not make a difference everywhere, but it can make a difference sometimes, and it is generally nicer to have more overall top-notch components.

If you want a better explanation about RAM timings than I have given, I urge you to read a guide around about them, as I over-simplified to give you a basic understanding of things. I also wish to point out that for most day-to-day tasks (reading e-mail, etc) the RAM won't matter all that much, so you will not have a distinct performance difference you can feel there.

And lastly, I don't think my suggestions are saving you money, but rather making your money do the best it can while you spend JUST a little bit more XD.

 

Thanks for the RAM info. Didn't know what the CL stuff meant

 

There's a lot more to it, but I just gave an incredibly basic idea. Usually it goes like "11-11-11-31" or something. If I was buying RAM right now I wouldn't use that kind of RAM unless it was at 2133MHz.

 

Just checked and i bought 2xCorsair Vengeance DDR3 @ 1600MhzCL9.
Are they nice?

 

Should be just fine. I prefer higher speeds myself, but nothing is wrong with it.