*** Official Clevo P157SM-P150SM / Sager NP8250 Owner's Lounge ***

Yes, tried every layout. The problem is present also in bios so it doesn't depend on key layout :/ why would they be incompatible if "p17s0" is written in part number? Don't know these codes
Maybe someone at xoticpc can help understanding

 

Because there were many adjustments to the range over the years over many cpu generations with different ec setups.

 

It works!
Done a cmos reset and reflashed Prema bios.
Ceg hardcustom helped me through that, being very kind

 

Good to hear you sorted your problem

 

So I don't mean to bother you guys, but I am having issues with my Control Center / HotKey software for windows 10.
I installed Windows 10 unto my Sager NP8258 (Clevo P157SM) I bought from xotic about a year and a half ago. Now the issue is basically that the hotkey works for the whole keyboard but I can't change the color of my mouse pad. In Windows 7, I could select the mouse pad and each individual section and make them all different colors. Now I can only do the 3 sections and not the mouse pad.
Anyone have any idea as to what the issue could be? The mouse pad is not even a possible selection of what you can change the color of.
http://imgur.com/awfGiQx

Theres the image of the software and what I see.


EDIT:
Sorry for posting this without actually looking around the software suite. Apparently they made it so that you have to check off a few boxes to enable colored touchpad. If you are like me that has the same issue... Do this:
Right click the circular Control Center v5 icon in your hidden icons tab in the bottom right corner,
Select Keyboard Settings, and check off the colored touchpad.
Thats it

 

I think they did that as most people wanted the touchpad light off.

 

Hi all, Does any one know how is it possible to use maximum capacity of lap top (RAM/CPU/GPU)? actually I have (Sager NP 8265/ clevo p150SM) which has a 24 Gb RAM but when for example I want to render a file in Lumion software, it won't use it's maximum capacity and just use 9 or 10 Gb of that and the render will take about 5-6 hour to be done, but as I see there is other lab top with lower CPU or GPU and also RAM but they can render the same file in a lower time! so I wanted to If there is any way to use the maximum capacity.Any help would be appreciated .

 

That sounds like a library/software issue. As long as Windows x64 sees all 24GiB in Control Panel -> System, it's available.

BTW, it's not a 32-bit version of your Windows software (Win32), correct?

Win32 programs and their libraries, which are shown with "*32" after the process name in Task Manager's "Process" tab on Windows x64, either cannot use more than 2-3GiB of "useful" memory, and, optionally, can only usually page in another 8-12GiB of memory, maximum. Now the OS itself, which using full 64-bit, might use several more GiB for buffering/caching. But that's about all it'll use.

Longer story ...

Spoiler

Windows x64 runs in an AMD64 (aka Intel IA-32e) mode known as "long mode" (which has to do with how the address registers are used) which uses 48-bit "flat" (effective, from a platform level, not 64-bit) addressing (with required 52-bit paging extensions). But this requires the program to be actually built with Win64 libraries, which can use the full 48-bit "flat" addressing mode.

If the program does not, Win32 on Win64 (WOW) is used to execute the Win32 program on Win64, including loading the appropriate Win32 libraries, with the same memory model (addressing, etc...).

WOW provides the legacy "segmenting mode" (again, how the address registers work -- and part of the reason why AMD64 "long mode" uses 48-bit, not 64-bit, so it can run legacy libraries/programs at the same time) which rationalizes them to 32-bit addressing (with optional 36-bit paging extensions). Windows x64 actually ships a lot of legacy Win32 libraries as well. In fact, early on it didn't ship a lot of Win64 libraries, which caused a lot of memory constrains and issues (I won't go into details, but let's just say graphics and high-end gaming development houses went Linux/x86-64 in the mid '00s, because Linux has always been "64-bit clean" thanx to Alpha).

So ... what does that mean?

It means a 32-bit program can only address 4GiB RAM (the program usually only 2-3GiB "user space" RAM, with kernel, memory mapped I/O, etc...) directly (aka "low memory"), and has to page in any more (if it's built to use the 36-bit paging, the "high memory"). Between 8-16GiB RAM, paging starts "breaking down" where you get a point of "diminishing returns." I.e., the kernel in "low memory" has to keep track of pages in "high memory" that is in use by the program, and this quickly adds up beyond 16GiB of "high memory" (which can use 0.5-1GiB alone, depending on OS, page size, etc...).

So ... again, get a native Win64 program, if at all possible. Most Windows software in the '10s offer native 64-bit versions, precisely for performance considerations, especially as Microsoft has made many Win64 libraries available these days (and quite peer to Multilib Linux/x86-64 these days, sans many MS Office components and some select C libraries still being heavily Win32/x86-only).

 

More RAM does not mean a job will finish faster, that's up to the CPU/Hard Drive/GPU/etc, it just means the job has more access to RAM IF it needs it, which it clearly doesn't. You are not going to get faster rendering times once you reach a certain amount of RAM, it will just plateau. The same is not true for something like CPU/GPU where the actual rendering comes from.

 

Thank you so much for your useful explanations,I checked all my software and my Windows too,all of them are x64, and still cannot understand where the problem could be from! for ex. I open 4-5 program simultaneously and try to work with them, the system not use more than 40% of it's capacity but becomes very slow! So do you know if the problem cause from other source? (I attached here also my system character)

 

Preface: Responding to multiple posts, but using "Spoiler" tags to keep the length down for those that commonly complain about such.

Er, um ... that seems a bit conflicting.

E.g., DRAM still kicks the pants off of NAND (SSD), especially commit/write speed (NAND heavily relies in DRAM for buffering writes), and latency is not comparable either (although NAND destroys platter's latency 2 full orders of magnitude).

Spoiler

Heck, the whole reason we're getting away from ATA/AHCI commanding, and to NVMe, is because NAND is being held back by it (even when over a native PCIe bus).

Of course, NVMe really is a Windows 10 thing, with some issues even on 8.1 (let alone 8), but it's already heavily utilized in the Linux world ... including some "vertical-hardware" OEMs offering seriously horrendous markup.

So if you can avoid paging, and keep things in RAM, it will vastly improve your performance -- sometimes 50x over. Of course, your application has to be able to take advantage of it.

Spoiler

If it's avoiding paging heap -- between "low memory" and "high memory" -- than that can hurt performance even 2-3x. Memory models make all the difference, especially in the age of x86-64 "long mode" versus traditional x86 segmented paging.

That is a "gamer's" view. That is wholly inapplicable to many "professional graphics" applications.

Spoiler

There are desktop applications that can consume 32-64GiB RAM, and have been for a full decade-plus. These are using huge heaps of allocations that could never run at 30-60fps in a gaming application, but RAM makes the difference between 50-100ms response, and >1s response.

In the '00s, this was heavily the Linux/x86-64 world, but in the '10s, Windows x64 has been getting more and more "64-bit clean" with the required libraries.

Again, you're thinking "gaming applications." In those cases, the vector SIMD is the major limitation, let alone to its dual-ported (Video RAM) framebuffer.

Spoiler

Anything outside of that is just to hold secondary textual data, along with other operations that are not related to the rendering. Windows x64 is only get some of the first Win64 libraries to do this, hence why most "games" never needed more than 2GiB userspace (plus maybe 2-4GiB for buffer/cache), and we're not just getting to 8GiB+ requirements.

Professional graphics are entirely different! We don't need 20ms response times for end-to-end, but 100ms+.

We were running with 3GiB back in the late '90s, and definitely 16GiB+ on Linux/x86-64 platforms in the mid '00s. We were, and still are, also doing things at much higher precision than gaming. Although gaming is finally getting double-precision more and more, thanx to the advanced SIMDs and GPUs.

E.g., marketing-wise, both AMD and nVidia now market their GFLOPS in 2 ways, no longer just as assumed single precision, but now also double as its own benchmark.

Spoiler

One of the best chapters I ever read on x86-64 came from the AMD pre-release Programming Manuals -- very first Chapter 1 explained all of the issues, and how AMD addressed them.

The 48-bit "Long Mode," and its limitations and requirements in the memory model, while being 100% backward compatible (including to the all-important 36-bit PAE of the i686, by adding another layer to 52-bit), taught me 90% of what I needed to know.

That was 12 years ago.

So no "*32" next to your program's name under "Processes" in "Task Manager" under Windows x64?

Also ... what programs, if I might ask? There could be tunables or other considerations. Out-of-the-box, Windows is tuned very, very conservatively, which is expected for legacy applications.

Spoiler

After all, the #1 Windows application is MS Office, and it's still very much Win32. The Win64 releases still have limitations, but they are slowly porting it, and rewriting (or writing wrappers around) Win32/x86-only C libraries, especially those that are used in .NET objects.

Although some will be emulated on a bytecode run-time (long story, don't get me started).

Wait?!

Are you talking a single application's benchmark? Or a combination of programs?

Context switching on its own can have an impact, more than memory can accommodate on its own. And, again, tunables. Windows, by default, isn't very aggressive in using memory for buffer/cache, and outside of MS Office and some of Microsoft's own applications, prefetch is also very conservative.

Spoiler

That's why it's not uncommon for Windows desktops to make very poor usage of more than 8-16GiB RAM, unless the software installer for various programs also offers to tune your Windows x64 installation too.

E.g., there's a reason why you don't usually need to tune Windows to run on VMware ESX and get a high ratio, but the first thing VMware has one do to Linux is tune its VM (that's kernel virtual memory/paging, not virtual machine) to be far more aggressive in reclaim and not being so liberal in using all memory.

Most high-end x64 applications for Windows try to use all the memory they can, and even change how Windows x64 is tuned -- hence why you get the UAC "prompt" at times to do privileged operations and/or the programs require you to run with "Administrator" or at least, what used to be called, "Power User" privileges.

But if you're talking 4-5 concurrent applications, then you're talking tuning beyond what a single program might do for Windows for itself.

 

Clearly he's reached a plateau if his program does not want to use more than 9GB - that has nothing to do with the amount of RAM he has, or his system, or his configuration, and everything to do with the program he is using. I am sorry you are having problems understanding this. Everything you said is true - but in this case, it's completely irrelevant and unnecessary because clearly he is at a plateau if his program does not want to use any more. It doesn't matter if he has 24GB, 32GB, 64GB or even 128GB of RAM. If his program does not want to use more than 9GB of RAM, the extra RAM will just be wasted and unnecessary. THAT is the point I am trying to make here.

You can try and break it down to it's most abstract layer but that still does not change what I am saying here.

 

Any chance of finding some cool Steelseries keyboards for this grandma?

I have one SM-A which I currently use it for travelling and stuff, and would like to have a new keyboard.

 

They don't share the same electrical layouts so can't be moved over.

 

I've used SteelSeries keyboards on MSI laptops in the past, and honestly - I prefer Clevo keyboards. This is coming from someone who owns like 4-5 different mechanical keyboards so I'm a pretty good judge of keyboards.

 

Thanks! Anyway, was just looking at the N150 keyboard. Almost the same as this old ones some exceptions in the F compartment and keypad.

Meh

 

It's not worth the risk really.

 

P157SM GTX880M

Any method to lower the temps of the GPU/CPU and noise? Playing games the temps get to 80-90+ degrees and the fans at full speed. The games aren't that demanding either.

Any custom BIOs or automatic fan controllers available?

 

You can make sure the heatsinks are clean. If they are you can try a re-paste with a high quality thermal compound (like IC diamond) and even experiment with things like replacing the thermal pads to get the best contact.

Also when in use prop the back of the machine up an inch or two to get the best airflow.

 

Was going to post the same thing. GTX970m is WAAAY cooler than this 880m. Seeing 80-90 degrees

Heatsinks clean, new thermal compound.

 

If you can afford the upgrade to a 970M that is also an option if you feel secure about performing it.

 

I am the only person in the world to downgrade to 880m. My original configuration was with 970m. 72% ASIC vs 83% for this 880m, and was perfect for everyday gaming and CADing.

Now, I am looking to find some tried-on undervolt settings so I can run quiet.

 

ASIC readings can only really be compared between identical chips, even then it is a little dubious.

 

I even tried in 15 minutes of gaming to run -.135 GPU voltage and 900 Core Clock. Still low 90s.

 

Yes the 880M is running at the limit of the core pretty much.

 

Not a good one tho`

 

What's the performance cap reason given?

 

Thermal and Power

 

In the meantime I tried running some 2133MHz DDR3 to 2400. Would not post.

Oh. And I dont feel the Intel MX processors are a good match for the SM versions. Cooling cannot cope with high overclocks.

4900MQ series are perfect. Can go to 4GHz, temperatures in XTU Stress don't exeed 83. MX will throttle.

Stock voltages for both.

 

The stock voltage on the MX can be a little higher and they do need a perfect paste job, the 17 inch versions come with a copper heatsink, they fit in the 15 inch if you can get one and it may help.

 

Already got the copper one. I tried with -40mV. When it goes to full 4GHz on all 4 cores, game over.

 

How did you guys manage to overclock a 4900mq? Just raising boost clocks?

 

Changing multiplier in XTU and adjusting the Boost power and time window. Changing the reference clock grants you a BSOD.

Non-MX processors have a limited time window for example (96 seconds) vs indefinitely on a MX. And are capped at +10 TDP (57W) max.

The P37*SM series have a slightly beefier power stage but also lack adequate cooling for the MX.

I can say the best in terms of Voltage regulation are the ZM series. Identical to the DM series, but Haswell has it's own Voltage regulator inside. Skylake doesn't.

As for the 4940MX I might say it is a downclocked 4790K.

 

Then again Haswell had to cool that part of the power circuitry on the hot die which led to its own issues.

 

Is it possible to unlock multiplier on 4800mq?? Didn't know about that! I use xtu but didn't see such an option. Only boost multipliers

 

Yes, that is true, but compare the power circuitry on a ATX motherboard to the one on a notebook.

Only boost multipliers are unlocked.

 

With premas bios all multipliers even reference clock are unlocked and can be changed, i got my 4710mq to 3,6ghz 4 core stable

 

You cant change the "Non Turbo Boost" multiplier to a higher value.

Just the Turbo ones.

Changing reference clock higher than 100.1 get's you nothing apart from a crash.

 

Reference clock changes are always very finicky, not something to be done by beginners really.

 

As a pure curiosity, a 680m will work on this platform?

 

Yes it should.

 

Update on the 880m matter:

It seems there is uneven pressure on the heatsink due to a number of factors. Used some longer screws and did a repaste now I get to max 80 in games, which is nice.

The original Maxwell screws are a size smaller for the 880m. Used the ones from my Alienware, but somehow there was a gap. Put some plastic shims from the CPU screws fixed the problem a bit. Now I`ve used a combination of shims, long and short fasteners and voila. Works as it should.

 

Typically the screws/mounting have been different between clevo/sager and alienware.

 

Hi all. This is the first time encountering this issue after using my machine for about a year. It's a Clevo based laptop with an Intel i7-4710MQ 2.5Ghz - 3.5Ghz Turbo Processor and a nVidia GTX970 GPU and running windows 8.1 64bit.

Desktop was running fine last week and I was playing games for about an hour or so. When I quit the game, I was greeted by artifacts in the windows desktop. I restarted my game again and I didn't see any artifacts in the game at all.

I believe the issue has got something to do with the integrated Intel HD4600 GPU, since no artifacts are present in the game (Grand Theft Auto 5), which uses the nVidia GPU. Rebooting the laptop didn't help, nor did shutting down the laptop and trying it again the next day.

I rebooted my machine into safemode and there were no artifacts to be seen in the Windows desktop, which leads me to think that drivers are the culprit. The same happens (artifacts disappear) when I disable the Intel HD 4600 display adapter.

I've tried installing the original drivers from the supplied CD and I'm still getting artifacts. I've also done a fresh install of Windows 8.1 OS, which didn't fix the problem.

I've no other ideas on how to fix the issue.

Anyone with suggestions? Thanks in advance.

 

You can disable the HD4600 via task manager. See how it works.

Are you stock? Freq/Voltages for the CPU/IGP?


-----------

Managed to install the 680m modded inf file with some help.

But now, there are some things I cannot understand. Default Clock is 719, but when I am gaming, I do not know if the GPU is active, as GPUz is reporting spikes of 135MHz on the core.

Another bit, the Subvendor appears Clevo, but the card has only 2GB of VRam? The VBios is a Dell one.

Oh, and before installing the drivers GPUz was reporting 4GB of VRam. How is that possible?

---------------

After some testing, the card get's ON and OFF

Seems that there is a compatibility issue with the SM-A. By disabling the IGP I noticed that there is a terrible performance deficit (1-2 FPS), thus the GPU was working.

This is a DELL card and it seems no love with Clevo.

 

Note the Intel GPU is always used in this model, even in 3d applications the frame buffer of the IGP in system memory is being used to send information then on to the display.

Investigate the system memory.

 

If I disable the HD4600, the artifacts disappear, but the nVidia GPU won't be recognised after that. I get the error "you are not currently using a display attached to a nvidia gpu".

edit: yes, everything is stock.

You say investigate the system memory. I would assume you're referring to the mainboard RAM rather than the video RAM?

 

Yes the system ram, not the video ram.

 

Any input on why the 680m appears as Clevo subsys (1558) but the Bios is from a Dell and the 2 vs 4 GB mixup?

 

Yup. I've performed the built-in Windows Memory Diagnostic test. Standard test, 2 passes. No errors detected.