*** Official Clevo P770ZM / Sager NP9772 and P770ZM-G / Sager NP9773 Owner's Lounge ***

Alright cool. Also, how much risk would you say it is to install prema vBios for my 970m? I may not overclock at this time, but would like the advantage of not having my card throttle

 

minimal risk once you follow the instructions. Go for it.

 

It just says "Double click 'ModMe.bat' in Windows and follow on-screen instructions", do I need to uninstall video drivers, etc?

Edit: Ah ohkay just have to disable it in the device manager. When using nvflash I get:




should I accept or no?

 

Well if it says follow instructions... I would assume there are instructions to follow XD. As I don't have a Maxwell video card, I cannot check for you. Mine was done by booting into DOS.

 

More info above ^

 

Nooo I did that, I meant this haha:

 

Seems to be okay. In your "adapter" above it says "10DE, 13D8, 1558, 7700" , so it shouldn't be a huge issue.

Look how close the "current" and "replace with" are. If you're really afraid, ask @Prema

 

@Prema will see it soon enough I suppose, he seems to be in this thread a decent amount, I will wait for his approval lol

Beforehand though:


original score was 882 with a little lower min FPS but I forgot to save it haha

 

That's normal on ALL Clevos.
Reason being the BIOS patching the ID on-the fly and adding its own device ID into RAM.
So NVFLASH is unable to read correct IDs as they actually are in vBIOS.

 

^ See? I figured it was normal.

 

For anyone interested, here is a succinct chart based page by cpu-monkey showing the 4790K vs the 6700K in some benchmarks (gives percentages).

http://www.cpu-monkey.com/en/compare_cpu-intel_core_i7_4790k-411-vs-intel_core_i7_6700k-518

 

Very interesting. It seems as though it will all come down to how the chips will perform when operating at 4.4 on all cores (for me) the bench values are very close so I would think with the increased TdP I'm just going to stay with the 4790K

 

Exactly. My main point here is skylake is all hype, no spark. It is what you usually receive from a die shrink, not a new architecture (and this should have the benefit of both). Now, it supports interesting new standards, if those fit your needs, I'd say get it. But skylake is underwhelming. It will be interesting to see what cannonlake will do using quantum tunneling finfets for lower electron leakage. Currently no fab has said they are adopting the new EUV lithography for the 10nm process (developed in joint ventures for the new lithography standard beginning at 10nm and below). This may be to save on costs as the cost of the new machines will be quite large at first. Also, TSMC and GloFo already have guidance out for producing 10nm chips in 2016 (late) and having them to consumers in 2017 (probably by 2H).

 

8% improvement, nothing to consider for a Haswell user. For an ivy bridge user, okay time to upgrade. Maybe wait for cannonlake. A sandy bridge user? A fine time to upgrade. The performance boost will be closer to 25% from Sandy Bridge. Maybe 30%.

That being said, we knew skylake was crap with IPC... but if it has no iGPU on the desktop side and has higher TDP, then it's definitely a hell of a power hungry chip. That's no joke. I don't know how mobile skylake is going to work out.

 

Another consideration is if they do a p570wm skylake-e replacement next year since it can utilize cannonlake-e. But definitely agree. Still waiting for oc benchmarks in the next couple months. Note - numbers are from an engineering sample, they may not be fully representative of final product.

 

Double post. Phone didn't update that posted. Sorry!

 

Alright cool, just wanted to make sure

I figured you were right, just wanted to be sure.

Anyways, thank you both.

 

Haven't seen this roadmap... Thank you for it! I've been working off of estimates for skylake and cannonlake for awhile now, but now I have graphic guidance... Also says that if Clevo updates the p570wm, it should be Q3 2016. Did you notice the cannonlake mq? Wondering if they screwed up on the heat with skylake and didn't get it back until the die shrink.... Thoughts?

 

Mm not sure haven't given it a full look through yet. @D2 Ultima Ult wanna chime in ? However I'd hope not since Intel supposedly has had test chips since a couple years ago.

 

"The flagship Core i7-5775C processor comes with a core clock of 3.3 GHz and 3.7 GHz turbo frequency. The chip is a quad core model with hyper threading enabled and features 6 MB of L3 cache, DDR3 memory rated at 1600 MHz native support along with the mentioned TDP of 65W."
Also, they're priced at well over $400 for the i7 and over $300 for the i5.
In other words... Intel are being dicks.

"This is for overclockers. It's what the i7-5800MQ would have been, as denoted by the 6MB of L3 cache, low base clock and low turbo clock, which is below even Sandy Bridge's mainstream unlocked chips. But like, we unlocked it for you, so you can overclock this clearly inferior chip with an iris pro GPU for "small form factor PCs" (lolno we just wanted to needlessly overcharge for it) to chase numbers because you probably don't have haswell yet, and look guys, BROADWELL IS SO GOOD, WE ONLY MADE TWO CPUS OUT OF IT!"

This is basically what I'm getting out of that second article. Skylake does not impress me. It's ~8% faster than haswell, which is no upgrade to anyone on haswell. Maybe people on ivy bridge could wait for cannonlake. Sandy bridge upgrades should be powerful enough, but waiting till cannonlake probably would do better for the same price.

No idea what to think about with Intel anymore. They seem to be getting more and more high on their horse and anti-consumer. Even if they care about their business-side first, it's no reason to specifically neuter what's already there for the market, especially without official chipset support (considering P770ZM-type laptops, etc).

 

But! But! But! 65W TDP!

Yeah... I know Apple was all about performance per watt when they switched from PowerPC to x86. And Intel has made some impressive performance per watt improvements since then. But... yeah. Everything you just wrote. Intel's gotten lazy since AMD started making mistakes.

 

This is why I'm praying Zen is good! If it is, we'll have an 8 core with HT and performance between Haswell and skylake (or greater hopefully) (if the estimates of 40% IPC are true (2 node shrinks to Samsung 14nm process and major architecture changes)) at lower prices which can put pressure on Intel.

 

@Prema I installed your vBios and my lappy won't sleep anymore.. it seems like it hibernating instead, and even opening the lid does nothing, I have to click the power button and then it takes a good 5 seconds to get to the login page vs less than 1-2 seconds it used to

Edit: Nevermind, switching the power plan to a different one and back fixed the problem... weird though.
(Yes I made sure all of the settings were on sleep, including in the control panel, advanced power options under sleep, lid close, etc.)

 

Re-install the driver.

 

@Prema See above edit^, should I still reinstall?

 

Probably just ClevoHotKey acting up then...

 

@Prema Do we need hot key?

Also, you should add a pause at the end of nvflash if you could (since you edited it Im assuming you can)
After the flash it shows a bunch of info/instructions and then just closes. All I had time to read was "Computer may need to be restarted"

 

He makes it do what needs to be done

 

yes but when something says follow all the instructions (especially with something like flashing your gpu) then you would generally want to be able to read all of it and not just spit out a **** ton of info and close faster than anyone could read it lol.

I'm not saying he has to add it, just saying it would be beneficial.

 

It's the way nvidia times it...The newer versions close even almost instantly, that's why I keep using the older version for Clevo.
It's not an end-user tool after all and only made for OEMs.

 

Awe damn, so theres no way to add like a system("pause"); at the end?

 

There probably is, but no time to look into it atm...

 

sorry for the quite late reply but didn't checked the forums for quite some time . According to the question "do you need windowed mode in some games?" No at least every game for me i tried used the Overclocked refreshed rate i'm using. First you have to use CRU cause it does something in the registry i guess (EDID Override?) to let windows think your display actually supports the refresh rate. Secondly you have to put the oced refresh rate at first in CRU, my guess is that this causes to use the higher refresh rate to be used as primary refresh rate.

And yeah i can oc with the nVidia drivers about 2hz more too, but it lacks the support that games actually use the higher refresh rate and i don't want to use borderless everytime, with CRU refresh rates are used like i set i to.

This is the setting i'm using atm, i went down from 98% stable 111hz to 105hz, cause i saw a very feint flickering line while browsing the Elite Dangerous forums, i know it's very picky but i want to be sure that it's absolutely perfect.


in nVidia drivers i can go down to Hor 1965 / Ver 1081 Total Pixels, but it seems in CRU you have to use Back porch Hor / Ver 1 as minimum so it setteled to 1977/1089

 

Fair enough, was just giving you a suggestion anyway. I don't really care if you ever get to it, I just think it would be beneficial

 

@ajc9988 how is the new windows build? I've been so busy I haven't had a chance to download it and play around.

 

Smoother. Features are added to edge browser so you now have history and favorites (so I can actually start browsing and playing with it). On the fresh install, it did not recognize or load a chipset driver, airplane driver, and the finger print driver loaded stock will not allow you to use it. I'm running benchmarks now, but they are so far the worst for the os to date. I thought the added functionality caused too much overhead in 10074, and the benchmarks agree. This is worse than benchmark performance in 10074. With that said, I like the added bits to Cortana. If you fresh install from the .esd file, you are on video driver 349.88 again. Also, realbench has an error that carried through to 10122 from 10074. If you are changing values in XTU, then test without a reboot, there is a major hit to photo editing performance. I recreated and sent the info to ms under feedback. I'll have numbers for no win 8 drivers (except chipset, airplane, and fp) up tonight added to the previous pdf I posted. I'll then try the new nvidia driver for comparison tomorrow.

Edit: they added the openCL driver from cuda 7.0 (something I had requested), so no one has to install cuda 7 for certain openCL apps to run.

 

I never had an issue with games using a higher refresh rate with an nvidia custom resolution.

 

I had, at least DX9 games didn't notice the custom refresh rate but DX11 games did. I noticed it at the last LAN Party we had, we played HL² DM, WC3, UT99 and UT2004 and i instantly felt that there is only 60hz working, after i put some games in window mode and started HL² with the -borderless cmd extension the higher refresh rate was used, of course due to the fact i used it in windows. I could even doublecheck it with WoW, running in DX9 Mode the game only noticed 60Hz, in DX11 Mode 111hz were seen. The Day after i fiddeled around with CRU and got to the result that with it the refresh rate was recognized even within DX9 games.

 

Interesting, i'll keep that in mind if others have issues with it.

 

Nvm figured it out...

 

Has anyone tried a custom resolution as well, say 1440p? Im curious how it looks and fares in games.

 

1440p on a 1080p monitor? Unless I'm missing something why would you want to do this? You are still limited to 1080p due to pixel count so switching resolutions in my mind would force the graphics card to work harder to display the same 1080p picture????

 

The scaling would not be the best. It's super sampling or dynamic super scaling from nvidia that does that if nvidia launch it on the mobile side.

 

Because these machines can do it without DSR, and Im curious how it looks and if there are any improvements.

I was supposed to test it last night but I fell asleep was too busy jet-setting on VIP-parties.

 

Thanks @Mr Najsman I get how it works but what I don't understand is the benefit. No matter what you command you still are limited by pixel count

 

Hi all,

Forgive my ignorance. Took a while but I finally finished reading all 234 pages of this thread and a good chunk of the SSD threads in NBR. I've tried asking this question in the Flight Simulation community and I can't seem to get a straight answer with respect to the m.2 drive.

OK, here is my rather long winded question:

First off, this is the configuration that I am waiting for from Mythlogic:

CPU: Intel Core i7-4790K 4.40GHz, 8MB Cache, Quad Core Processor Unlocked Devils Canyon (88W)

m.2 Slots: Samsung 850 EVO 500GB m.2 SATA III 6GB/s Solid State Drive
m.2 Slots: Samsung 512GB SM951 m.2 PCIe x4 Solid Sate Drive (Windows 7 Pro installed here)
Hard Drive: Samsung 1TB 850 PRO Series SATA III 6Gb/s Solid State Drive
Hard Drive: Samsung 1TB 850 PRO Series SATA III 6Gb/s Solid State Drive

This machine is going to run Microsoft Flight Simulator FSX (actually, the Steam version but the engine core is the same) exclusively. I do know that FSX is CPU dependent so I did get the 4790K. I also understand that the real difference between an HDD and an SSD is boot time or application load time. I really don't care about that. I usually start FSX then go get a cup of coffee while the system loads textures. The FSX gurus state that in game FPS does not increase just because it's on an SSD. For me, the goal is not to increase FPS but rather to increase "smoothness".

"Smoothness" is a function of texture loading while the airplane is flying 250kph over scenery files. My biggest problem on my P170EM is "stutters" and mini freezes while the sim tries to load textures on the fly.

Now here is my dilemma. Some of the high resolution airport files must be installed within FSX itself. However, the bulk of my photorealistic scenery files can reside outside of FSX and can simply be pointed to it. That's a good thing because I have over 600GB of scenery files.

Someone here has suggested that I put FSX on the SM951 because of the speeds and I agree but I still would have to put all those scenery files on one of the 850 PRO drives due to size.

Also, for some reason, people in the flight sim community say to keep FSX on a different drive than the OS. I don't know why.

Here is my question: If an SSD really doesn't help the base application run better and if I have to make the application point to the textures on the 850 PRO anyway, then is there really any advantage to installing FSX to the SM951? If not, then I'll put it on one of the 850 PROs and that will give me space headroom for more addons in the future.

I know this is really a long winded question and I'm sorry about the noob questions on SSD's and m.2's but this will be my first machine ever with SSD's. Thank you all in advance.

Bobby

 

I was hoping maybe for some sharper image as the sampling is like an AA. I just tried in Witcher 2 and it works without problem aaand.. I don´t see any difference Apart for lower fps it looks the same to me.

 

@Captain_Bobby
The first reason for putting OS and applications on separate physical devices is to avoid head thrashing. When two or more processes are trying to read or write from different areas of a disk it causes the heads to spend more time seeking than doing I/O. That's thrashing. It's a non-issue for solid state media.

The second reason for putting OS and applications on separate physical devices is bandwidth. If a device can saturate the I/O capacity of the data channel and you have N processes performing sustained reads then each processes will be throttled to approximately channel bandwidth/N less any overhead from thrashing, processor context switching, etc. This is an issue for SSDs which can saturate SATA 3's 6Gb/s bandwidth when dealing with sustained reads or writes.

Here are my suggestions. Don't take these as gospel because I don't fully understand what you are doing and I'm pulling from high-powered compute server setups which isn't your work load. Don't be afraid to try it then rip it apart and try something else to see if it works better, and take notes so that you can go back to a known good configuration if a try doesn't work out.

Put the OS on the slowest drive you have. You're shaving off at most a few second from your boot times between the slowest and fastest drives and this performance can be better used elsewhere. This is probably the M.2 EVO. It's a SATA drive so it's 6Gb/s and it has fewer flash chips than the 1TB models so it doesn't benefit as much from parallel I/O behind the controller. I'm not certain of this; I haven't looked at the high capacity EVOs lately.

Put your load-on-the-fly data on the fastest, lowest latency drive you have. This will minimize the stuttering you get when hitting disk. You can't make it go away completely but you may be able to reduce it below your threshold of awareness. This will be the SM951.

If you have enough system RAM to avoid paging then create a minimally sized page file on the system drive (NT wants this for crash dumps) and be done with it. If you are paging then get more RAM to avoid paging. No matter how fast the drives are, system RAM is significantly faster. If this is not possible then you're going to have problems no matter what you do but you may be able to minimize them by putting the page file on the SM951. This may cause contention with the on-the-fly data but maybe the drive will be fast enough to minimize it. If not then try different drives until you find what works best for you.

Finally, put all of the front-loaded data wherever you like. Performance here is a non-issue; you measure load times in units of "getting a cup of coffee". It's all front-loaded so it won't interfere with any other I/O.

 

@ratinox,

Thank you very much, that is exactly what I needed to learn. That also makes a lot of sense to me. I installed the m.2 EVO drive just because the slot was there and I wanted to fill it while the machine was being built instead of deciding to add something later. Putting the OS there makes perfect sense. I just chatted with the Mythlogic chat person and set it up to install the OS there.

I will use the SM951 strictly for FSX and the textures that were causing me grief on the P170EM. Between that and the 4790K, I believe that will fix my stuttering issues.

I do have 16GB RAM ordered and the FSX gurus advise to lock the page file at 3072MB which seemed to work on my previous systems. I can add 2 more 8GB RAM sticks later if you see value in it.

Anyway, thank you very much for your explanation and advice. I appreciate it and will follow your suggestions.

Bobby