*** Official Clevo X170SM-G/Sager NP9670M Owner's Lounge ***

No, they are totally different systems. The only common element between them is the LGA 10th Gen CPU. Like systems will have similar performance with exceptions relating to artificial impediments like firmware cancer and environmental impediments like thermals. When all factors are equal, which is often not the case, the hardware performance between systems with the same core specs will fall within a small margin of error.

There is the inability to upgrade or repair, but also the cost factor that makes it extremely unattractive unless money is no object. If you keep the system longer than the warranty and the GPU fails (which is the one part most likely to fail; second only to a mechanical HDD, which is too inexpensive to even care about) the problem is you can't just grab a used MXM GPU off of eBay, slap it in and keep going. The motherboards with integrated GPU are almost as much as a new system to replace if it dies outside of warranty. Motherboards with the CPU and GPU both soldered are even more ungodly expensive. And, if my understanding is correct, MSI charges a ludicrous premium for their replacement motherboards. It is a great way to discourage upgrades and repairs and drive the sale of new systems for only a little more. The leftover junk is also great for landfills.

 

Nope.

What is the cost for a new GT76 MB with 2080 Super when it run into hardware heaven?

Not so sure it will be so easy to obtain a used board and to an ok price on internet.

 

That fancy Minecraft screen saver is provided at no extra charge.

 

As weird as this might sound, the magical CCC we all love and worship has rewritten the default xtu profile of my 10600k and I can't find it online.
Thus I would be ever so grateful If anyone here could provide me either screenshots with the default settings or the profile itself, thank you

 

By your own admissions its more space efficient for a unified heatsink. But done right linking the thermal masses does not take much material and improves idle/single load.

 

Thanks for everyone's input.

 

BGA and or proprietary design is disgusting! Even worse if they are tied together... http://forum.notebookreview.com/thr...ll-13-15-and-17.770323/page-468#post-11054940

People have started run more tasks while gaming. Streaming is one of them and will push Cpu harder while you run your Gpu to the limits. Unified heatsink is an disadvantage for many out there. Not a good design with more and more cores on 14nm+++++++ and higher and higher power draw. + more difficult to make perfect to fits on double up of the hardware. And with Unified heatsink we will see a change to the worse... New tech is already here. And more will come...

Currently, many gaming notebooks come with powerful GPUs and CPUs that share the same unified cooling solution, such as a 45W GPU and 80W GPU (total 125W) housed under the same cooler. However, the cooler may only be able to dissipate 90W of heat in aggregate, which leads to constraints on either the CPU or GPU during heavy use.
http://forum.notebookreview.com/thr...laris-navi-gpus.799348/page-750#post-11000607

This tech was in first place born for thin and slim. Not powerbooks. Just look at Applebooks. Without it, they couldn't run the Cpu at more than barely base clock speed in most tasks. Saved by the GPU cooling capacity. Run max load on both cpu and Gpu and it will fail hard.

 

Yeah, CCC is a mess. It totally screws up all of the CPU settings. If you are applying custom settings it seems to do OK with that, but not the defaults. You'd be better off without it if you do not mind the default LED keyboard and light bar scheme. The version for the X170 is about as good as I have ever seen CCC. It's somewhat less buggy than it used to be, but still something you're better off without. This really applies to the software bloat all of the vendors are using now. MSI Dragon Center and Alienware Command Center are trash as well. They're all cancer now that they are UWP filth that is reliant on Micro$lop Store.

Uninstall CCC. Uninstall XTU. Remove the AC and battery, then unplug the CMOS battery. This will clear NVRAM and ensure the real BIOS defaults are applied. Boot Windows normally. Reinstall XTU. Or, better yet, just use ThrottleStop instead. It should work fine with your 10600K. Many of these crappy OEM control center applications are built on XTU and that, in and of itself, tends to be problematic. But, XTU tends to work better without the OEM garbage causing conflicts.

If you feel you need CCC, be sure you install XTU first and save a new profile with the BIOS defaults and name it something other that "Default" such as "Normal" or "BIOS Defaults" and then install CCC. You may find it necessary to apply that profile at every reboot. Otherwise, CCC is going to molest the settings and screw things up automatically every time you boot Windows. Using ThrottleStop instead solves all of those problems.

 

Regarding BGA...

It's repairable. But holy hell.... it sucks (used to reball 360 GPU's).

 

Yes, most things are with sufficient effort. And, that is assuming you can find a place that is able and willing. In practical application, it generally turns out to be not repairable due to the lack of a competent and readily accessible resource.

The logic behind BGA CPU and GPU is no less retarded than the notion of having the engine and transmission being permanently grafted to the chassis of your car, designed to be not-removable, and having no ability to remove either one with ordinary tools. You could use a plasma cutter to separate them for servicing the drivetrain and re-weld them into the chassis when you are done. But, who in their right mind would think that building it that way is a good and acceptable idea. Answer: Nobody.

 

Going by that, cpu and gpu (10700k and 2080super) together you have almost the entire price of a new bga laptop that runs not too much slower if both systems ran by default of course.

For example if you get your 4-5k€ x170 trashed youd have to pay about another 1500€ for the card and cpu plus addionally delidd and LM. Meanwhile laptops starting to get the LM treatment for free at the same price. 10850h with liquid metal plus 2070 costs barely 2k.

If that one fails you can salvage the ssd and save that on the next, possibly RAM as well.

Overall you should run much cheaper while losing not too much performance.

Maybe im wrong, always open for insight.

 

I have never seen an example of a situation where the CPU and GPU both died. Not saying it could not happen, only that I have never seen it and the likelihood is such that it would be some kind of very freakish event. The part that is most likely to expire before you are done using it is the GPU. Second most likely is the motherboard, unless you own an Alienware; in which case the motherboard will likely die based on historical evidence. So, repairs are easier, more economical, and you have an overall superior product with fewer compromises. There is no argument for BGA that can offset those points of consideration. The concept is designed with one purpose in mind: benefitting the seller and putting their interests ahead of the buyer.

 

Doesn't the 2080 super mxm cost a fortune tho? What about socket and or power limitation, how do you stay strong or justify and say okay i want this even tho upgradeability is "limited"? While some kid next year or in 2 posts benchmarks of a awful bga laptop that outruns that machine by margin while being alot cheaper?

 

I've not seen an example of that yet. And, even if we do someday, it won't change the underlying problem of inferior throw-away products becoming the norm because it's better for the interests of the manufacturer, and even more of an issue, public acceptance and embrace of mediocrity. With the latter we can expect the problem to worsen and the bar to be lowered even further.

 

True that,
this can however only be fixed by having entry and midrange non bga laptops where also battery life isn't ass.

 

In my experience, the main issue is finding the right stencils so you can get the balls to line up properly.

And even then, that's assuming that the part ITSELF isn't dead... because if that is, you're completely screwed (see macbooks with dead CPU's being unrepairable).

I do agree though. Like... the laptop I just sold (old Clevo), that had a socketed CPU but BGA GPU. So I could replace and upgrade the CPU if I wanted, but can't do anything about the GPU. The laptop BEFORE that I still have upstairs... same situation (I actually upgraded it from an i5 to an i7).

I can only assume BGA is cheaper than MXM. But you'd expect high end machines to have as much socketed stuff as possible.

 

A shared heatsink does not decrease your total dissipation unless it interferes with the fins and fans. If it's just a thermal bridge then it has no impact on thermal max performance.

 

BGA is much cheaper, its an extra PCB(being made, stuffed with components and then reflow soldered), its an extra connector on the main PCB, an extra assembly step that takes time, all that cost is passed to the end buyer upmarked at at least 2.5x the raw cost.

 

Thanks to your suggestions I did the resets , used throttlestop and gotten a more stable undervolt + overclock out of it.
CCC was and is kinda needed for me, but haven't touched anything related to cpu/ gpu tuning in there ( also xmg has a newer version of it so i downloaded the latest)

 

Awesome. Glad to hear that it helped. ThrottleStop is awesome and @unclewebb is under-appreciated.

What version of CCC did you download? Maybe the newer one is not as buggy. I can test it and see (if it is newer) if it works better.

 

That's fine. There is a place for that. But, not in the space where they are trying to pass off cheaper junk as being something special. Therein lies the problem. They market the cheap stuff as though it is something amazing and beguile those who do not know they are being misled.

 

@Entropytwo You have perfectly valid points.

However, you don't have to go crazy with specs. Usually the i7 + xx70 combo or i5 + xx60 combo gives you the best price/performance ratio. This is true of BGA laptops as well. You get diminishing returns as you keep going higher with specs.

The X170 and laptops like it are meant to last much longer than BGA laptops, which is why their upgradeability is a strong selling point. Normally the part that starts showing it's age first is the GPU, since that's generally the most important component for games. Even if you get a xx80 MXM card to upgrade your laptop with, it's still a better value proposition in the long run than getting 2 BGA laptops within the same timeframe you keep your upgradeable one (assuming all three of these laptops have the processing power you need when you get them). I will admit this does depend on how you specced out your LGA laptop when you got it, because the top end offerings are ludicrously expensive (but this applies to BGA laptops too).

CPUs improve at a pretty slow rate these days, so I don't fancy double spending on a CPU I practically already have. If I buy a new system, everything about it should be significantly better than my last system. If you go the economical route with an upgradeable system, it will be much cheaper in the long run. The xx60 and xx70 MXM cards generally do not cost a fortune, especially if you get them used.

For my personal use case, I would have had to get 3 different BGA laptops over the lifetime of my current laptop had I went the BGA route. The total cost of owning computers over this time period would have totaled up to around $6000 for me (estimated). Instead, I've only spent a little less than $3000 on my current upgradeable system over it's lifetime, including the cost of buying it.

Upgradeable machines can be much cheaper in the long run if you play your cards right. Don't go for the highest end parts because you pay an exorbitant amount of money for a small increase in processing power over the second best offering. This applies to BGA laptops as well. Get an upgradeable system with an i5 or i7 and pair it with a xx60 or xx70 card, then upgrade with a newer xx60 or xx70 card when needed, and you'll have significantly reduced the cost of owning computers during the time you have that laptop. Don't double or triple spend on a CPU and motherboard you already have with BGA machines.

BGA was never meant for performance machines, and it's sad that that's how most performace laptops are nowadays. The build quality on them has taken a huge nose dive too.

 

Most unified heatsink is created to remove less heat than the total heat source. Same also for the unified heatsink in X170.


Nothing is changed vs. Jokebooks todays cooling


10900K PL2 250W. Gpu is 200W. This means an unlocked Cpu adverticed/promoted with Over-clocking technology has to lower the power consumption to avoid boiling. People say the Cpu normally isn't maxed out in gaming. But if you max the cpu on other tasks simultaneous the performance will be hit by not adquate cooling. This is the nasty truth with Unified/shared heatsink.

If they absolutely has to use it due lack of space in the chassis... Then follow hardware specs... 200w for Gpu and 250w for the Cpu (X170). + the normal 10-15% thermal headroom (+495w cooling capacity). But will not happen now or in the future. Because this HS design was invented as a fantastic final solution for thin chassis design with limited inside space.

And with this we will see the screw will be tighten further. More sensors will be welded on the MB calling back to mama to regulate the power draw as mentioned in my previous post. This with help of nice firmware and software. Dell is good at it as you can see ... Others will follow same paths.

 

If the OEMs will not design, engineer and install appropriate cooling solutions, the only other recourse for them is to lock down CPU and GPU power limits. An X170 with a 340w cooling solution gives you only a 140w PL2. I see this behavior in my P870 where the GPU has been artificially throttled to 150w. With that restriction, I can't even get my GPU much above 70 c in Time Spy. i pretty much have wide open CPU power limits but I chose to cap them to keep from thermal throttling.

 

Clevo RTX 2080 mobile MXM is capped at 150w. Clevo saw the light with the Super card and increased it to 200W.

 

https://download.schenker-tech.de/package/xmg-ultra-17-idxul17m20/
The one listed here, well, updated or not, I learned throttlestop and I like it CCC is just for fan profiles if needed and those perstering LEDs which of course are still zombiefied

 

Nice write up, however seeing that amd and nvidia just threw out their new cpu and gpus, you'd buy a pretty badly outdated and overpriced system now.

Clevo is also extremly stubborn regarding high level amd cpus.

My girlfriends bga turd lenovo which is 5 years old now and who has seen endless games didn't snare any operation besides being trashed by windows.. a clean format cured that cancer and now it runs just as new, no repaste at all.

 

Cool. ThrottleStop is amazing. @unclewebb is amazing. We are lucky to have both.

Yes, that looks like an older version of CCC. Older is often better. I wish there was a way to have only the custom fan profile editor as a stand-alone non-UWP software application. My keyboard is solid white by default, which is EXACTLY what I want. I don't need the LED controls. I am living without the fan controls just to not have CCC screwing things up. If I need it to be cooler I just use FN+1 for max fans. The stock fan profile shouldn't be so lazy and slow to react to temperature changes.

Had I been the one responsible for designing it, there would have been a discrete heat sink for each, with a 250-300W capacity for each. Always better to go at least a little bit further than the bare minimum. The chassis would be built around accommodating the needs of a show-stopping thermal solution rather than the thermal solution being constrained by a limitation in chassis dimensions. But, nobody actually builds laptops that way. Everything from the best to the worst is engineered backward instead of forward.

 

To be honest having had the x170 myself there is alot of unused space for a improved sink. You would start to think that they are doing this on purpose. What is the point building the best? Nobody will buy the next newest innovation. It's same as car facelifts, always keep something in your backhand so the fishes come to your bait.

 

Greatness has become relative rather than definitive. The landscape is cluttered with so much utter rubbish that being the best option does not require excellence. Superiority is achieved by being the least broken option available.

 

Love you.

 

I was musing about upgrading my ram, however, between PC-Specialist reporting corsair's 3000mhz being "unstable" and some reports here as well, I was wondering if anyone here was running buy-able and workable 3200mhz ram (2x16 possibly, also would be nice , given the lack of xmp options, if such sticks would actually get to that speed).

 

Doesn't the regular samsung 3200mhz cl 22 work?

 

as long as it does at that speed without problems, if I recall there were some issues explained by some users months ago

 

Im running 3000mhz corsair CL16 right now 32GB no issues 2x16GB

 

@Mr. Fox Or you know it would be good to have chips actually manufacturered on a modern process so 400w plus cooling was not needed

 

Yes that would be very good. Especially if it gave it plenty of extra headroom for more aggressive overclocking.

 

I honestly wish they'd gone AMD instead and paid for the thunderbolt licence.

 

That's another big point on why you should not pay this much for a inferior product, you show them it's fine to put those outaged cpus into modern laptops and charge a fortune. People should swallow it up and rather order million amd cpu plus crap bga and let them suffocate on the intel batches.

I promise if it wasn't for the influence the xmg apex would have a socketed gpu as well.

Let's wait 1 - 2 years for a proper amd nvidia x170.

 

That's fine if you do not enjoy overclocking. If you do, AMD doesn't offer anything worth looking at. I know that I would not be interested. They fail hard at CPU, memory and GPU overclocking.

I agree on the Thunderbolt. I am not into the novelty of the eGPU thing and really see no use for it apart from that. I disable it in the BIOS.

 

Doesn't some recent user made ryzen tool boost all those processors?

 

None AMD powered notebooks come with higher GPU SKUs than 2070 Super graphics. And AMD processors still offers worse gaming experience in at least 1080P. Why would Clevo shoot themselves in the head? X170 is still promoted as Gamingbook. Would it be wise max out the gaming experience at 2070 Super for a gamingbook flagship? Nope. + AMD haven't been known for offering good tech support for the Notebook ODM/OEMs. Tech support/guidelines doesn't come for free. Being a guinea pig in the tech world is a risky project. Even for big companies as MSI etc.

 

I'm not so sure they would have much headroom anyway. Boost eats most of that these days.

 

Nothing that makes a respectable CPU and/or memory overclock that I am aware of. Only relatively modest overclocking potential, but not anything worth writing home about. Rather mundane.

 

Agreed, smart boost is here to stay and getting more and more refined each cycle. The days of monster OC headroom are quickly becoming a distant past.

 

Do you really think it's that, or just that Intel has milked their technology to the point that there is nothing left to extract from it and nothing left for it to give? It would be a shame for the "everyone gets a trophy for participation'" mentality to ruin overclocking as it has literally everything else that used to be good and awesome.

 

I think it is a combination of Intel having extended 14nm even farther than they even originally planned , AMD and the "new" battlefield of smart boosting naturally right outside of redlining.

The Intel of old could leave plenty on the table and sell and resell and re-resell some more just binning up for the most part. Then we would swoop in and even applying an average overclock realize insane gains (50% or more back in the day) and with water even higher. Then we would cackle maniacally while Intel offered "new" versions of the same chips binned up knowing we had already overclocked the originals well past those new bins.

Now? It's a game of inches and AMD has no desire to play the headroom game and they lay it all on the line on their CPUs and GPUs giving you just about everything their chips can clock on air with smart boosts laying bare your individual chips' capability as much as possible. That leaves what? 5%? 10% maybe? for many over clocks after putting in some real work? Meh. Look at Rage Mode and SAM. It banks on a finely tuned smart boosting synergy between the CPU and GPU. I wonder what will be left for overclockers who don't want to go ln2 and stick with air or even water? Yes, you can overclock but there isn't much left. Depending on your perspective, this is a blessing or a curse.

Intel is following suit namely for survival at this point. I don't know if they can or will ever be able to go back to the old model. Maybe if 11th (or 12th) gen is such a revolutionary leap that they leave AMD in the dust again but AMD has become a scary, real force in the market against both Nvidia and Intel and market forces have set a new trend of smart boosting chips as a result.

For Participation Paul or Tammy Trophy, this is a blessing they may not even realize they're experiencing. All they know is they buy a system and it runs like a bat out of hell right out of the gate. They're getting almost everything the chips have to give due to smart boosting with no effort.

How about Ampere?

Look at GN, JayZ2c and Paul. Pulling out a lot of the big guns, the general conclusion is outside of bragging rights, we've reached a point of its not even really worth it to apply extreme daily compatible methods to achieve at best modest frame rate boosts when smart boosting is doing so much for you naturally.

If you're an enthusiast, tinkerer and especially avid overclocker looking for appreciable gains, the last several years have been a sobering event. A kick in the teeth as a large portion of what you used to do to get those large overclocks have been swallowed up by smart boosts not leaving much on the table for you to achieve.

 

Yeah, that really sucks. It is very deflating and eliminates the incentive to spend any money on new tech. What's the point if you're going to end up with the same product as Joe Blow and Jane Doe? There isn't one. Bore me to death. It makes whatever gets pitched as "flagship" totally meh and pointless to purchase.

 

I wouldn't call it inferior. I wouldn't have gone above the 12-core anyways. I would've just liked the option and then tossed a 5000-series in later.

Only just getting into the eGPU thing, but this machine was bought so I can still do my coding and stuff in bed if I need to lie down, AND to replace the huge tower that was in the living room (still have a desktop upstairs though)

As someone who HAS a Ryzen desktop... no... not really. My 2700X performed at 1080p absolutely fine. If there was any performance loss, I didn't notice it on a 1080p144 panel... at all.

As for the 2070 Super, I bought the model of this machine with the 2070 Super... and it's completely fine as far as i'm concerned. I would've been perfectly happy with an 8 or 12 core ryzen and this GPU for my gaming experience, and then just throw the new 5000 chip in there when they're out later this year.

 

It's not just Intel, AMD and Nvidia chips are similar.