New Clevos with Max-Q?

Hmm, what was I referenced for ? I feel important. If is weight, then it's the official weight advertised by Asus, I know reviews say something else, but on their website that's the weight, we only take official producer data into account and revert to reviewer data when official data is missing.

 

I don't have it yet installed on the laptop with the pascals (because INF hack driver install)... plus, SLI.

 

You have also the other options... AcerBook Predator Triton 700 with even more crippled solution. Offer thin and flimsy due portability, but use them on the go is almost impossible. Expect more of similar Crippled design. One more crippled than the other. Nice with maximum portability!! $130 dollars JokeBook's is much better choice on the road. And you pay correct price as well.

 

@D2 Ultima, is your review up yet, or is the one on NBC the final one?

 

Search for the title "maximum efficiency, minimum performance?" and you should find it. Not 100% a-ok yet but should be close enough. Warning: article at least 50% longer than original non-watered down version.

Sent from my OnePlus 1 using a coconut

 

this?

https://www.notebookcheck.net/Opini...-efficiency-minimum-performance.232038.0.html

 

Yes

Sent from my OnePlus 1 using a coconut

 

The first Max-Q Gaming Notebook: ASUS Zephyrus GX501

 

@hmscott Julia is hot

On a more serious note, there's a very simple TL;DR. It depends on what one prioritises. If they want maximum performance at minimum weight, then they will have to forgo value, and get Max-Q. If they want maximum value, and minimum weight, they will have to forgo performance, and get something cheaper like the Inspiron. If they want maximum performance and maximum value, they will have to forgo weight, and buy at least the P650HS.

The choose-two-out-of-three triangle for laptops has never changed, and Max-Q won't suddenly make the magic happen.

 

You can get maximum performance at minimum weight for good value with the p950. For 1650 you can get p950h4 1070mq + 16gb + 500gb ssd + 1tb hhd. Good deal from imho

And even if 1070mq is close to 1060 and you want save more money you can get the normal 1060 for 200$ cheaper

 

Linus did a video too with some tests compared to a desktop and mobile reference.

 

Honestly that's what made me so annoyed. Before, you didn't put a 780M or a 880M into something 0.8" thick and be like "heyo, here you go!" at full price and people bought it (nothing razer makes counts, they can't even get a 1060N to run normally). You took a weaker GPU, usually a 960M. Maybe a 970M if lucky. The Aorus with the 980M I think was the thinnest and since their cooling design was made for (and failed to) handle 970M SLI when they switched to single card it worked fairly well. But even then it wasn't THAT small. Gigabyte had sub-1" 980M afaik and it took a lot of work to make them not overheat from what I remember, and that was in the UK where it's relatively cool too, remembering Cakefish.

I'm not saying if the tech can exist don't do it, but to lower card performance like that just... to me, was just a real slap in the face. I'm glad I got the opportunity to show that a 1070N could have done the job without losing performance in a public place that OEMs and Nvidia read and take note of.

There's only so far they can go with their pricing and marketing.

 

I just wish it had a 120Hz full hd screen to better utilize the 1070MQ....otherwise the 1060 is plenty.

 

Exactly. Why the hell would I pay $200 more for the same performance? I'd just buy the P950HP and be done, perhaps OC it a little bit, rather than stuff a 1070MQ and not get 1070 performance, while paying way too much for it.

 

The 1070MQ does seem to have a better performance increase over the 1060 compared to the 1080MQ over the 1070, but yeah I agree charging the same for the 1070MQ as the 1070 is too much.

Sent from my SM-G930F using Tapatalk

 

Because most people don't overclock for one and they don't want the increased heat and noise that brings.

 

Yeah, his G701VI and AW 15 scores were too low. I think he was running in non-Sport mode for the G701VI - I've seen much higher results from a G701VI, and using an older AW 15, not the new 1070 OC out of the box model that started shipping with the Max-Q 1080 to match it's performance.

Those weren't good comparisons unfortunately, but a lot of people will think they are

Asus Max-Q 1080 Laptop Review by LinusTechTips
http://forum.notebookreview.com/threads/asus-max-q-1080-laptop-review-by-linustechtips.806763/

 

Indeed. I noted earlier in the thread that the 1070MQ performs significantly better than the regular 1060 in the MSI GS73 shell (cracks 15000 GPU score in Firestrike from memory). It's faster than all but the highest desktop 1060 OC while operating at the 40dBa limitation of Max-Q which "thin" 1060 machines (including the GS73 7RE) cannot do. The 1070MQ really seems like the one where they actually got the "point" right.

 

For about $1600, I can purchase a notebook with a full GTX 1070, achieving ~20000 on Fire Strike, with plenty more expansion options, more robust cooling, power delivery, etc. The weight of the notebook plus power adaptor is something a lot of people overlook, and the P650HS is a very thin and light machine already. If I wanted to reduce temperatures and hence noise even more, I'd undervolt the 7700HQ and the 1070, while maintaining the same (or even higher) clocks, because the power and thermal limits would not even be reached yet.

The fact remains: Why would I spend more for less? Why is the so-called improved voltage delivery, reduced voltages, and quieter fan curves, not a thing for regular non Max-Q notebooks, but restricted to Max-Q? I mean, you see it with turbojet engines... The higher-performing ones are more efficient than the ones lower on the performance scale.

 

Well it really depends. Lets say you want a MSI GS63 laptop. Maybe you like the look of it, or maybe you really need that Thunderbolt3 (none of the thin Clevos have TB3 which blows my mind given they already have the Alpine Ridge chip... ). Before you topped out at a GTX1060 on the GS series. Now you can pay more and get the 1070MQ which as noted, ticks all the boxes. This seems fine to me as there's no full 1070 version of it already. You're paying more for more in that case. That is, if "more" for you is less weight, less noise and an small boost over the 1060. To some people this is the case.

To some people it also isn't about the money. In my opinion, somebody who can afford a GS63/Aorus/Blade with 1070MQ probably doesn't have the same kind of perf/$ battles inside of their heads that most of us mere plebs have to. They've probably also got a monster desktop rig in the background somewhere. Price-wise a GS63 7RG costs $3400 and a roughly equivalent spec P650HS-G costs around $3000 (prices for Australia and we always get screwed equally on prices).

The P650HS is a great machine for it's size, but it's still a long way away from the likes of the GS63, AorusX5 or the P950 which all sit just shy of 2KG. Most people forget the P650HS uses the much fatter bottom plate for the larger copper heatsinks which puts it at 28.8mm and 2.65kg. For comparison the GS63 7RG (1070MQ) is 17.7mm and 1.8KG.

To put that in perspective, the P650HS is closer to the P750DM in size and weight than the GS63!

That 850g difference can be rather significant if you have a habit of "one-handing" your laptop or if you're on weight-restricted flights. For example, for weekend flights I can just get away with my P650HP as carry-on without the need for checked baggage. If I wanted to carry a bigger camera or lenses I'd be stuffed. Alternatively, I could fit an entire DJI Mavic drone (thinking about getting one) in that weight differential if I had a GS63.

 

But the the Aorus X5 already has a full-fledged 1070 in it, and manages to cool it well, to boot. And yet it is still the same price as the GS63VR with the 1070 Max-Q, where both have approximately the same dimensions, but the Aorus X5 significantly outperforms the GS63VR.

Why neuter performance further by making the 1070 'Max-Q', and then asking more for it? I concede that the GS63VR has got Thunderbolt 3, but to me, and most other rational people, it isn't worth a $400 premium—nearly A$530 (roughly equal in S$, too). For which I may as well jack up my tangible performance, or disk drive capacity.

I don't disagree with the concept of Max-Q. In fact I think nVidia ought to implement it universally as an optional feature, built into the hardware, firmware and drivers as an option for users to choose and customise. But what I do disagree with, is the deceptive marketing, the ridiculous prices and the expectations that nVidia sets for the average consumer, only to lead them to disappointment.

 

Thunderbolt 3, as-is, is pretty useless all around. I cannot think of a single device (sans eGPU) that does not have a non-thunderbolt alternative. Also, the HM175 chipset has less lanes than the CM238, I think the GS63VR has less I/O overall and can afford TB3 on it:
GS63VR = 3*USB 3, 1*USB 2, TB3/typeC, 1*SATA III, 1*M.2 NVMe
P650HS = 3*USB 3, 2*USB 3 gen 2 (10Gb/s), 2*SATA III, 2*M.2 SATA or 1*M.2 NVMe

The layout of the PCI/e lanes in the chipset is different. Even the P870KM1 shares bandwidth if you use two TB3 at once; there isn't enough lanes in the system.

 

Depends entirely on where you are. The Aorus x5 is more expensive where I am. The new Aorus x5v7 will have TB3, run a full 1070 and cost $3599 here ($200 more than the GS63 7RG).
At the same time they're releasing an Aorus x5 MD which is identical except it'll have a 1080Max-Q and only the 4K panel (no price atm). Even has the same dimensions and weight so I'm guessing the cooling system will be the same.

As far as cooling with the AorusX5, it sure as hell doesn't do it quietly though. The only way it looks like the temperature is ok is because the fan is absolutely screaming. It registers 53dBA in unigine valley test (ie gaming-like with average CPU load, not prime95 like the other tests)!! dBA being logarithmic means that is more than 4X increase in noise over a 40dBA system (each 6dB increase equates to roughly double the perceptible noise level). To put that in perspective, my P650 using the FN+1 shortcut tops out at around 50dBA which in real terms, can be heard through my headphones WITH music playing through them.

Hopefully NBC can get their hands on both the x5v7 AND x5-MD above as they should provide a great A/B test platform for 1070 vs 1080MQ. Identical hardware and power supply, one with 1080MQ and noise limits etc, the other with a full 1070 at full screaming fan mode.

Indeed, a "quiet mode" or similar has been something done in software for years now, although mostly with bad results as it always feels tacked on by the manufacturers. ie, Clevo has a quiet mode but it sucks because it's controlled by ClevoCC and the fan table only has 3 or 4 speeds. It might just be a case where if Nvidia don't set some sort of rules the manufacturers all just muddle around getting nowhere.

Thunderbolt could pick up speed with audio engineers as a high end Firewire replacement (if Intel could sort their collective butts out about opening it up with add-on cards etc). Most engineers favoured Firewire originally because it being a parallel interface afforded it lower latency for multiple in/out streams compared to USB2.0, regardless of the fact USB had similar bandwidth. Firewire (or TB for that matter) can also run with lower CPU load at smaller buffer sizes (which equates to lower latency) since they can use DMA. I'm actually considering getting a basic TB3 box for my desktop just so I can stick my old firewire card in it to connect my old interface up. My "new" USB interface can only get down to around 12ms safely whereas my old firewire interface can run around 4ms.

 

I mean, that all sounds well and good and all, but realistically, it doesn't matter. Add-in cards are still limited by (and share I/O bandwidth with) the DMI, which is limited to a PCI/e 3.0 x4 connection from Skylake onward. Even if you wanted to run an add-in card for TB3, if you had a M.2 NVMe SSD in there, they'd fight for bandwidth should you be using sequential transfer data on them.

As I said, I cannot realistically see how Thunderbolt 3 DIRECTLY improves a notebook over a good compliment of USB ports. eGPUs are worthless unless you shove a 1080Ti in one and even then it barely succeeds a 1060N with a voltage curve frequency adjustment downwards. If you had a 1050/RX 460 notebook or something, it might make more sense, but we're not considering those kinds of notebooks in this particular discussion.

Either way we take it, Apple forced thunderbolt and everybody jumped at it on promises of great things, and thus far there's a huge race where everybody wants it but nobody really wants to do anything with it. I keep asking people why they want TB3 and every answer has an alternative solution that doesn't require it, or ends up being an eGPU (which as we talked about it is pretty worthless).

 

This is actually a limitation of the Alpine Ridge chip, not the PCH.

Since it can have either 1 or 2 external ports connected to it, you would theoretically need 8x PCIE lanes back to the PCH if you wanted to fully feed both ports. The HM175 and similar could theoretically do this is you dropped some SATA/USB channels.

The only exception to this would be if you have 1 port as TB3/USB and 1 port as pure DisplayPort mux, as the DP port draws bandwidth with it's own link directly to the GPU and thus, doesn't share resources.

Apple actually had very little to do with it this time. I think the main problem is Intel failed to open the standard early enough and simultaneously failed to make it ubiquitous. Part of creating a new standard is you end up with "N+1" standards. The only way to get around that in the IT world is either A) force it, B) flood the market with it by making it cheap. They failed on both accounts.

The difference here, is Intel controls Thunderbolt3 directly and could very easily have built it straight into EVERY Intel platform since Skylake. You can be damn sure there would be more USB 3.1 and TB3 devices if they just built it straight into the platform itself. Hell, they could've taken it a step forward and enforced it into the Ultrabook standard (which all use SoC designs) so that TB3 chargers and docks became the norm for them.

 

Technically it isn't really Thunderbolt that's the true be-all end-all, it's actually the USB-C port and plug, the physical standard. USB-C supports the three fastest, transfer standards available—USB 3.1v2 (10 Gbps), Thunderbolt 3 (40 Gbps), and even DisplayPort v1.4 (7680 × 4320 pixels, 30 bits per pixel, 60 Hz, effective bitrate ≈ 60 Gbps). All in one plug. The absolute cherry on the cake, is that USB-C is minute compared to any other connector there is, from USB-A, (Mini) DisplayPort, HDMI, Ethernet, etc. By 2020 I fully expect USB-C to come standard with all new add-on storage like USB drives or smartphones, which would then have bidirectional USB-C.

I look forward to the day when we really have just one connector for every single standard available. But then Apple will be like 'no, we want to be special'.

 

Even then, it's still a bandwidth issue. KM1 has 24PCI/e lanes. 3 x M.2 NVMe drives, 12 lanes left. One TB3 40Gb/s is 8 lanes left. This goes to SATA/USB/etc. If you cut another and used two full-speed TB3 without bandwidth-sharing with a different controller, how'd you run wifi/ethernet/sata/USB/etc. I mean, you theoretically COULD, but it'd be really silly.

Apple was pushing Thunderbolt for a good while. Some of their screens afaik only worked via thunderbolt 2, because they said so, so screw off etc. Apple most certainly started it, and removed a lot of ports in their newer devices for Thunderbolt to replace it, so that people have to buy their dongles to use their old/existing equipment with the new standard, that has nothing. I think @Galm it was who put the questions to an Apple employee in like a geek bar or whatever they're called and they couldn't really justify TB in their gear now.

 

My point was it's not really the Alpine Ridge controller or Thunderbolt 3's fault that Intel's DMI channel is too small AND they don't have enough lanes on the PCH.

For most non-SLI laptops (ie everything but the P870), you've also got the potential to cut 8x lanes off the CPU itself. Almost all single-GPU systems have all 16x PCIE lanes going to the GPU which is pretty much entirely unnecessary for the current generation of GPUs.

The other thing, is the Alpine Ridge controller can be used as a shared USB 3.1 controller with those extra lanes. ie, you could run the majority of type-A USB ports off the Alpine Ridge chip instead of directly off the PCH which reduces lane wastage since you're unlikely to have 2 TB3 ports running full 40gbps given the DMI limitation.

For Thunderbolt 1+2 you can definitely pin it on Apple for being an objective failure as it was always built to be an Apple exclusive. So therefore it was on Apple to push the standard out. Thunderbolt3 however, was always designed to work on PCs and was built to integrate into the Skylake PCH from the start. It's entirely on Intel that it fell over. That being said, TB3 will be royalty free in 2018 which *may* allow it to see better adoption. Royalties were probably half the issue depending on how high they are. The only reason DisplayPort has had such a good run recently is because it is royalty free compared to HDMI.

 

Fair enough.

You can cut 12 out, to be honest. PCI/e 3.0 x4 is more than enough for even a Titan X Pascal. @tgipier did testing with his at PCI/e 1.1 x16, which is PCI/e 2.0 x8, which is PCI/e 3.0 x4 (effectively, anyway).

Fair enough, though neither intel nor AMD can run a single TB3 port at 40Gbps... the DMI limit on Intel *AND* the chipset <--> CPU connection on AMD's Ryzen (the name I forget) are both PCI/e 3.0 x4. The maximum is 32Gbps, or a bit under 4GB/s. And once again it's eaten by any NVMe drives, and your wifi card/ethernet card/capture cards/whatever-in-the-world-you-have-attached-to-any-expansion-port. You could give it a direct-to-CPU connection like Apple did on one of their newest macbooks, but let's be honest here... that's an entire x8 of lanes for ONE port, with 3 of those lanes purely wasted.

"May". HA. Intel doesn't understand any longer how to compete. That's why Skylake-E is such a disaster, and the naming scheme is so idiotic. Maybe they'll learn by 2018 though. But really I'll blame AMD more than intel for this, and laugh as intel drives itself into a corner once AMD learns how to bite back.

 

That's just PCIe 3.0, IIRC. AMD hasn't exactly named that connection. There is Infinity Fabric, which is a CCX-CCX and die-die connection, and seems like a fairly big deal...

 

Yeah I know infinity fabric well. 22GB/s (or was it per cycle? probably per cycle) and ticks at RAM actual speed. Any program with a thread in one CCX that's dependent on a thread in another CCX needs to transfer by the relatively slow Crossbar on Infinity Fabric, and the faster that thing gets the data across the lower the in-processor latency, and thus the faster the CPU performs. Think of it like IPC directly scales with RAM speed (a plateau near 3200MHz) for anything that isn't incredibly parallel.

 

HID has 120hz for p950 but not a better deal than gentech pc

 

personally as a user who is looking for light high performance laptop, i am waiting for the reviews. i just pray day and night the performance difference is minimal (10%). if it happened then it'd be one perfect fit deal. maybe overclock it a little to reach normal 1070 performance.

 

They basically said they were pushing the envelope, but had no answers when I asked ok so why couldnt you just put on a couple and also give regular ports?

 

Overclocking a MQ will never achieve perf of a normal, undervolted notebook card. The undervolted part is key there

Sent from my OnePlus 1 using a coconut

 

As above, and for $1600 I could build a 1080/1080ti desktop.

But form factor requirements are what give notebooks the space and the larger machines are still there for those who don't mind.

 

Wrong thread.

 

Then maybe the next better option is GL502VS? value+full 1070+thin and light. i find it better option than p650
2.65kg vs 2.2kg

not sure how is P650HS-G any better tbh..

 

And that is precisely why my next high-end machine will not be a notebook, but a Ryzen desktop. I will keep my current notebook for portability's sake, and primarily because it still works. The HDD is likely to start failing soon, too, so I'll swap that out with a 1 TB SATA 2.5" SSD next year, too. Notebooks are lousy for price/performance, and I don't think I'll buy one again any time soon.

 

Because you're not an ignorant person...some usually wants a ultra thin powerful gaming laptop so when Ngreedia puts a 1070MQ, 1080MQ into those chassis some thinks they have it all when little did they know they've been ripped off by paying more for less performance...usually a person who lacks research.

I also gotta admit this same mistake when I purchased the ASUS G501...I fall to that mistake of having a light thin gaming machine would be the greatest choice of all. lol

 

True, I saw that yesterday.

 

GL502VS is garbage, and as far as I know, should weigh more than the P650HS? It doesn't handle the 1070N easily either and a LARGE number of people did the undervolting thing there. I think that machine is largely what started it, because people had so many heat issues.

 

Gl502vs is not 2.2kgs. Check notebookcheck reviews

 

I went desktop/tablet and have not yet regretted it. Though if I have to start being on the road more, I might have to go back to some sort of laptop.

 

the source i've seen was from NBR too but checked the actual review now and it's 2.6kgs ... sorry

 

Official figure puts it at 2.33 Kg . I know in reviews it's more: some reported 2.5 - 2.6, others placed it at 2.34 http://www.anandtech.com/show/10885/the-asus-rog-strix-gl502vs-review-mainstream-gtx-1070
I will just stick to the official figure of 2.33Kg : https://www.asus.com/us/ROG-Republic-Of-Gamers/ROG-GL502VS/

 

If you don't need the portability then yeah, it has been like that for the last 20 or so years.

All of a sudden everyone seems so shocked that you pay a premium for form factor.

Official figure is wrong, notebookcheck is right. Go ask @link626 (eg users who own these machines) or @Donald@HIDevolution the resellers

 

I think it's mostly that people look only at performance and nothing else. They see a system that's engineered to perform while being thinner and lighter and quieter and look right past it to say it's trash because it doesn't perform like a version that's 3x the size/weight, like none of those other factors can possibly matter to anyone.

 

I see every notebook as a shoe, you have those big sizes for people like me and the small sizes for those like my wife. It's all about finding the perfect fit for you.

Just because you take a size 10 that does not mean a size 6 is trash (even though it costs the same) but wont fit you.

 

I believe both have your merits and there are people that undoubtedly label everything they don't like or agree with as trash. But there are others that see a problem with a size 6 sneaker being marketed as a size 10 trekking shoe "Max Confort". I believe @D2 Ultima is in this camp, for example.