New macbook Pro USB C kills WiFi

[dragosudeki]

28 minutes ago, NeoZeon said:

I find it funny that you call yourself a the forum's "scientist" and have such a closed mindset, when looking at evidence you would see that many scientists use apple products. Almost my entire social circle is filled with scientists. I don't know every scientist, but when I go to conferences with 20k+ PhDs and see a sea of apple computers and ipads, I have to think that you are just incredibly wrong.

 

I'm going to copy and paste my post from the LED thread on macs:

 

Keep talking people... it only serves further illuminate who the "pros" are. 

I agree that some people here can act like they know stuff and doesn't really keep in mind of every kind of user out there. I also agree that some people overreacted to the transition of physical function buttons on the top to a touch display instead (which is not this topic, but a different one).

Aside from the first paragraph, the rest has literally nothing to do what people have been disliking about Apple's recent design changes (especially this post). Most of what is talked about in your quote is the performance/size of Apple's products, in which people are mainly paying for in any laptop. The entire use case is almost just terminal, so Apple can literally remove every single I/O port from the laptop and it would be fine anyway because its not needed for that specific use case.

This entire topic is about the fact that if that person were to plug in a hard drive or other specific devices, the WiFi may (or may not depending on the severity of this issue) start to slow down. This can be a little disruptive if you're doing things remotely through ssh or something. Other than that, most logical complaints have been about the case design, which is crucial more so for a laptop because the internal hardware isn't something you can mix/pick/choose from like a desktop.

[goodtofufriday]

6 hours ago, VerticalDiscussions said:

Without the usual bonehead hate in this post from some users, could this be an issue that may be fixed with the top tier models and just another way to get more money from the performance users (unless those cpu's dont suppor additional pciE lanes or they were not added phyisically)? 

 

Also, Louis Rossman is someone that lives with hate and its not just because of Linus either xD. But this is just the extra... Have you seen how that fool keeps the replacement motherboards on the ground, all stuffed and pilled up on each other? Ridiculous, this guy has the least sense of organization ive ever seen from an "Elite Fixer". Fancy gear, good basis, bad habits and even worse component safeguard. Would surely not leave any machine with this guy lol.

As a repair guy, i leave donar parts in a box as well. Why? cause it dont matter. Any large components wont be getting damaged on a laptop pcb. Any smaller ones can be gotten anywhere. 

He doesnt use those boards as entire replacements, as your comment seems to imply.

[79wjd]

30 minutes ago, goodtofufriday said:

As a repair guy, i leave donar parts in a box as well. Why? cause it dont matter. Any large components wont be getting damaged on a laptop pcb. Any smaller ones can be gotten anywhere. 

He doesnt use those boards as entire replacements, as your comment seems to imply.

Still he should be more organized, although thats not my problem with Rossman. He knows what he's talking about with board repairs, but he kind of makes an ass of himself when he starts talking about other things.

 

From just this live stream: 

• $30 adapter must be better than the $5/$10 adapters (and should be good) because it's more money and has good reviews.
  • Unfortunately that's not how things always work. Just because an adapter is more expensive doesn't make it better and just because it has good reviews doesn't mean anything. 
• This 1 year old XPS 15 is the same price and beats the 13" brand new pro
  • maybe with a non-equivalently spec'd Dell (i.e. 1080p display?) and/or with a sale price, because the current price points put the XPS 13 @ $1400 and the 13" Pro @ $1500 while the base XPS 15 with an >1080p display starts at $1850. (and of course the OLED 13" starts at $1800/15" @ $2400, so that's a pretty shit price point due to the OLED Bar tax, but that's not what Rossman was comparing here).
• Only apple separates CPU performance based on screen size 
  • Except you also can't get a quad core XPS 13. 
• "Up to 130% faster graphics" and then bitches how the 13" cheapest retina is only 30%~ faster.
  • But that 130% number is specifically listed as referring to the Pro 460 vs. M370x 15" (in Maya).
• Expecting worse performance from a one year old computer (XPS 9550)
  • Even though Skylake has and still is the main laptop platform.....so, no shit, the performance shouldn't be worse on the "1 year old" laptop.  

 

That's not to say he doesn't make some good points, he does, but he still talks out of his ass a bit too much. 

[LAwLz]

1 hour ago, NeoZeon said:

I find it funny that you call yourself a the forum's "scientist"

Good that you find it funny. I am doing it ironically.

 

1 hour ago, NeoZeon said:

and have such a closed mindset

You think I got a closed mind? Well geez I don't know what to tell you. I try to be open minded but I think Apple is going in the wrong direction. My statement was an hyperbole but maybe that did not come across well enough in my post. Yes a lot of professionals use Macs. Hell, at my job like 1/3 uses Macbooks. Apple is clearly making this less of a "pro" computer and more of a "this is what we think casual users will want" approach.

You should get off your high horse.

[Zodiark1593]

Off Topic, but I just realized it had been well over a year since Skylake launched. 

[goodtofufriday]

15 minutes ago, djdwosk97 said:

Still he should be more organized, although thats not my problem with Rossman. He knows what he's talking about with board repairs, but he kind of makes an ass of himself when he starts talking about other things.

 

From just this live stream: 

• $30 adapter must be better than the $5/$10 adapters (and should be good) because it's more money and has good reviews.
  • Unfortunately that's not how things always work. Just because an adapter is more expensive doesn't make it better and just because it has good reviews doesn't mean anything. 
• This 1 year old XPS 15 is the same price and beats the 13" brand new pro
  • maybe with a non-equivalently spec'd Dell (i.e. 1080p display?) and/or with a sale price, because the current price points put the XPS 13 @ $1400 and the 13" Pro @ $1500 while the base XPS 15 with an >1080p display starts at $1850. (and of course the OLED 13" starts at $1800/15" @ $2400, so that's a pretty shit price point due to the OLED Bar tax, but that's not what Rossman was comparing here).
• Only apple separates CPU performance based on screen size 
  • Except you also can't get a quad core XPS 13. 
• "Up to 130% faster graphics" and then bitches how the 13" cheapest retina is only 30%~ faster.
  • But that 130% number is specifically listed as referring to the Pro 460 vs. M370x 15" (in Maya).
• Expecting worse performance from a one year old computer (XPS 9550)
  • Even though Skylake has and still is the main laptop platform.....so, no shit, the performance shouldn't be worse on the "1 year old" laptop.  

 

That's not to say he doesn't make some good points, he does, but he still talks out of his ass a bit too much. 

ehh Cant really disagree here. I do believe point 2 focuses more on I/O. For people like him, and myself, I/O is a large part of what makes a laptop better or worse. My purchasing decisions always end up being on th I/O side of things.

[79wjd]

5 minutes ago, goodtofufriday said:

ehh Cant really disagree here. I do believe point 2 focuses more on I/O. For people like him, and myself, I/O is a large part of what makes a laptop better or worse. My purchasing decisions always end up being on th I/O side of things.

It wasn't about the I/O, start watching around 37 minutes: https://www.youtube.com/watch?v=VPYEw-Ig_JE

 

The I/O is going to be hit or miss for users -- I saw a survey earlier today and it was about 50/50 for/against the new I/O.

 

 

 

 

Unrelated notes: 

Other things to note from Rossman's review about the new Mac:

• The keyboard is good, surprisingly, although it's obnoxiously noisy (but that's probably something that becomes less of an issue as you get used to it and stop bottoming out the keys so much) 
• Doesn't seem to throttle
• Trackpad is perfect as far as current trackpads go (unsurprisingly) 
• Speakers are a noticeable improvement over previous Macs (albeit still laptop speakers....so not "good")
• 13" base retina (HD540) is 30%~ faster than 2015 13" (HD6100)
• CPU performance is about the same (2ghz Skylake (3.1-2.9ghz turbo) vs. 2.7ghz Broadwell (3.1ghz turbo))

[Belgarathian]

I love Apple, but if this turns out to be true... Wow. Fail. 

[RexinOridle]

Apple never really upgrades the products that matters in time. Like Mac Pro.

[Monkey Dust]

4 hours ago, NumLock21 said:

No, you're reading it wrong. Look at desktops, gpu already eats up 16. If there is only 16 lanes, a lot of people will be pissed they don't 4 lanes to spare for their M.2

Well high end GPUs aren't compromised by only having 8 Gen 3 lanes available. The most powerful GPU available in the new Pros is the AMD 460, and AMD let you run their cards in a desktop with just 4 lanes, in a laptop you might get away with just 2?

 

Wifi interference from USB3 seems like something they would have picked up in development.

 

I suspect he has just got a faulty example, if this was widespread I'd have though this would be everywhere by now. Unless no one else tried plugging a USB3 device in because no one owns one... 

[NumLock21]

1 minute ago, Monkey Dust said:

Well high end GPUs aren't compromised by only having 8 Gen 3 lanes available. The most powerful GPU available in the new Pros is the AMD 460, and AMD let you run their cards in a desktop with just 4 lanes, in a laptop you might get away with just 2?

 

Wifi interference from USB3 seems like something they would have picked up in development.

 

I suspect he has just got a faulty example, if this was widespread I'd have though this would be everywhere by now. Unless no one else tried plugging a USB3 device in because no one owns one... 

On desktop you have some boards that does x16 / x4. This indicated the 1st slot uses all 16 lanes from the CPU, leaving the rest to be handled by the chipset. I know the GPU can run at x4, which is why CFX supports it even at x4 mode, but that all depends on how the board is configured. And you only get 3 choices, x16, x8 or x4 and nothing in between.

[79wjd]

19 minutes ago, NumLock21 said:

On desktop you have some boards that does x16 / x4. This indicated the 1st slot uses all 16 lanes from the CPU, leaving the rest to be handled by the chipset. I know the GPU can run at x4, which is why CFX supports it even at x4 mode, but that all depends on how the board is configured. And you only get 3 choices, x16, x8 or x4 and nothing in between.

Well, assuming Apple's engineers aren't a bunch of mentally challenged monkeys, one would make the assumption that they wouldn't give a full 16 lanes to the GPU since then there wouldn't be enough lanes for the remaining ports, SSD, and wifi. Although, only time will tell. I'm waiting on a review from someone(s) who actually tries many configurations and tries to cause what Rossman showed (and what eventually wasn't even an issue in Rossman's own video). 

[NumLock21]

Just now, djdwosk97 said:

Well, assuming Apple's engineers aren't a bunch of mentally challenged monkeys, one would make the assumption that they wouldn't give a full 16 lanes to the GPU since then there wouldn't be enough lanes for the remaining ports, SSD, and wifi. Although, only time will tell. I'm waiting on a review from someone(s) who actually tries many configurations and tries to cause what Rossman showed (and what eventually wasn't even an issue in Rossman's own video).

Well some can fire up their laptops with discrete gpu and see what it's running at. My netbook's APU can't be detected, says N/A.

[jagdtigger]

Just now, djdwosk97 said:

Well, assuming Apple's engineers aren't a bunch of mentally challenged monkeys, one would make the assumption that they wouldn't give a full 16 lanes to the GPU since then there wouldn't be enough lanes for the remaining ports, SSD, and wifi. Although, only time will tell. I'm waiting on a review from someone(s) who actually tries many configurations and tries to cause what Rossman showed (and what eventually wasn't even an issue in Rossman's own video). 

Actually the ones using the "U" CPU's have less than 16, the 7200U is only has 12 for instance...

 

Just now, NumLock21 said:

Well some can fire up their laptops with discrete gpu and see what it's running at. My netbook's APU can't be detected, says N/A.

[79wjd]

6 minutes ago, jagdtigger said:

Actually the ones using the "U" CPU's have less than 16, the 7200U is only has 12 for instance...

 

 

Good point, here's the PCIE config for the 6360u/6267u/6287u: 1x4, 2x2, 1x2+2x1 and 4x1

[NumLock21]

Spoiler

 

1 minute ago, jagdtigger said:

Actually the ones using the "U" CPU's have less than 16, the 7200U is only has 12 for instance...

 

 

 

 

Okay so the mobile GPU does take away 16 lanes. And yes looks like mobile dual core only have 12 lanes not 16.

[79wjd]

5 minutes ago, NumLock21 said:

  Hide contents

 

 

 

 

Okay so the mobile GPU does take away 16 lanes. And yes looks like mobile dual core only have 12 lanes not 16.

It CAN take all 12/16 lanes. It doesn't have to though. 

 

And remember, the 13" has an iGPU only. So it only needs 4 (ssd) + 2 (Wifi) + 4 * 2/4 (USB-C) = 22 lanes (and there is a minimum of 12+6 with the cheapest chipset, and 12+16 with a more likely chipset). So there should be plenty of lanes. 

[SansVarnic]

Apple ... always to easy to hate on.

 

Always

Purposefully 

Proprietary [and]

Lacking

Exceptionalism

 

(sorry its still a work in progress)

 

 

[NumLock21]

18 minutes ago, djdwosk97 said:

It CAN take all 12/16 lanes. It doesn't have to though. 

 

And remember, the 13" has an iGPU only. So it only needs 4 (ssd) + 2 (Wifi) + 4 * 2/4 (USB-C) = 22 lanes (and there is a minimum of 12+6 with the cheapest chipset, and 12+16 with a more likely chipset). So there should be plenty of lanes. 

which chipset only as 6 lanes?

[SlipperyPete]

10 hours ago, Dackzy said:

so you don't like that a person shows a problem with a product? You sir have a very weird attitude to this. 

If you think that it is okay for a company to sue a person because they find a flaw in the product, then you have a fucked up way of thinking.

 

He does not go out of his way at all. he tests a laptop and he sees that there is a problem with you use a USB port, how often do you use a USB port while surfing on the web or using some form of net.

a mouse

[79wjd]

1 hour ago, NumLock21 said:

which chipset only as 6 lanes?

H110 supposedly only has 6, although I really doubt H110 is what's used here. But that would be an absolute worse case scenario. 

 

 

 

 

 

Also, I got a change to use someone's 13" non-oled MBP, and a few things. First off, there was no usb-c/wifi related issues at all. Secondly, the keyboard isn't nearly as loud as I originally though it was from Rossman's video. I don't really find it any louder than my 2013 rMBP. And supposedly the battery life is even better than his previous 2013 13" rMBP. So yeah. the usb-c/wifi may just be an isolated incident. 

[NumLock21]

1 minute ago, djdwosk97 said:

On one of the previous pages someone listed the lanes by chipsets (I just assumed they were accurate). 

 

 

 

 

Also, I got a change to use someone's 13" non-oled MBP, and a few things. First off, there was no usb-c/wifi related issues at all. Secondly, the keyboard isn't nearly as loud as I originally though it was from Rossman's video. I don't really find it any louder than my 2013 rMBP. And supposedly the battery life is even better than his previous 2013 13" rMBP. So yeah. the usb-c/wifi may just be an isolated incident. 

Both mobile skylake chipset have 16 lanes not 6. H110 has 6, now that is for desktops. M.2 on those boards can run at x1.

 

[79wjd]

2 minutes ago, NumLock21 said:

Both mobile skylake chipset have 16 lanes not 6. H110 has 6, now that is for desktops. M.2 on those boards can run at x1.

 

Ah, alright. So yeah, definitely not a physical lane issue then (unless they just wired them poorly).

 

I'd really hope M.2 isn't wired for x1 though as that would be a pretty big bottleneck. It should be wired for 3.0x4 to not get bottlenecked. Either way, that's x4 (SSD) + x2 (wifi) + x4*2/4 (usb-c) -> 22 lanes for the 13" (out of 28 possible lanes), and 22+8(gpu) for the 15" (out of 32 possible lanes).

[Qwerty2323]

While so many go to praise apple for usb c and their push to future seems like laptops anit ready to handle power of these ports.

[VerticalDiscussions]

18 hours ago, goodtofufriday said:

-snip-

Im sorry, what were you saying about a box? Check at 5:41 to see what i mean, even Jerry got mind blown that this guy just piles up all the motherboards on the freaking ground!