Best laptop for Music production: very curious - Macbook?

[GusHage]

Wow, loving the discussion!

Do you guys know how big a performance difference there is between the previous generations of the i7 quad-cores, specifically for music production, if any?

I am asking because I'm considering bying a 15" rMBP second hand,and I wanna know if there is a significant performance boost in the newer models?

Any imput is appreciated (:
EDIT: for example, is there a difference between a "2012 2.7 GHz i7" and a "2015 2.7 GHz i7"?

I am aware that the SSDs have received a speed increase..

[fishmonkey]

you can search and compare models on this page:

http://browser.primatelabs.com/geekbenc ... ro+15-inch

[GusHage]

you can search and compare models on this page:

http://browser.primatelabs.com/geekbenc ... ro+15-inch

Thank you.
Wow, those stats are impressively thorough.. Unfortunately, the numbers do not make a lot if sense to me, unless "higher is better", and I simply want to know if that same principle then applies to Ableton performance?

[Tarekith]

You can see more real world examples here:

viewtopic.php?f=1&t=191549&start=0&hili ... mance+test

Not totally indicative of the exactly how fast a computer is, but will give you a rough idea.

[H20nly]

Reason for the wattage, Xeon server chips, originally two four core 2.66ghz, now two six core 3.33ghz monsters.
Live of course doesn't use them as effectively as other DAWs but still not bad.

Surprisingly the whole thing is really quiet. I'm effectively out of the race for more power for probably another 4 years, on a 2009 mac.

nice. sounds like a beast.

did you take the El Capitan plunge?

i avoided Yosemite on my iMac, i like Mavericks... mostly because it and everything else installed on it or attached to it works, but now i'm a little torn because i don't know if i should kick myself for not at least downloading Yosemite.

on the Windows side, i've been waiting for a copy of 10 to try out on my laptop. i don't know if it's Lenovo or Microsoft that are holding me up... but the upgrade email never comes. our tests at work with the demo show that it has the promise of being a great OS to jump from 7 to. i want try it with my Apogee interface and see how it fares.

[Machinesworking]

You can see more real world examples here:

viewtopic.php?f=1&t=191549&start=0&hili ... mance+test

Not totally indicative of the exactly how fast a computer is, but will give you a rough idea.

100% disagree here, Geekbench scores map out very closely to real life power, it's mostly CPU stress tests.

The Live performance tests here are not to failure, and Live's CPU meter has always been a rough estimate. You simply are not going to get a real world idea of the power of a computer by running a test that isn't right up against it's failure rate, i.e. when it starts to affect the audio.

For instance the way that Lives CPU meter reads with machines with more than a few cores is misleading in the tests here, right away it will jump to 60% for example, then barely be affected by additional stress. Probably due to the way Live deals with multiple cores. So a machine for example with two 3.2ghz chips will look like it's stomping on an eight core machine running 2.2ghz chips, but in ANY stress test to failure the 8 core will load twice to three times as many plug ins before the audio is affected. Love you all man, but the "Live 8-9 Performance Tests" here have always been incredibly bad science, not at all real world.

Any test to see if a computer can handle x amount of plug ins and audio tracks etc. has to be to failure. Otherwise it's just not going to report anything that will be useful what with modern computers, 2, 4, 8, 12 core systems, and the way applications interact with those cores.

[Machinesworking]

Reason for the wattage, Xeon server chips, originally two four core 2.66ghz, now two six core 3.33ghz monsters.
Live of course doesn't use them as effectively as other DAWs but still not bad.

Surprisingly the whole thing is really quiet. I'm effectively out of the race for more power for probably another 4 years, on a 2009 mac.

nice. sounds like a beast.

did you take the El Capitan plunge?

i avoided Yosemite on my iMac, i like Mavericks... mostly because it and everything else installed on it or attached to it works, but now i'm a little torn because i don't know if i should kick myself for not at least downloading Yosemite.

on the Windows side, i've been waiting for a copy of 10 to try out on my laptop. i don't know if it's Lenovo or Microsoft that are holding me up... but the upgrade email never comes. our tests at work with the demo show that it has the promise of being a great OS to jump from 7 to. i want try it with my Apogee interface and see how it fares.

I'm running El Capitan on my laptop, it's OK. Doesn't seem like that big of a jump from Yosemite. Windows do load quicker, TRIM support is across the board in El Capitan I hear. The app snapping feature is garbage. It's been a while since I used Windows, but I remember it being better than what Apple did here.

Yosemite on the desktop. There are a ton of caveats for me with El Capitan: NI Rig Control, NI in general, MOTU hardware, older software (Office 2004 etc.) that may not make it. To be fair MOTU and NI both have beta drivers out for El Capitan, but that's not fun. I will have to get a new USB3 card with El Capitan, the PCIe card I have installed isn't even supposed to work correctly in Yosemite but I seem to be lucky. I'm booting from a M.2 SSD on a PCIe card and there's that fear that El Capitan will screw that up somehow as well. I still have my old hard drive in there with OSX on it though, so I could use it to test drivers anyway.

Windows 10 has a lot of improvements for audio from what I've read. It's possible that Microsoft really get it right with the touch tech and their OS, looking at what Bitwig did with it. Aren't Apogee one of the few OSX only audio interface developers out there?

[Tarekith]

You can see more real world examples here:

viewtopic.php?f=1&t=191549&start=0&hili ... mance+test

Not totally indicative of the exactly how fast a computer is, but will give you a rough idea.

100% disagree here, Geekbench scores map out very closely to real life power, it's mostly CPU stress tests.

The Live performance tests here are not to failure, and Live's CPU meter has always been a rough estimate. You simply are not going to get a real world idea of the power of a computer by running a test that isn't right up against it's failure rate, i.e. when it starts to affect the audio.

For instance the way that Lives CPU meter reads with machines with more than a few cores is misleading in the tests here, right away it will jump to 60% for example, then barely be affected by additional stress. Probably due to the way Live deals with multiple cores. So a machine for example with two 3.2ghz chips will look like it's stomping on an eight core machine running 2.2ghz chips, but in ANY stress test to failure the 8 core will load twice to three times as many plug ins before the audio is affected. Love you all man, but the "Live 8-9 Performance Tests" here have always been incredibly bad science, not at all real world.

Any test to see if a computer can handle x amount of plug ins and audio tracks etc. has to be to failure. Otherwise it's just not going to report anything that will be useful what with modern computers, 2, 4, 8, 12 core systems, and the way applications interact with those cores.

How did I know you were going to post this? FWIW, I agree with you, and have for some time. People just keep copying and pasting my orginal test from ages ago each time a new version comes out. Still, while it might not be to failure, he can see roughly how many tracks we're talking about on the new machines compared to some others.

[stringtapper]

Aren't Apogee one of the few OSX only audio interface developers out there?

Metric Halo are still all Mac on the hardware side.

[deva]

Aren't Apogee one of the few OSX only audio interface developers out there?

Metric Halo are still all Mac on the hardware side.

And they are Firewire which Macs no longer have...

[Machinesworking]

How did I know you were going to post this? FWIW, I agree with you, and have for some time. People just keep copying and pasting my orginal test from ages ago each time a new version comes out. Still, while it might not be to failure, he can see roughly how many tracks we're talking about on the new machines compared to some others.

I'm saying, no he can't see a rough anything. The Live performance test would only give similar numbers for similar machines. Mostly it's only going to be even close to accurate at comparing machines with similar amounts of CPUs. Comparing a 3.2ghz dual core Windows i5 to a quad core macbook pro 2.4ghz i7 machine with the Performance test for instance, the dual core i5 will very likely look like a faster machine with the test, whereas in real life when using it up to 80% cpu etc. the macbook will easily outperform the dual core.

Now compare that to doing a stress test where you get right bellow crackling audio by adding plug ins with a few tracks running. A straight real world test.
With that sort of test Geekbench scores perfectly match up, a computer scoring 75% compared to another is going to be able to use 75% of the plug ins in a DAW the other one can.

There is absolutely no benefit of knowing that your computer scores 70% compared to some else's 50% if in the real world your computer can load twice as many plug ins before crackling than theirs. That's what I found to be the case with the Live test. Live tends to show CPU use as high when it maxes out your first CPU, but it trudges on using the other ones and adding only marginally to the CPU count in Live, getting to twice as many plug ins in some cases with an only 5% jump in CPU. That isn't a small problem, or a minor flaw in the test, it completely negates it's usefulness.

You're an incredibly intelligent good natured person for sure and I'm not harping on this because I have some sort of agenda, we all make mistakes sometimes and honestly it wasn't even you who started these tests, it was a user who doesn't even post here anymore. If you agree with me, get rid of that thread and post a real stress test, then you'll see real numbers.

[Machinesworking]

Aren't Apogee one of the few OSX only audio interface developers out there?

Metric Halo are still all Mac on the hardware side.

And they are Firewire which Macs no longer have...

It's a $25 converter to use Firewire with Thunderbolt. Thunderbolt is so much faster than firewire that there is absolutely zero latency added to do this.
They will probably eventually switch to USB3, but that won't be any improvement or big deal.
No reason to use Thunderbolt, the cables alone are $50.

[fishmonkey]

the problem now is that Apple is dropping development and maintenance of their Core Audio Firewire driver from El Capitan onwards. for now it seems that most stuff still works, but for how long?

my next interface will be Thunderbolt for sure (i have one USB 2.0 and one Firewire interface now)...

[H20nly]

Aren't Apogee one of the few OSX only audio interface developers out there?

And they are Firewire which Macs no longer have...

It's a $25 converter to use Firewire with Thunderbolt. Thunderbolt is so much faster than firewire that there is absolutely zero latency added to do this.
They will probably eventually switch to USB3, but that won't be any improvement or big deal.
No reason to use Thunderbolt, the cables alone are $50.

i have on of those converters and it worked GREAT with my TC Electronic interface.

regarding Apogee as Mac only... doh! that's right... completely spaced that.

the only way i'll be able to A/B test the audio on Windows 7 with my iMac is to install my TC Electronic interface on both. problem there is i have to use firewire card, at least the one i have has a TI chipset...

[Machinesworking]

the problem now is that Apple is dropping development and maintenance of their Core Audio Firewire driver from El Capitan onwards. for now it seems that most stuff still works, but for how long?

my next interface will be Thunderbolt for sure (i have one USB 2.0 and one Firewire interface now)...

If you got to the system profile, you will still see support for Parallel SCSI, that should be of some comfort.