Teardown of New Mac Pro Reveals Surprising Amount of Accessibility, Circular Daughterboard

[milo]

... you're not letting me make the point and instead are speaking condescendingly to me as an attempt to make me realize my view of hardware requirements are misguided..

You give examples of people using midrange hardware and jump to the conclusion that nobody needs high end hardware. If you're not making a point, it comes down to your communication skills.

except for the part about 1 single gpu in the nmp uses all the bandwidth(16lanes) that the haswell has to offer...

The high end i7 have 40 lanes. And there are additional lanes on the motherboard, the nMP has that as well. You really think that nobody with i7 is running even a single high end GPU?

That's the best excuse for the $2999 quad core? PCIe lanes?

...the million dollars he spent...

You don't need anything close to that dollar amount or what is shown in the picture to do high end audio work. The beauty of computer based production is that a ton of things can be done without all that expensive hardware...but many things do require more than a midrange computer.

 

[flat five]

So, what’s more logical, to design a computer expecting great leaps in software or to just design a computer for what current software needs?

if you were building computers, i think it's more logical to realize moore's law has run it's course and multi core cpus was never truly a viable option.. only a stop gap solution.. not only are cpu's barely getting faster these days, their prices aren't dropping either..

most current software relies on noticeable gains coming from increased cpu clocks at prices their customers can afford..

and we've hit a brick wall.. it's time for a change.. if software devs want to give their customers great increases in performances at a cost they can afford then they're going to need to switch up the mindset and programming techniques and it's not as if there's some untapped algorithm which is going to unleash this kind of leap under the current cpu based train of thought..

so two things need to happen.. different software and different hardware..
apple is selling hardware so they've done/are doing their part.. the ball is in the devs court now..

----------

That's the best excuse for the $2999 quad core? PCIe lanes?

not an excuse but 'the best excuse' i can come up with for apple's $3000 quad is that $1000 of it is in the chassis/design.

 

[wallysb01]

if you were building computers, i think it's more logical to realize moore's law has run it's course and multi core cpus was never truly a viable option.. only a stop gap solution.. not only are cpu's barely getting faster these days, their prices aren't dropping either..

most current software relies on noticeable gains coming from increased cpu clocks at prices their customers can afford..

and we've hit a brick wall.. it's time for a change.. if software devs want to give their customers great increases in performances at a cost they can afford then they're going to need to switch up the mindset and programming techniques and it's not as if there's some untapped algorithm which is going to unleash this kind of leap under the current cpu based train of thought..

so two things need to happen.. different software and different hardware..
apple is selling hardware so they've done/are doing their part.. the ball is in the devs court now..

That’s just naive. Plain and simple. Computers are all temporary. If multicore CPUs is a stopgap, every computer ever is a stopgap. That goes for the nMP too. So why design today’s stop gap to use code not yet written? We all know the improvements GPGPU can have if the task is up to it, adding one computer model doesn’t increase the pressure for them to code for that.

But I’m glad you dodged the vast majority of the points in my previous post. I’ll assume you are still internalizing it.

not an excuse but 'the best excuse' i can come up with for apple's $3000 quad is that $1000 of it is in the chassis/design.

BTW, that made me laugh. $3000 quad core, and the best you can think of is the chassis/design.....haha, maybe you’re coming around huh?

 

[flat five]

You don't need anything close to that dollar amount or what is shown in the picture to do high end audio work. The beauty of computer based production is that a ton of things can be done without all that expensive hardware...but many things do require more than a midrange computer.

sure you do.. more than a million dollars in many locations.
that building in the video i linked to earlier which houses a lounge/bowling lane, a music studio, and a skatepark is a $3.2million facility..

you're the one that initial brought up the high-end studio thing to which i simply pointed out that $1000 difference in a computer is nothing in that scenario.. which it isn't.

so either i don't know what a high end studio is and generally how much one costs or you're completely using the phrase wrong..

i mean, what is the type of studio in the image i posted called? "super duper mega high end" or something?


but since you've said high end studio to which i've pointed out there's possible semantics/communication issues when using that term to describe your usage, i've subsequently been told i've lost the argument, i'm out of touch with real world requirements, i'm blind, i'm biased, etcetc by multiple people...

so are you in the high end studio business or not? or am i at fault for incorrectly assuming "high end music production" is a term which would describe places like flux? (and @everyone else, feel free to retract all the stuff you've said to me which stems from the original posts about it once you see milo's response to this)

----------

BTW, that made me laugh. $3000 quad core, and the best you can think of is the chassis/design.....haha, maybe you’re coming around huh?

it's not the best i can think of.. it is what it is.. the design/chassis truly is a big percentage of the cost of a quad Mac Pro.. (of course, this percentage drops as you get into the higher configurations)

----------

So why design today’s stop gap to use code not yet written?

the code is already written.. it's not as if apple made this computer then said "we're going to start writing openCL now".. they started writing OpenCL nearly a decade ago

----------

adding one computer model doesn’t increase the pressure for them to code for that.

apple isn't necessarily pressuring developers to code for the nmp..
the developers want to improve their programs.. they pressure themselves or they're going to get out-competed within the next couple of years..
it's not even a win/mac thing.. it's a performance thing and not platform specific.

 

[wallysb01]

it's not the best i can think of.. it is what it is.. the design/chassis truly is a big percentage of the cost of a quad Mac Pro.. (of course, this percentage drops as you get into the higher configurations)

Right, but the problem is that the design/chassis doesn't really do the work, does it? So if I put a low value on aesthetics and just need a basic 4-core workstation without dual GPUs, its a really terrible deal.

the code is already written.. it's not as if apple made this computer then said "we're going to start writing openCL now".. they started writing OpenCL nearly a decade ago

Sure, openCL is writen, but the code that interfaces with openCL to get specific jobs done is NOT written in almost everything I do (much of the GPGPU that is available in bioinformatics is CUDA), and isn't there for a great many other things.

apple isn't necessarily pressuring developers to code for the nmp..
the developers want to improve their programs.. they pressure themselves or they're going to get out-competed within the next couple of years..
it's not even a win/mac thing.. it's a performance thing and not platform specific.

Which is my point, actually. If the code can be written to use GPGPUs via openCL, there already exists that pressure to make it so. So Apple forcing the hardware that uses openCL doesn't add anything.



I will again point out that you've now dropped many off the topics we were previously discussing.

 

[milo]

sure you do.. more than a million dollars in many locations.

Falling back on the same lack of argument over and over. You keep pointing out that people have spent over a million dollars on studios. That's not evidence that you need to spend that much to do high end work.

you're the one that initial brought up the high-end studio thing

No, I brought up doing high end work. It's not necessary to spend a million dollars to do that work, but in many cases it does require a high end computer.

so are you in the high end studio business or not?

Go back and read what I wrote. I'm doing detailed orchestral mockups. Which have high hardware requirements but no, don't require a room like that. And for the record, I have worked in million dollar rooms like that as well as much more modest facilities so I know the capabilities and what is required for various kinds of projects. What exactly have you done in those studios you keep talking about?

or am i at fault for incorrectly assuming "high end music production" is a term which would describe places like flux?

You are at fault for assuming that. Does high end music production happen at places like that? Yes. Does it happen only at places like that? Or in other words does high end music production (which is a broad term that would include many very different things) require a studio like that? Nope, that's where you are getting confused.

...they started writing OpenCL nearly a decade ago

You mean creating the standard? He's talking about updating apps to use OpenCL which is going to take a lot of time and effort.

the developers want to improve their programs.. they pressure themselves or they're going to get out-competed within the next couple of years..

Only if a critical mass of computers get powerful enough GPU, otherwise they would be doing code that only benefits on machines like the MP. Some apps are aimed at that niche anyway but others may not have enough of a user base with the right machines to make it worthwhile.

 

[flat five]

Falling back on the same lack of argument over and over. You keep pointing out that people have spent over a million dollars on studios. That's not evidence that you need to spend that much to do high end work.

lol. I'm not even arguing anything.

you - "I do high end audio production"
me - "oh.. then you don't have to worry about minuscule costs such as computers"
others - "you're out of touch ..$1000 is not something to be taken lightly"
me - "I know.. I was being facetious and pointing out a semantics issue with the term "high end"
you - "$1m is not necessary to produce high quality work"
me - "yeah, I know"
you - "your argument is wack that audio work requires that type of investment"
me - "right. I understand that"
you - "you just keep repeating the same argument over and over"
me - "hmmm"

etcetcetc

No, I brought up doing high end work. It's not necessary to spend a million dollars to do that work, but in many cases it does require a high end computer.

see above

Go back and read what I wrote. I'm doing detailed orchestral mockups. Which have high hardware requirements but no, don't require a room like that. And for the record, I have worked in million dollar rooms like that as well as much more modest facilities so I know the capabilities and what is required for various kinds of projects. What exactly have you done in those studios you keep talking about?

hang out.. smoke weed.. look at naked girls.. listen to some jams.. even some cocain but that one was quite a while ago.
you asked...

You are at fault for assuming that. Does high end music production happen at places like that? Yes. Does it happen only at places like that? Or in other words does high end music production (which is a broad term that would include many very different things) require a studio like that? Nope, that's where you are getting confused.

dunno.. there's no shame in saying smthng like "I produce music which demands high quality" (or whatever).. not sure why you're getting so (seemingly) defensive that you're not working in a high end studio..
my design studio is at a desk in my apartment.. that says nothing about the quality of my work.

You mean creating the standard? He's talking about updating apps to use OpenCL which is going to take a lot of time and effort.



Only if a critical mass of computers get powerful enough GPU, otherwise they would be doing code that only benefits on machines like the MP. Some apps are aimed at that niche anyway but others may not have enough of a user base with the right machines to make it worthwhile.

here's a problem in this thread..
all of you are more/less arguing me about what apple should do.. as if I'm arguing back about what I think apple should do.

but I'm not.. I'm just pointing out the obvious and saying what apple IS doing.

in an effort to not make a longer than necessary post to further explain that, do you understand the communication breakdown which is occurring because of this?

 

[flat five]

.



I will again point out that you've now dropped many off the topics we were previously discussing.

what's there left to discuss? I was saying you should use window for at least the next couple of years.. you're saying you're doing exactly that.
we're in agreement, right?

 

[MagnusVonMagnum]

One might consider that even if a Windows machine can do it cheaper, some tools are Mac only and in the case of Logic Pro, it's the primary reason I bought my 2008 Macbook Pro to begin with and frankly, given Logic's tools and quality, it was worth every penny. I think between the computer, ram and hard drive upgrades plus Logic and PreSonus/Midi-Man, it cost me around $3000. Throw in my instruments and my high-end audiophile monitoring speakers and it was probably closer to $10k, still a bargain compared to paying for any real amount of studio time and I think I got Pink Floyd Wish You Were Here level sound quality in the final product (yes the album is rather Pink Floyd-like).

More importantly, I could take my time. I think with 10 songs on the album averaging 4.5 minutes each overall (album is 45 minutes and some odd seconds), the average time for each song from start to finish plus mastering might have been on the order of 60 hours each (guess based on latter tracks that went faster). So, 60 * 10 * $50 an hour average = $30k (and I may be WAY off on my estimates since I didn't really pay attention to a lot of the mastering and monitoring time since I listened to each track in the car/headphones, etc. before tweaking things to fit all environments which wouldn't be possible in a studio environment. In any case, I probably saved $27k by doing it all at home with Logic (yeah some studios includes instruments you can use, etc. but I had most of these for some time over the years anyway; you have to write and practice on something).

Of course, the only thing you really need to do professional quality audio recording is a Firewire 400 or 800 port (to plug in a multi-port audio/midi card), a good DAW (I use Logic Pro) and enough power to handle the number of real-time channels + effects you'll need (plus whatever music gear, monitoring speakers and optionally hardware mixing/mastering consoles; I just used a mouse and the keyboard here and it worked fine, if a bit of extra work).

My old 2008 Core2Duo with an upgraded 500GB 7200 RPM hard drive can typically do over 20 tracks at once with a lot of effects and probably at least twice that for just straight recording. I recorded all the parts myself so I went with a Pre-Sonus interface which only has 4 inputs at once, but I'm only using one at a time typically anyway. I have a separate MidiMan MIDI interface for synth controller input. If I did it over again, I'd get an all-in-one unit even if it cost more. The MidiMan has had driver issues for awhile now (I typically have to unplug/plug it once in awhile after waking from sleep to get the odd lag out. The PreSonus has performed flawlessly. Sound quality on my album turned out similar quality to say Wish You Were Here by Pink Floyd without all the analog effort). Logic pretty much went from recording to editing/effects to mixing and mastering in one package.

My point is that a 2008 MBP was more than adequate for the task of creating a Pink Floyd style album from scratch using loads of electric guitar effects and software synths in addition to acoustic guitar and mic inputs. I can only imagine how many channels the new Mac Pro could potentially handle with even 6-cores and those SSD drives. I'd imagine 80 tracks at once with 4-6 effects each on average would be no problem for it and that's more tracks than anyone on earth should need for a given song, IMO, let alone at the same time. I doubt I ever had more than 12 tracks playing at the same time at any given moment. The point is that the Mac Pro would probably be overkill. My 2012 quad-core i7 Mac Mini would blow away the MBP for recording save the fact I have a mobile setup and can record just about anywhere with it.

 

[milo]

you - "$1m is not necessary to produce high quality work"
me - "yeah, I know"
you - "your argument is wack that audio work requires that type of investment"
me - "right. I understand that"

"Yeah I know"? Then why do you keep saying just the opposite?

sure you do.. more than a million dollars in many locations.

.. there's no shame in saying smthng like "I produce music which demands high quality" (or whatever)..

Why would there be, that's what I keep saying.

(seemingly) defensive

Not defensive at all, just disagreeing with the notion that anyone doing high end music production has a million dollar studio so the cost of the computer is just a rounding error. But way to try and change the subject.

Thanks for making clear that you've never actually done audio work, in a "high end studio" or otherwise. It makes sense that you haven't the foggiest idea what audio production actually means or what resources it actually requires.

 

[wallysb01]

lol. I'm not even arguing anything.

you - "I do high end audio production"
me - "oh.. then you don't have to worry about minuscule costs such as computers"
others - "you're out of touch ..$1000 is not something to be taken lightly"
me - "I know.. I was being facetious and pointing out a semantics issue with the term "high end"
you - "$1m is not necessary to produce high quality work"
me - "yeah, I know"
you - "your argument is wack that audio work requires that type of investment"
me - "right. I understand that"
you - "you just keep repeating the same argument over and over"
me - "hmmm"

Oh, so basically you were trolling, or you're just making this story up at this point because you can't back up what you said any longer. I can't really tell which, but I won't call you a liar.

what's there left to discuss? I was saying you should use window for at least the next couple of years.. you're saying you're doing exactly that.
we're in agreement, right?

Well, there was that whole xMac thing. I guess you figured out you were wrong? There was that part about you trying to tell me 2xd700 is a cluster...... But I guess you figured you got that wrong too?

 

[flat five]

Not defensive at all, just disagreeing with the notion that anyone doing high end music production has a million dollar studio so the cost of the computer is just a rounding error. But way to try and change the subject.

what's the subject? that is a real question..

you're literally sitting here arguing me post after post that it doesn't require 1 million dollars to produce music..

let me ask you a question.. this is a yes or no question so please answer either 'yes' or 'no':

do you honestly believe i think it costs $1m (or anywhere remotely close) to buy equipment in order to produce high quality audio?

 

[flat five]

Oh, so basically you were trolling, or you're just making this story up at this point because you can't back up what you said any longer. I can't really tell which, but I won't call you a liar.

no, i can back it up.. if you set out to build a high end studio (as most people in the industry or related industry would interpret 'high end studio') it's going to cost you 7figures (location dependent-- as in- if you plan on doing this in nyc, you're looking at a possible $10m)

if you're looking to build a low end home studio then you're not looking at anywhere near that amount.. your design fees, real estate, and construction costs are relatively nonexistent in that scenario..

Well, there was that whole xMac thing. I guess you figured out you were wrong? There was that part about you trying to tell me 2xd700 is a cluster...... But I guess you figured you got that wrong too?

lol..

the xmac -> you're right! they exist and i'm a fool for suggesting apple won't build one.. i'm such an idiot.

2xd700 = 4000 processing cores -- ie- cluster

 

[wallysb01]

no, i can back it up.. if you set out to build a high end studio (as most people in the industry or related industry would interpret 'high end studio') it's going to cost you 7figures (location dependent-- as in- if you plan on doing this in nyc, you're looking at a possible $10m)

True or not (this isn't my field, but it sounds like you have people that would dispute you), spending 7 figures doesn't mean you should waste 4 figures. Incase you forgot the point isn't that you can spend alot of money on audio work. The point is spending a lot of money, doesn't justify frivolously spending what seems like a lot less money. That's why you can't back it up. There is always an opportunity cost. Do you know what that is?

Also, think of this way. Many businesses see profit margin in the single digit ranges. So if they spend, say, $2M, then lets say profits are 4% (after equipment, salaries, etc.) that is "only" $100K for the owners. Suddenly, frivolously spending a couple grand extra on something that doesn't increase your profits doesn't sound so small does it? Instead of being 2K/2M (.1%), its really 2K/100K (2%). You don't have to make that kind of mistake very many times for it to add up.

the xmac -> you're right! they exist and i'm a fool for suggesting apple won't build one.. i'm such an idiot.

You are acting like quite the idiot, yes. You know that point wasn't if they exist or not.

2xd700 = 4000 processing cores -- ie- cluster

Case and point. 4000 GPU cores a cluster is not. Even going by the somewhat crude Tflops 4000GPU cores is nothing like 4000 CPU cores (which would be a very small cluster at that). For example, just looking down the top 500 list, one IBM cluster has 7560 CPU cores using the 2680v2 (so 756 of them since they are 10 core processors), and it gets 130TFlops, divide roughly in two, and 4000 modern Intel CPU cores is 65TFlops. The new Mac Pro gets about 7TFlops. Its impressive what GPGPU can do if your software can use it, but that's big if for many folks. Plus, you should read up on the short comings of just using TFlops as the measure of computation. Basically, its testing for tests that are great on GPUs and don't hit bottle necks or play impossibility due to the nature of a GPU. BTW, that cluster is 413th on this top 500 list. Respectable, but its not hard for a researcher like myself to gain access to many clusters much bigger than that.

Oh, and guess what, a lot of clusters have those 4000 GPUs per node! So they have maybe 12, 16, 20 or 24 CPU cores, PLUS those 4000 GPU cores on each node and they have 4000 nodes.

Your ignorance is shinning. I know you think you're right and there is probably nothing I can say to make your understand you're wrong, so I'm just going let your willful ignorance continue. The Mac Pro might have the computational power of a cluster 10 years ago (and would rank around 7th IIRC), which is awesome when you step back and think about how far we've come, but compared to today's clusters, meh.... no so much.

 

[flat five]

True or not (this isn't my field, but it sounds like you have people that would dispute you), spending 7 figures doesn't mean you should waste 4 figures. Incase you forgot the point isn't that you can spend alot of money on audio work. The point is spending a lot of money, doesn't justify frivolously spending what seems like a lot less money. That's why you can't back it up. There is always an opportunity cost. Do you know what that is?

yes. i know what that is..

and i think it's pretty evident in the situations i've (attempted to) outline earlier.. even amongst the megarich.

the people i'm familiar within the recording industry are using midrange computers (iMacs & mbps) in their personal studios and are recording albums with mac pros.. (well, they're not actually recording the albums.. they're paying pros to do it and those pros are generally using mac pros)..

which translates to this.. the personal studios (experiments/songwriting/rehersal/etc) are using computers which are suited to their needs and they're not overspending on computers even though they have millions of dollars..

this, to me, is an example of smart spending even though the extra $grand appears as if it may be a drop in the bucket.

but, at the same time, the pro studios are also spending smartly when purchasing a mac pro over an imac even though both scenarios require the same type of computational power..

how so? because the mac pro is more of a workhorse and less likely to break down or slug during peaks.. and it's being used during time_is_money times in which a hold up could end up costing much more than the $1000 they apparently overspent on..

a personal analogy from my own work is this.. i've donated time/skill/labor/etc quite a few times on down_for_the_cause type scenarios.. in situations like that, i'll rent a 1ton forklift in order to offload/relocate/assemble 1ton stacks of wood or components..

however, when i'm on a contract job, i always rent 5000lb lifts to move those same sized stacks because i don't want to be sitting there dicking around or getting stuck on a rock or navigating a slight slope (or whatever) because i'm operating at max capacity.. i want more power than necessary so everything runs as smooth as possible.

what i'm saying is that opportunity costs can't (or shouldn't) always be viewed based on a price tag alone.. there's often more to be considered in the equation in order to arrive at the smartest decision because sometimes, the best decision is to pay for a more expensive item up front even if it could be considered as overspending within a singular line of thought.

(how this relates to this thread in particular, i'm not sure.. because i have a good feeling that people are reading the above as me saying "hey, you-- you should buy a mac pro and don't worry about the cost".. because i'm really not saying that at all.. i'd much rather see everybody in the thread make smart purchases that suit their needs or even wants in this case-- somehow, that is getting lost in this mashup of a conversation)

Also, think of this way. Many businesses see profit margin in the single digit ranges. So if they spend, say, $2M, then lets say profits are 4% (after equipment, salaries, etc.) that is "only" $100K for the owners. Suddenly, frivolously spending a couple grand extra on something that doesn't increase your profits doesn't sound so small does it? Instead of being 2K/2M (.1%), its really 2K/100K (2%). You don't have to make that kind of mistake very many times for it to add up.

yeah.. i get that.. but you're talking to me as if i've been saying "oh.. it's only money.. buy a mac pro.. you'll love it"

You are acting like quite the idiot, yes. You know that point wasn't if they exist or not.

i don't really have much of a choice other than just turn it into fun&games..
my points or anything i have to say on the matter are simply being ignored by everyone i've been talking to.

and it starts getting really weird because the gist of the arguments when combined come across as "it's stupid to expect developers to adopt/incorporate openCL in their software but it makes perfect sense to sit around and expect apple to create a new line of desktop computers".. there's a lack of consistency depending on which angle people try to call me stupid from..

how about this- just so i know we're clear.. instead of you telling me what you think about the xmac, repeat back, in your own words, the things/views i have on the idea.

as in- you (you personally) would like apple to create a computer which lays off the apple flair -or- puts more emphasis on function than it does form.. it's xeon based as you want the memory capabilities of a xeon and it (likely) has at least a couple of pcie slots.. one of which accepts a gpu of your choice.. i think you'd still want it to have a couple of thunderbolt ports and usb3 (i.e.- the i/o has been updated beyond the mp1).. i imagine you'd want more than a single drive inside and would probably be ok with it having two bays but wouldn't mind it having four..
the base model is a quad with maybe 8GB ram, an acceptable but nothing too fancy gpu, and 128GB ssd which starts at, say, $1200..

does that sound about right?

so what have i been saying about the xmac?

4000 GPU cores a cluster is not.

it is a cluster if you view it as intended.. that being- we're talking about personal computers and and not some enterprise situation where a true render farm is going to set me back $100,000

because i can certainly utilize the power of a renderfarm but there's no way i can justify buying one.. gpgpu coupled with (relatively) lots of vram/cores allows single seat to small studio users to have (the idea of) a cluster inside their computers..

if you think i'm ignorant for seeing it like that then so be it.. when i'm cranking out renders next year in 10minutes which previously took overnight and you say to me "whatever bro.. that's not even a real cluster" ..do you think i'm going to care?

 

[milo]

do you honestly believe i think it costs $1m (or anywhere remotely close) to buy equipment in order to produce high quality audio?

If you don't believe that, why did you say that in the first place? That has been the main basis of your whole argument for post after post.

no, i can back it up.. if you set out to build a high end studio (as most people in the industry or related industry would interpret 'high end studio') it's going to cost you 7figures (location dependent-- as in- if you plan on doing this in nyc, you're looking at a possible $10m)

And there it is again. I'm baffled by your posts, you ask if you think that, then follow it up by saying the same thing again. WTF?

if you're looking to build a low end home studio then you're not looking at anywhere near that amount.. your design fees, real estate, and construction costs are relatively nonexistent in that scenario..

Still clueless about music production. Really, the only ways to produce music are "high end" for a million (or now ten million) or "low end" in the basement?

the people i'm familiar within the recording industry are using midrange computers (iMacs & mbps) in their personal studios and are recording albums with mac pros.. (well, they're not actually recording the albums.. they're paying pros to do it and those pros are generally using mac pros)..

Well that's a new spin on your argument. Before you were making it sound like those guys were doing albums with iMacs and laptops. Now we hear that the less powerful machines are just in their personal studios.

but, at the same time, the pro studios are also spending smartly when purchasing a mac pro over an imac even though both scenarios require the same type of computational power..

Your ignorance of audio production shines through again. If someone is doing demos in a home studio, that doesn't require the same power as a final production mix. And "less likely to break down or slug" is nonsense, there's no difference in reliability in a recording studio situation. There are two reasons to use a mac pro (previous generation) for audio production. One is to get more processing power (six or more cores), and the other is to use PCIe cards like Pro Tools (which often include DSP chips to provide even more processing power). Those "million dollar studios" are the ones most likely to be using PT PCIe hardware, and those are the guys most negative about the new Mac Pro since it won't handle their cards without expensive external hardware. You won't see nMP in those studios any time soon, not for price reasons, but because the new machines aren't compatible with their hardware (and that particular hardware is way more expensive than the computer itself).

it's stupid to expect developers to adopt/incorporate openCL in their software

Someone said that? Who? Quote please?

What I have been saying is that the algorithms used in audio processing can't be translated to OpenCL. The software simply doesn't translate to that particular hardware.

Let me fix that statement for you:

it's stupid to expect developers to adopt/incorporate openCL when it's a technical impossibility

...it's xeon based as you want the memory capabilities of a xeon...

Who said they want that? High end i7 can handle 64 gigs of ram and that would be enough for me in a true midrange machine.

Has a single person here said that they want a xeon based midrange mac? Has anyone said they need a midrange mac with more than 64 gigs of ram? Who?

when i'm cranking out renders next year in 10minutes...

Oh, so now you're actually going to buy one of these? That will be a nice change of pace from being the guy arguing about hardware you don't own and haven't used. Be sure to report back what configuration you ended up getting, and of course what software you're running on it.

 

[flat five]

If you don't believe that, why did you say that in the first place? That has been the main basis of your whole argument for post after post.



And there it is again. I'm baffled by your posts, you ask if you think that, then follow it up by saying the same thing again. WTF?

you're talking as if the only thing necessary to make a studio is a computer.. in your scenario, i understand this is all you need to be concerned with.. ok- are we clear on that?

but now, you have to listen to what i'm saying and quit mashing everything together.. if you set out to build a high end studio, the computer becomes maybe the 100th most important or most expensive thing.. for instance, buying acoustic wall tiles:
http://www.amazon.com/Acoustic-Wall-Tile-12-Box/dp/B0002ZPTVS
..for a 20x20 room costs more than a mac pro.. you're looking a $6000 just for wall coverings.. i could give you 50 more examples of this and once you read through the entire list, it should become perfectly clear that when you set out to build a high end studio, the cost of a computer or the gap in apple's lineup becomes the least of your worries.

i'm sorry, if you can't understand what i'm saying then so be it.. i'm not going to explain it to you again because it's something that should be easily grasped by most people.. especially someone in the high end sound production business... so if you don't grasp it then i'll just sit back and find humor in the fact that you think someone is trying to convince you that "computers for high end studios cost $1,000,000"

Still clueless about music production. Really, the only ways to produce music are "high end" for a million (or now ten million) or "low end" in the basement?

and everywhere in between.. try to move your studio off your couch and you're going to find out real quick how much studios actually cost..


------------------------------------

Well that's a new spin on your argument. Before you were making it sound like those guys were doing albums with iMacs and laptops. Now we hear that the less powerful machines are just in their personal studios.

aside from the part about me saying this:

another rock star (that i actually put to work while he paid me )
same thing though.. imacs and macbooks in the home studio and macpros in the main studio..

in the very first post i made about those people.. sure, good point #

----------------------------------------------

Your ignorance of audio production shines through again.

and your mastery amazes me..

Someone said that? Who? Quote please?

are you even reading the thread?

What I have been saying is that the algorithms used in audio processing can't be translated to OpenCL. The software simply doesn't translate to that particular hardware.

great.. now you're an audio software developer too.

Who said they want that? High end i7 can handle 64 gigs of ram and that would be enough for me in a true midrange machine.

again, i can help but conclude that you're not even reading the thread.. who said that? the exact person i was talking to in the post you quoted..
do you really not realize that everybody's 'xmac' is different? all of you people keep saying xmac as if it's one singular idea when in fact, if there are 50 people saying xmac, there are 50 different xmacs.. i don't know of a way to get you to read what's actually being written here so i won't try.

Has a single person here said that they want a xeon based midrange mac?

lol.. yes. this is getting ridiculous

Has anyone said they need a midrange mac with more than 64 gigs of ram? Who?

hmm.. wally.. the same guy you're teaming up with to tell me how ignorant i am.

Oh, so now you're actually going to buy one of these? That will be a nice change of pace from being the guy arguing about hardware you don't own and haven't used. Be sure to report back what configuration you ended up getting, and of course what software you're running on it.

jesus christ man.. i feel like i'm in the twilight zone or something.. didn't you already ask me this 2 days ago and haven't i already answered?

 

[wallysb01]

again, i can help but conclude that you're not even reading the thread.. who said that? the exact person i was talking to in the post you quoted..
do you really not realize that everybody's 'xmac' is different? all of you people keep saying xmac as if it's one singular idea when in fact, if there are 50 people saying xmac, there are 50 different xmacs.. i don't know of a way to get you to read what's actually being written here so i won't try.

This is bs. If there are 50 different people talking about an xMac, there are about 2 different types of computers that could cover them all, maybe 3. For almost everyone, the xMac is the “box-with-slots” desktop, just like every other PC, just with OSX and prettier. For others it would be the “box-with-slots” entry level workstation.

hmm.. wally.. the same guy you're teaming up with to tell me how ignorant i am.

I’m not the typical person talking about an xMac.

 

[flat five]

I’m not the typical person talking about an xMac.

yeah.. nobody is.

you guys are talking as if you know best how to develop a line of computers which will accommodate the vast majority of people.

are you really so blind to the fact that apple has much much more data/usage scenarios/spending limits/etcetc to analyze and have in fact been analyzing this data for over 30 years..

and that they've developed 3 separate products to accommodate everybody.

they already made the xmac.. it's the mac mini/imac/mac pro.. ($1000/2000/3000)

that's their solution.
if you don't like it, gtfo.
you see how that works?

 

[wallysb01]

how so? because the mac pro is more of a workhorse and less likely to break down or slug during peaks.. and it's being used during time_is_money times in which a hold up could end up costing much more than the $1000 they apparently overspent on..
...

what i'm saying is that opportunity costs can't (or shouldn't) always be viewed based on a price tag alone.. there's often more to be considered in the equation in order to arrive at the smartest decision because sometimes, the best decision is to pay for a more expensive item up front even if it could be considered as overspending within a singular line of thought.

(how this relates to this thread in particular, i'm not sure.. because i have a good feeling that people are reading the above as me saying "hey, you-- you should buy a mac pro and don't worry about the cost".. because i'm really not saying that at all.. i'd much rather see everybody in the thread make smart purchases that suit their needs or even wants in this case-- somehow, that is getting lost in this mashup of a conversation)

We’re getting lost in how this started, I believe. You’re now claiming that it can be wise to seemingly overpay on a Mac Pro vs. an iMac because the Mac Pro will be faster, and deliver projects faster which turns into money. But the issue is that if the iMac can work, there is almost certainly a PC workstation that will deliver the same, if not better performance boost for cheaper than the Mac Pro. The only place the nMP has an advantage over PC workstations in performance or performance/dollar is with openCL and OSX specific apps. If you can do your work on Windows and you tasks are CPU bound, the better deal is in a PC. You can get a 6-core PC for almost half the price as the 6-core nMP. Or for the price of the 12 nMP, you could have a 20 core duel socket PC workstation.

That’s where the overpay really sits. Not between the iMac and nMP, but nMP and PC workstations.

yeah.. i get that.. but you're talking to me as if i've been saying "oh.. it's only money.. buy a mac pro.. you'll love it”

i don't really have much of a choice other than just turn it into fun&games..
my points or anything i have to say on the matter are simply being ignored by everyone i've been talking to.

If you’re going to sarcastic on line, you had better be ready for people think your serious. Its hard to read that sometimes, especially when so many things you are saying in all seriousness also seem so completely idiotic.

and it starts getting really weird because the gist of the arguments when combined come across as "it's stupid to expect developers to adopt/incorporate openCL in their software but it makes perfect sense to sit around and expect apple to create a new line of desktop computers".. there's a lack of consistency depending on which angle people try to call me stupid from..

I don’t see that at all.

how about this- just so i know we're clear.. instead of you telling me what you think about the xmac, repeat back, in your own words, the things/views i have on the idea.

Basically, you’re saying its hopeless for Apple to attempt to make an xMac because everyone has a different idea of what an xMac should be. Therefor, even if Apple does make something one might call and xMac, another person will still be asking for an xMac.

That would be true to an extent. No computer will please everyone. But Apple could make a computer to please at least most of those asking for an xMac, if not an overwhelming majority.

it is a cluster if you view it as intended.. that being- we're talking about personal computers and and not some enterprise situation where a true render farm is going to set me back $100,000

because i can certainly utilize the power of a renderfarm but there's no way i can justify buying one.. gpgpu coupled with (relatively) lots of vram/cores allows single seat to small studio users to have (the idea of) a cluster inside their computers..

Most people that use clusters don’t buy clusters, they "rent" time on them. And you can have the idea of a cluster on your desk, but its not a cluster. Its hype. The IO, the RAM, the TFlops, the versatility, are nothing like a modern cluster.

Ultimately, I suppose, this is semantics. You want redefine the meaning of cluster to fit this new GPGPU in a single computer thing, just because there are lots of cores in there that, in some cases, can get a tremendous amount of work done in a highly parallel way, similar to a cluster. But its just not a cluster. To believe that, is to swallow the GPGPU hype hook, line and sinker.

if you think i'm ignorant for seeing it like that then so be it.. when i'm cranking out renders next year in 10minutes which previously took overnight and you say to me "whatever bro.. that's not even a real cluster" ..do you think i'm going to care?

Cool. I’m happy you’re getting your work done quickly. Sometimes the speed of progress goes beyond Moore’s law for lots of reasons. But its not because you suddenly got a cluster in your personal computer. If it feels like that to you, great, but it doesn’t make it true. Sorry.

----------

yeah.. nobody is.

Believe it or not, yes, there is a common ground for the vast majority.

you guys are talking as if you know best how to develop a line of computers which will accommodate the vast majority of people.

Not the vast majority of people. The vast majority of people wanting an xMac.

are you really so blind to the fact that apple has much much more data/usage scenarios/spending limits/etcetc to analyze and have in fact been analyzing this data for over 30 years..

They have made choices to abandon the xMac idea for a lot of good reasons. That doesn’t mean they aren’t leaving a specific class of user behind that can’t be defined by one or two computers that would suite them.

and that they've developed 3 separate products to accommodate everybody.

Just no. Not “everybody”, nothing, nohow, noway will ever accommodate “everybody”.

they already made the xmac.. it's the mac mini/imac/mac pro.. ($1000/2000/3000)

Pure silliness.

that's their solution.
if you don't like it, gtfo.
you see how that works?

Exactly, that’s the thing. You just proved my point, its that gtfo part. If you’re too blind to see it, so be it.

 

[flat five]

That would be true to an extent. No computer will please everyone. But Apple could make a computer to please at least most of those asking for an xMac, if not an overwhelming majority.

you need to put this xmac thing into perspective..
this is a mac centric forum with nearly a million members.

far less than 1% of people here are asking for an xmac.

so you're suggesting apple build a new computer to please an overwhelming majority of 1%.. so maybe .8% of people will be satisfied with this wonderful new xmac.


i know it sounds so smart to you but step out of your bubble and look at it from afar.

 

[milo]

you're talking as if the only thing necessary to make a studio is a computer...

Again, it's all black and white to you. I said you don't need to spend a million dollars and what you hear is "the only thing necessary is a computer".

You list wall tiles and room sizes. Really, hanging out with guys who do recording makes you an expert on what is necessary for audio production? Thanks for enlightening those of us who do audio production for a living, it's an eye opener to find out it's not "just a computer".

Just use common sense. You're actually trying to convince people that "high end audio" can't be done unless you have six thousand dollars worth of wall tiles!!!!. That's going to come as quite a shock to all the guys doing film, tv, commercial, game scores in rooms that don't have six thousand dollars worth of wall tiles!!!!

Don't get me wrong, I know that there are studios with six thousand dollars worth of wall tiles!!!! and what the advantage is of six thousand dollars worth of wall tiles!!!! But I also know from actual real life experience that not having six thousand dollars worth of wall tiles!!!! is not going to magically make an audio production not "high end".

if you can't understand what i'm saying

I understand, you have no experience or actual knowledge of audio production. I understand what you're saying, you hung out in a studio a couple times and you're spouting off what you saw there.

"computers for high end studios cost $1,000,000"

More straw men, nobody has said anything like that in the thread. But I guess that's all you have to offer, no real case to make, just putting words in everyone else's mouth.

are you even reading the thread?

So humor me. Where's the quote.

...now you're an audio software developer too.

And you are? I'm just passing along what I have read from audio software developers. I have yet to see anything from a single one disagreeing with that. If you dispute that OpenCL isn't suited for audio, you need to do it with facts.

the exact person i was talking to in the post you quoted..

RAM is capped at 32GB, which is good but ideally it would be at least 64GB if not 128

So he said 32 is good, ideally at least 64GB. i7 can do 64GB. You read that as he needs more than 64 and thus needs xeon? And you keep saying I'm not reading the thread?

do you really not realize that everybody's 'xmac' is different?

Not really that different, you're exaggerating because you need to to try and make your point. You can never make 100% of people happy but there's a sweet spot that would please a hell of a lot of the folks wanting an xMac.

hmm.. wally..

Wally mentions 1650v2. There is an i7 available with the same specs about the same price. Somehow if apple shipped an xMac with i7-4930K at a fair price I doubt he'd consider "not a xeon" to be a deal breaker. But he can clarify for himself.

haven't i already answered?

Looking back, I guess you did. Sorry I missed the actual answer, I didn't make it to the end of that paragraph of bloviation. I look forward to hearing about your experience once you actually have one.

 

[flat five]

Exactly, that’s the thing. You just proved my point, its that gtfo part. If you’re too blind to see it, so be it.

i'm not blind.. i don't know how much simpler/smarter this equation can be from a consumer's pov.

if you don't like a company's product.. don't buy it.

the fact that you're sitting here arguing me over this and telling me how stupid i am.. well.. so what.
that's the way i shop. that's the way most people shop. it's not even worth arguing about.

 

[milo]

To address the topic of OpenCL for audio, here's some commentary on it from the programmer behind the fantastic Valhalla reverbs. Feel free to dispute it with contrasting opinion from another audio software developer. Also, it seems like convolution may translate well to GPU hardware but that is only used in some reverbs which leaves out most other kinds of audio processing.

I think the odds of me porting to CUDA are about zero.

The reason for this: GPUs are geared towards massively parallel processing. This is the sort of thing that you see in images, which are essentially NxM arrays of data. Image processing tends to be feedforward, and not reliant on previous outputs of the filters.

Most of the audio processes I work on are based on delays, feedback processes and filters, that sort of thing. This doesn't work well in massively parallel systems. I process audio in parallel whenever I can get away with it, for efficiency, but the fact is that delay lines of different lengths are inherently difficult to parallelize. Feedback doesn't work well with massively parallel systems, unless you have feedback with pretty big blocks, and that tends to lose a lot of the cool aspects of feedback.

Convolution processing (including Nebula) works well with massively parallel processing. Beyond that, I could see GPUs used for other things that work well with lots of parallelism, like modal synthesis. For general audio applications, I don't think that GPUs will prove that useful. I wouldn't mind being proven wrong, as there is a LOT of power in the GPU.

----------

...people here are asking for...

Bzzt. Idiotic to try and pretend to know what the general public wants based on postings on an internet board.

 

[flat five]

Again, it's all black and white to you. I said you don't need to spend a million dollars and what you hear is "the only thing necessary is a computer".

You list wall tiles and room sizes. Really, hanging out with guys who do recording makes you an expert on what is necessary for audio production? Thanks for enlightening those of us who do audio production for a living, it's an eye opener to find out it's not "just a computer".

Just use common sense. You're actually trying to convince people that "high end audio" can't be done unless you have six thousand dollars worth of wall tiles!!!!. That's going to come as quite a shock to all the guys doing film, tv, commercial, game scores in rooms that don't have six thousand dollars worth of wall tiles!!!!

Don't get me wrong, I know that there are studios with six thousand dollars worth of wall tiles!!!! and what the advantage is of six thousand dollars worth of wall tiles!!!! But I also know from actual real life experience that not having six thousand dollars worth of wall tiles!!!! is not going to magically make an audio production not "high end".



I understand, you have no experience or actual knowledge of audio production. I understand what you're saying, you hung out in a studio a couple times and you're spouting off what you saw there.



More straw men, nobody has said anything like that in the thread. But I guess that's all you have to offer, no real case to make, just putting words in everyone else's mouth.



So humor me. Where's the quote.



And you are? I'm just passing along what I have read from audio software developers. I have yet to see anything from a single one disagreeing with that. If you dispute that OpenCL isn't suited for audio, you need to do it with facts.





So he said 32 is good, ideally at least 64GB. i7 can do 64GB. You read that as he needs more than 64 and thus needs xeon? And you keep saying I'm not reading the thread?



Not really that different, you're exaggerating because you need to to try and make your point. You can never make 100% of people happy but there's a sweet spot that would please a hell of a lot of the folks wanting an xMac.



Wally mentions 1650v2. There is an i7 available with the same specs about the same price. Somehow if apple shipped an xMac with i7-4930K at a fair price I doubt he'd consider "not a xeon" to be a deal breaker. But he can clarify for himself.



Looking back, I guess you did. Sorry I missed the actual answer, I didn't make it to the end of that paragraph of bloviation. I look forward to hearing about your experience once you actually have one.

i love you. wally too.
bye