Will there be a new Mac Pro in 2015?

[H2SO4]

I now believe the nMP will come only next year with SkyLake, TB3 and Fiji.
And as much as I would want one now, I rather wait and have a better machine, since DMI3, PCIe 3 on the PCH will allow a much better use of the hardware that can even be further tweaked.
Think more available PCIe3 lanes on the PCH means no internal switches for TB3 (although DMI could be a bottleneck if connected to PCH instead of CPU available lanes), no compromises with internal drives (full speed NVMe drives, possibly even more than 1 internal drive - 2 for space constrains).

I’d be very disappointed if that’s the case. I’ve wanted to buy this machine since day 1 to replace my Mac Pro 1,1. Instead I bought a 6 core from Create Pro as the nMP was just too many compromises for me. I want a little bit of future proofing and versatility at least.

Tim if you’re LISTENING, not hearing - listening, (there is a difference). I’d like you to upgrade HDMI to revision 2. Fit DDR4. Make dual drives an option. More GPUs and option with Mac support for Xfire etc. TB2 i fine, USB3 is fine also but a single USBC port.
You would then have another few thousand dollars on your bottom line.

 

[Demigod Mac]

Best case scenario for 6,1 is that Apple offers the 7,1's video cards as upgrades for outrageously expensive prices. They issue a machine ROM update at the same time the new stuff comes out, essentially locking 3rd parties (OWC, Sonnet) out of the mix unless they come up with some clever solution.

Then they don't touch the line again for another 1.5 years.

 

[flat five]

You say this as if a dual CPU machine couldn't also house multiple GPUs. And you also say it as if GPU computing is the be all end all answer. It's not and you know that.

i said it as it was written at face value.
i was responding to mta2 who said pros want 36core rack mounted etc and i'm seriously questioning if that's what pros really want.

asking the question again that no one has answered:
do you really think they want 36 core computers? what if they had a choice between a behemoth (relativiely) of a machine or something small and quiet and efficient that puts out vast amount more computational power than possible with 36 cpu cores.. or a hundred cpu cores even?

gpgpu isn't the endallbeall.. especially currently.
but cpu based computing certainly isn't the answer.. it's way too expensive.
slow software is exactly that.. slow software. insane amounts of computing power are sitting idle in the nmp (or other computers) when using legacy software..

the general software mantra to date has been "if you want our software to run faster, buy faster computers".. that was fine 20years ago when processors were getting noticeably faster while staying moreorless same price. now, processor speed isn't developing a quick and the cost for multiple cores has an exponential increase in cost.

if i want a renderer to run at optimized speed, i need 96cores.. ?? yeah, i need $100,000 worth of equipment.. the software is just fine and is optimized perfectly.. it's the hardware that's the problem??

------------
and that gets at the gist of much of my issues with this particular forum.. so many people here talk about what pros need etc and only a handful of people who visit here are actually working pros.. only like 1 of the prominent naysayer voices are pros.. why can't people just be straight up honest about their computer usages? the conversations would be completely different then.. why can't MacVidCards admit "the nmp is way more computer than my needs dictate" instead of talking about this mythical pros out there and how apple has let them down?

the forum is a bunch of hobbyists who like running geekbench and getting nerdy about computer hardware.. that's fine. that's fun etc.. but if everybody could just admit what's going on then much of the 'what pros need.. what pros want' bs would disappear.

And it appears there are people who want these massive machines or I wouldn't be getting promotional emails for these sent to me all of the time:

http://www.boxxtech.com/products/apexx-5

http://www8.hp.com/us/en/campaigns/workstations/z840.html

Not to mention the other brands as well.

how many people buy those things? how many people here, if apple were to build the dream machine or whatever, would actually buy one? you're talking about $25,000 computers.. it's just stupid to think that's what pros want..

the computer isn't what's important.. the software is.. the computers are fine right now.. they're more powerful than necessary to get the job done. you want serious gains in computing speed? feed it better/more efficient algorithms.

apple knows way way more about how pros use their computers.. everybody loves to think apple is out of touch.. apple abandoned their pros.. an onandon..

if anybody knows exactly what resources in the computer are being used by the typical creative pro and how they use their computers, wouldn't apple know this info the most? it's their customers.. they receive tons of feedback both in conversation and via hardware analysis being sent to them..

it's borderline ludicrous to think 'apple abandoned their pros with this machine' and 'i know what's best for pros'.. if they abandoned their pros with the nmp, who did they make this machine for?

i get it that the answer will be ala 'for facebookers with more money than they know what to do with' etc but come on.. it's nearly impossible to imagine a scenario where apple engineers (or any computer engineers) would sit around the meeting going 'ok.. let's build an unusable computer, call it a pro machine, and sell it to rich people who will use it in the same way they'd use an ipad.. that's a great business plan"

You're right in that there's no room for a 2nd CPU. It's not happening and anyone waiting it out for Apple to go that route should abandon hope now. But you're still touting that "why not more" BS? As has been pointed out, 2 is kind of the magic number for a "desktop" solution and bang for your buck. But if we were to use your argument, why only 2 GPUs? 4 is surely better than 2, correct?

the point with using the 'why not more' example is that it's an avenue with no end in sight.. just keep building bigger and more expensive computers in order to increase speed? it's a dead end street and demanding that type of solution is pricing yourself out of more powerful computing.

i want affordable computing.. sub $6000.. with hundreds of times the power as what was available last decade..

not- expensive computing.. $20,000+ with twice the power as what was available last decade.

does that not make sense?

 

[handsome pete]

how many people buy those things? how many people here, if apple were to build the dream machine or whatever, would actually buy one? you're talking about $25,000 computers.. it's just stupid to think that's what pros want..

What makes you the arbiter of the "pros?" And stop using that word. It's lost any meaning it once had. I remember working on a couple of smaller productions back when HD was first getting traction. The DP had just bought a new HD camera for just over $100K. So yeah, I don't think $25K for a computer is outrageous. You're also using the extreme to make an example. There are plenty of different options at lower cost.

the computer isn't what's important.. the software is.. the computers are fine right now.. they're more powerful than necessary to get the job done. you want serious gains in computing speed? feed it better/more efficient algorithms.

Software often lags way behind the hardware. But if we're bound to the software that is readily available, then I want to use the hardware that will offer the most benefit.

if anybody knows exactly what resources in the computer are being used by the typical creative pro and how they use their computers, wouldn't apple know this info the most?

Wouldn't HP? Wouldn't Dell? Wouldn't Boxx?

If anything, it's been kind of the opposite. Historically, Apple has been the company that has told its customers what it wants, not the other way around. And I'm not even saying this is always a bad thing.

it's borderline ludicrous to think 'apple abandoned their pros with this machine' and 'i know what's best for pros'.. if they abandoned their pros with the nmp, who did they make this machine for?

The problem is your definition of "pros." That word is meaningless anymore. It covers everyone from some Pixar TD working on their next film to your distant cousin making some extra cash shooting bar mitzvahs on Saturdays. I don't think anyone truly knows who this machine is for. I can just attest to what I've personally witnessed. Since its release I've only known one person to get a nMP. An agency I worked with has migrated to Premiere and is looking at HP workstations once they squeeze everything out of their old Mac Pros. The university I teach at is no longer recommending Apple computers for their VFX and Animation students. Regular colleagues of mine, the ones who are most like the regulars on this forum who kneel at the altar of Apple, are buying iMacs or Macbook Pros.

i get it that the answer will be ala 'for facebookers with more money than they know what to do with' etc but come on.. it's nearly impossible to imagine a scenario where apple engineers (or any computer engineers) would sit around the meeting going 'ok.. let's build an unusable computer, call it a pro machine, and sell it to rich people who will use it in the same way they'd use an ipad.. that's a great business plan"

I don't think that at all. I think it's obvious that apple saw a dwindling market (workstation sales) and did something radical to reinvigorate it. They took a gamble. Who knows if it will ultimately pay off?

just keep building bigger and more expensive computers in order to increase speed?

What? Where are you getting computers being bigger and more expensive? We've been getting more power in the same footprint year after year, sometimes smaller even. And where has cost gone up?

i want affordable computing.. sub $6000.. with hundreds of times the power as what was available last decade..

not- expensive computing.. $20,000+ with twice the power as what was available last decade.

There are plenty of options in your "affordable" level. Not sure why you need to quote the extreme $20,000+ example. And again, the key word you used there is "I." Everyone has different needs.

 

[AidenShaw]

What $20K will buy....

does that not make sense?

No. The "twice the power" line is simply wrong, since a dual socket system has more than twice the available PCIe 3.0 lanes compared to the single socket system. More than twice.

More cores, more GPUs, faster GPUs - and 50 times slower than the MP6,1 ???

i want affordable computing.. sub $6000.. with hundreds of times the power as what was available last decade..

not- expensive computing.. $20,000+ with twice the power as what was available last decade.

• $6,499 - nMP - 8core Ivy Bridge EP- 3.0 GHz - 32 GiB - 256 GB SSD - 2*D700s - 1 yr warranty
• $5,244 - ML350g9 - 8core (2*4) Haswell EP - 3.0 GHz - 32 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $6,368 - ML350g9 - 16core (2*8) Haswell EP - 2.6 GHz - 64 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $7,242 - ML350g9 - 8core (2*4) Haswell EP - 3.0 GHz - 32 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty
• $11,624 - ML350g9 - 24core (2*12) Haswell EP - 2.6 GHz - 256 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $16,082 - ML350g9 - 32core (2*16) Haswell EP - 2.3 GHz - 256 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty
• $20,392 - ML350g9 - 36core (2*18) Haswell EP - 2.3 GHz - 384 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty

One of my ML350s with three TitanX cards, 28 cores and 768 GiB of RAM. (To add a 4th, I'd have to give up either the dual port 10 Gbps Ethernet card or the dual port 48 Gbps SAS RAID controller.)

 

[flat five]

What makes you the arbiter of the "pros?" And stop using that word. It's lost any meaning it once had.

hmm.. somewhere we've crossed wires.
that's what i was trying to say.

people keep talking about pros and their needs.. saying they need 36 cores and replaceable components all the time etc.. backtrack and you'll see me asking who are these 'pros' etc.. why don't people speak honestly of their individual needs in a computer and quit speaking for this hypothetical pro.
the conversations around here would be entirely different then.

Software often lags way behind the hardware. But if we're bound to the software that is readily available, then I want to use the hardware that will offer the most benefit.

somebody (devs or manufacturers) has to take the first step in order to make an industry wide radical change.. the developers aren't going to do it.. they're not going to change their code to run on a hypothetical computer.. if apple decides 'ok, let's make the hardware for these developers to design for' then i'm all for it.. further, they make an example for developers to strive for with fcp which runs very well on the nmp..
it's not like apple makes this new design then the entire industry makes a computer like theirs.. the computers the naysayers want are out there.. so many choices.. all workstations to date share the same design / theories. those choices are still out there and aren't going away in the near future.. is it really a big deal if there isn't a picture of an apple on the side?

The problem is your definition of "pros." That word is meaningless anymore.

heh.. again. i'm not the one touting pro this and pro that.. that's MacVidCards and his following.

i'm a pro and i currently use an iMac,mbp, and iphone and they're awesome for most of my stuff.

Regular colleagues of mine, the ones who are most like the regulars on this forum who kneel at the altar of Apple, are buying iMacs or Macbook Pros.

i bought an iMac too.. my software just can't utilize the potential power in the nmp yet.. it's coming though. and apple is actively assisting the developers of my renderer to tune it towards the mac pro.. patience. (but i've seen some previews and i'm telling you, it's way better than what one could expect from a dual socket machine / single gpu)

What? Where are you getting computers being bigger and more expensive? We've been getting more power in the same footprint year after year, sometimes smaller even. And where has cost gone up?

i'm just responding to people saying we need a bazillion cpu cores and always adding the latest gpus etc.. that's a lot of cash.

There are plenty of options in your "affordable" level. Not sure why you need to quote the extreme $20,000+ example. And again, the key word you used there is "I." Everyone has different needs.

there isn't an affordable solution for radical performance gains.. if i want to double the speed of my renders, i have to triple the cost of my computer..

oh.. but this boxx is 20% faster!! who cares.. that's nothing.. if you're pegging all cores, you're waiting.. if you wait 5 minutes or 4 minutes, there's zero difference.. your workflow is still not running smooth.

i want 2000% faster.. and i don't want to pay out the nose for it.

hardware won't give me that performance gain.. software will.. and it can.

 

[MacVidCards]

I physically met with 2 Directors of Photography in last 2 weeks. One just unloaded his 12 core/D700 to buy a Dual CPU workstation to cut his work time in half. I believe he told me 6K was choking on nMP.

The other DP was putting a Titan-X in his nMP 12 Core.

Are these the "mythical Pros"? They seemed pretty real standing in the sidewalk while I talked to them.

 

[flat five]

I physically met with 2 Directors of Photography in last 2 weeks. One just unloaded his 12 core/D700 to buy a Dual CPU workstation to cut his work time in half. I believe he told me 6K was choking on nMP.

The other DP was putting a Titan-X in his nMP 12 Core.

Are these the "mythical Pros"? They seemed pretty real standing in the sidewalk while I talked to them.

it's just a bs story is all. you're seriously trying to say these dudes worked 40 hours last week.. now they have a cmp and only need to work 20 hours. ??

either that or they still work 40hours and are now making twice the production/$$

it's simply not a real story.. that's all there is to it.

 

[MacVidCards]

it's just a bs story is all. you're seriously trying to say these dudes worked 40 hours last week.. now they have a cmp and only need to work 20 hours. ??

either that or they still work 40hours and are now making twice the production/$$

it's simply not a real story.. that's all there is to it.

I said nothing about how many hours anyone worked last week. And I don't appreciate being called a liar.

I understand that you have a rather tenuous grip on reality, and that may explain your inability to accept the truth.

I will email the guy who unloaded his 12 core/D700 to buy a more suitable machine for high end work. When he gets back to me I will post the details. The system he described was interesting, multiple x16 PCIE slots on opposite sides of the CPUs.

Not liking what someone says doesn't give you the right to call them a liar.

 

[flat five]

I said nothing about how many hours anyone worked last week. And I don't appreciate being called a liar.

I understand that you have a rather tenuous grip on reality, and that may explain your inability to accept the truth.

I will email the guy who unloaded his 12 core/D700 to buy a more suitable machine for high end work. When he gets back to me I will post the details. The system he described was interesting, multiple x16 PCIE slots on opposite sides of the CPUs.

Not liking what someone says doesn't give you the right to call them a liar.

"cut his work time in half"

maybe I'm misinterpreting what that's supposed to mean.

 

[AidenShaw]

When he gets back to me I will post the details. The system he described was interesting, multiple x16 PCIE slots on opposite sides of the CPUs..

The system that I pictured a few posts back has CPUs in the middle, 24 DIMM slots flanking the sockets, and PCIe slots on opposite sides of the chassis. Four double-wide PCIe 3.0 x16 slots.

 

[MacVidCards]

"cut his work time in half"

maybe I'm misinterpreting what that's supposed to mean.

I'm sorry that you are having so much trouble understanding. A render for 6K would take much less time, is what he told me. So rather than sit and wait, he could get next segment running. Why is that so hard to understand?

According to your logic, we should all be content with a G4 since the render will get done eventually anyway, right? Just take a nap and wait til it's done. Some people work on high end stuff, they need high end computers. Apple doesn't make any anymore.

 

[td2243]

my software just can't utilize the potential power in the nmp yet.. it's coming though.

Mine can't either which is why I went with an iMac. Sometimes I feel like it is going to blow up it works so hard.

My question is... The nMP was announced almost two years ago, if Adobe and other companies aren't yet geared up for the nMP, when will they? How do we know these companies will? It

 

[flat five]

• $6,499 - nMP - 8core Ivy Bridge EP- 3.0 GHz - 32 GiB - 256 GB SSD - 2*D700s - 1 yr warranty
• $5,244 - ML350g9 - 8core (2*4) Haswell EP - 3.0 GHz - 32 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $6,368 - ML350g9 - 16core (2*8) Haswell EP - 2.6 GHz - 64 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $7,242 - ML350g9 - 8core (2*4) Haswell EP - 3.0 GHz - 32 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty
• $11,624 - ML350g9 - 24core (2*12) Haswell EP - 2.6 GHz - 256 GiB - 240 GB SSD- 2*TitanX - 3 yr warranty
• $16,082 - ML350g9 - 32core (2*16) Haswell EP - 2.3 GHz - 256 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty
• $20,392 - ML350g9 - 36core (2*18) Haswell EP - 2.3 GHz - 384 GiB - 240 GB SSD- 4*TitanX - 3 yr warranty

those sound great.. the non apple machines that are most interesting to me are those overclocked boxx_es.. 8-core at 4.smthng ghz..

the best (for me) would be macpro 5ghz quad with 2x D7.1 or D8.1

idk, do you see xeons going to that type of speed in the next 3 yrs or so? or ever?

 

[flat five]

Mine can't either which is why I went with an iMac. Sometimes I feel like it is going to blow up it works so hard.

My question is... The nMP was announced almost two years ago, if Adobe and other companies aren't yet geared up for the nMP, when will they? How do we know these companies will? It

adobe will do it if--

smaller companies go though the trial and error etc and put out proof-positive products.

i doubt it's an easy task for a company as large as adobe to drop what they're doing and start rewriting big chunks of code.. i imagine they have a few people tinkering with openCL / gpgpu but probably not a central focus right now [/guessing]

right now, there's basically only one program which can be used as an example and it's final cut.. if it gets to 5 or ten then more companies will have to take notice.. then it's snowball from there.

i have faith things will play out that way.. but not completely convinced either as i bought an imac last year instead of macpro.

 

[flat five]

I'm sorry that you are having so much trouble understanding. A render for 6K would take much less time, is what he told me. So rather than sit and wait, he could get next segment running. Why is that so hard to understand?

it's hard to understand because you said two totally different things.. if i'm supposed to know "cut his work time in half" is meant to be interpreted as "A render for 6K would take much less time"
..then call me dense i suppose.

According to your logic, we should all be content with a G4 since the render will get done eventually anyway, right? Just take a nap and wait til it's done. Some people work on high end stuff, they need high end computers. Apple doesn't make any anymore.

no.. a G4 is a problem because it takes tooo long to render.. you want to set it up to where the computer is rendering at night.. with a G4, it would still be rendering the next day. (and night )

rendering can be likened to pouring concrete.. first you design it/form it/reinforce it/pour it/shape it/finish it.. that's all the work.. then rendering is the curing process.. it's doing something. the main thing which is turning into concrete.. but all your work is finished at that point and you don't sit there and watch the stuff dry.. comeback manana and it's done.

super powerful computers or optimized software can and will change that workflow (the overnight thing).. but the ultimate goal would be to eliminate rendering as a whole from a CAD workflow.. real-time rendering with real-time physically accurate sun/light simulation in the viewports.. with the right software and maybe a couple more iterations of the GPU, the nmp's base design does seem capable of producing those results.. i don't think it's too small or too limited ala 'compromised too much for aesthetics at the expense of power'

 

[AidenShaw]

Watch this, and reset your ideas about what "professional" is, and how the MP6,1 utterly fails this segment:

http://blogs.nvidia.com/blog/2014/03/26/how-pixar-uses-gpus/

The pro world is a CUDA world, OpenCL is for amateurs using FCP.

 

[flat five]

Watch this, and reset your ideas about what "professional" is, and how the MP6,1 utterly fails this segment:

http://blogs.nvidia.com/blog/2014/03/26/how-pixar-uses-gpus/

The pro world is a CUDA world, OpenCL is for amateurs using FCP.

that's cool.. and certainly the direction i hope to see coming to my own software (the realtime stuff).

it's gpu processing but it's not cuda (i don't think).. or opencl.
you don't need either of those to code gpgpu.. they just help the developer.. but if you're wiz enough to do it from scratch, more power to you.

----
didn't realize you meant watch a video.. (or there wasn't a video at the link).. this one?

that looks like openGL stuff.. my modeling software (rhino) is currently able to look like that realtime (minus the animation)..
the actual movie looks different so there's another rendering process that happens i think.. getting the final ray traced look in real time.. that's the goal. (or- i hope that's the goal.. it's just a wish for me)

 

[MacVidCards]

it's hard to understand because you said two totally different things.. if i'm supposed to know "cut his work time in half" is meant to be interpreted as "A render for 6K would take much less time"
..then call me dense i suppose.


no.. a G4 is a problem because it takes tooo long to render.. you want to set it up to where the computer is rendering at night.. with a G4, it would still be rendering the next day. (and night )

rendering can be likened to pouring concrete.. first you design it/form it/reinforce it/pour it/shape it/finish it.. that's all the work.. then rendering is the curing process.. it's doing something. the main thing which is turning into concrete.. but all your work is finished at that point and you don't sit there and watch the stuff dry.. comeback manana and it's done.

super powerful computers or optimized software can and will change that workflow (the overnight thing).. but the ultimate goal would be to eliminate rendering as a whole from a CAD workflow.. real-time rendering with real-time physically accurate sun/light simulation in the viewports.. with the right software and maybe a couple more iterations of the GPU, the nmp's base design does seem capable of producing those results.. i don't think it's too small or too limited ala 'compromised too much for aesthetics at the expense of power'

As usual, your arguments make no sense whatsoever.

"Cutting a render time in half only helps if it is under 8 hours, otherwise there is no benefit since it could happen overnight". Absolute rubbish, and I suspect you know that.

I find it humorous that the nMP's most strident defender is someone who doesn't own one.

I think mine is very cute. I lost my rolling pin so sometimes when I make an Apple pie I use it to roll out the dough. The smooth surface only gets a little dough stuck to it.

And I also think that when 7,1 comes out you are going to see a bunch of receptionist's desks have a lovely grey cylinder suddenly appear.

 

[flat five]

As usual, your arguments make no sense whatsoever.

"Cutting a render time in half only helps if it is under 8 hours, otherwise there is no benefit since it could happen overnight". Absolute rubbish, and I suspect you know that.

it's not rubbish.. i don't think you understand what i'm saying.
regardless of whether a render takes 4 hours or 2 hours or 1 hour.. i'm not doing any work.. if i work 50 hours one week to produce a set of 7 renderings.. say each one took 10 hours.. so the computer was on 24/7..

now say i used a xyz2000mega and each render took 1 hour.

how many hours did i work? 50.. nothing changed except the computer slept for the night.

do you get that?

I find it humorous that the nMP's most strident defender is someone who doesn't own one.

try not to confuse me arguing you & your philosophies with me defending the mac pro.

I think mine is very cute. I lost my rolling pin so sometimes when I make an Apple pie I use it to roll out the dough. The smooth surface only gets a little dough stuck to it.

if you could get one before they cut the i/o access, it'd probably be pretty cool to make a pie crust with.

And I also think that when 7,1 comes out you are going to see a bunch of receptionist's desks have a lovely grey cylinder suddenly appear.

lucky receptionists i suppose.

 

[MacVidCards]

it's not rubbish.. i don't think you understand what i'm saying.
regardless of whether a render takes 4 hours or 2 hours or 1 hour.. i'm not doing any work.. if i work 50 hours one week to produce a set of 7 renderings.. say each one took 10 hours.. so the computer was on 24/7..

now say i used a xyz2000mega and each render took 1 hour.

how many hours did i work? 50.. nothing changed except the computer slept for the night.

do you get that?

What I get is that you are more full of it then a porta-pottie on the last day of Coachella.

You sit and "set up" for an hour, it takes 2 hours to render or 4 hours. Then you spend another hour to set up next thing, it takes 2 hours or 4 hours to render. With which system can you get more done in a day?

You keep making up numbers where whatever happens takes 8 hours to render, therefore overnight is good.

But no rational person is ever going to argue that "It doesn't matter if a 2nd CPU could render in half the time since it all happens overnight", because it doesn't.

Not every render is 8 hours and not every day ends with a render ready to take place at closing time. You are creating a little fantasy world that doesn't exist.

I think the relevant thing here is that when it came time to pull out the American Express at the Apple Store, you said "iMac Please" because the nMP didn't fit your needs.

So why is it you argue so stridently that it DOES fit the needs of all of the people who say it doesn't fit their needs and that a cMP fits them better? You know better then them even though you aren't them and don't have either machine?

You don't find me in the rMB forum going on and on about what a glorious machine it is and that everyone should love it and use one even though I decided to go with an Air because it fit my needs better. I would look rather silly, as you do.

 

[AidenShaw]

it's gpu processing but it's not cuda (i don't think).. or opencl.

Actually read the article and links - it's CUDA and Nvidia's CUDA-based Optix framework.

you don't need either of those to code gpgpu.. they just help the developer.. but if you're wiz enough to do it from scratch, more power to you.

Do you have an inability to read? (and to capitalize the first word of a sentence?)

CUDA is both the low-level API for Nvidia GPGPU processing, and the base of many toolkits (some part of CUDA itself) and libraries for higher level processing.

If you can't find what you need from one of the builtin or addon toolkits, you can build it from the CUDA APIs.

I don't know if it's even possible to "do it from scratch" - if the Nvidia driver only exposes the CUDA APIs, you have to use CUDA.

Anyway, this whole argument is irrelevant for the MP6,1, since it is incapable of running CUDA - and unless the nnMP moves to Maxwell GPUs, the same applies.

The GPGPU world is an Nvidia world, and Apple is not part of it.

 

[flat five]

What I get is that you are more full of it then a porta-pottie on the last day of Coachella.

You sit and "set up" for an hour, it takes 2 hours to render or 4 hours. Then you set up next thing, it takes 2 hours or 4 hours. With which system can you get more done in a day?

You keep making up numbers where whatever happens takes 8 hours to render, therefore overnight is good.

i'm trying to make a point.. you should recognize that.

my renders take around 90minutes.. you set them up and launch them at once.. you don't make one.. wait for it.. make another.. wait for it.

i mean, you can do it like that but it's not efficient.. you have bigger problems than computer speed to worry about if that's your flow.

But no rational person is ever going to argue that "It doesn't matter if a 2nd CPU could render in half the time since it all happens overnight", because it doesn't.

dude, it's small potatoes and a waste of time to argue about.. "oh..oh.. the triple-zinger annihilated the powermac at geekbench.. 35k to 30k!!!".. so what.

i want to geekbenchpress 1 million.. to me, that's something exciting enough to hope for.. pay $thousands to get my render in 45 minutes instead of 90?
wow no thanks.

Not every render is 8 hours and not every day ends with a render ready to take place at closing time. You are creating a little fantasy world that doesn't exist.

i'm trying to explain what it's like to use a computer to complete a project.
8 hours is just an example number. 'overnight' is meant to show that there are way more hours in a week than you work.. 4 times as many. there's nothing wrong with making a computer do it's heavy lifting during those hours.. in fact, it's the smart thing to do. it's not fantasy.

my fantasy world is more about having access to next generation software in which all this arguing over rendering times becomes a non issue.

once things are real time, there's nothing left to argue about.. you can't get any faster.

I think the relevant thing here is that when it came time to pull out the American Express at the Apple Store, you said "iMac Please" because the nMP didn't fit your needs.

not really.. it's because a lot of it's potential would sit untapped for at least a couple of years.

So why is it you argue so stridently that it DOES fit the needs of all of the people who say it doesn't fit their needs and that a cMP fits them better? You know better then them even though you aren't them and don't have either machine?

i don't argue people saying the cmp fits them better than nmp.. i argue the stupid reasonings people give to back up their claim.

if someone were to step in and say (truthfully) - "i use octane".. or "i use thea".. "therefore, the cmp is much better for me than nmp".. i'm not going to argue them because they're absolutely right.

but if someone starts with the "cuda is king.. opencl is st00ps" and you don't use either well.. i'll argue it.

You don't find me in the rMB forum going on and on about what a glorious machine it is and that everyone should love it and use one even though I decided to go with an Air because it fit my needs better. I would look rather silly, as you do.

funny thing.. you don't find me in this forum doing that with the nmp either.

there's a lot of words by me up there ^ and i don't recall any of them being along "the nmp is tiptop bruh!!.. it's way better than ___" etc.

 

[MacVidCards]

Your arguments don't hold water.

There is no rational reason to say that renders at double speed doesn't help workflow. The artificial scenarios you have to create to make this be the case are ludicrous.

I stand by my Coachella comment.

And it seems like your grasp of the "quote " system is lacking as well.

Good luck.

Adios.

 

[flat five]

Actually read the article and links - it's CUDA and Nvidia's CUDA-based Optix framework.




Do you have an inability to read?

read what? this:

Van Gelder showed how Presto – Pixar’s proprietary GPU-accelerated animation system

if you mean something specific, say something specific.. don't just drop a link that talks about a few different things and run.. it's poor internetting.. which is then amplified when you come back after link&run and start insulting people "what are you, stupid?" etc..

you said 'watch this'.. i assume you meant the video i posted.. is that right? is that what you wanted me to watch? if so, that's not cuda based software.

(and to capitalize the first word of a sentence?)

No, I know how to write proper English.

CUDA is both the low-level API for Nvidia GPGPU processing, and the base of many toolkits (some part of CUDA itself) and libraries for higher level processing.

If you can't find what you need from one of the builtin or addon toolkits, you can build it from the CUDA APIs.

I don't know if it's even possible to "do it from scratch" - if the Nvidia driver only exposes the CUDA APIs, you have to use CUDA.

Anyway, this whole argument is irrelevant for the MP6,1, since it is incapable of running CUDA - and unless the nnMP moves to Maxwell GPUs, the same applies.

it's software.. nothing inherently special about nvidia hardware which allows cuda to run.. other than nvidia locking it down to their gpus.
same thing as osx running on macs and not dells even though the hardware inside is moreorless the same.

----------

Adios.

bye