FORUMS: list search recent posts

Phi

COW Forums : Apple Final Cut Pro X Debates

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Jeremy Garchow
Phi
on Nov 13, 2012 at 5:38:13 am

Interesting release dates for the "commercial" versions.

Interesting price, too.

http://news.cnet.com/8301-1001_3-57548761-92/intels-60-core-chip-ships-elit...

Jeremy


Return to posts index

Devin Crane
Re: Phi
on Nov 13, 2012 at 3:02:54 pm

But does it support Thunderbolt???



Return to posts index

Jeremy Garchow
Re: Phi
on Nov 13, 2012 at 3:12:52 pm

That would be one scenario, external to the main motherboard.

Or, this becomes an add on to current Xeon workstations via normal pcie configs.


Return to posts index


Brooks Tomlinson
Re: Phi
on Nov 13, 2012 at 3:42:40 pm

I hope to dream big, and I can only hope that this is in the new mac.

(please, don't spoil my dream with reality!)

Brooks Tomlinson
"I dream in 32bit float"


Return to posts index

Walter Soyka
Re: Phi
on Nov 13, 2012 at 4:23:26 pm

Looks cool!

Anybody know if this processing power is available automatically to the system, or if applications need to be written with Phi in mind (as with NVIDIA/CUDA or OpenCL)? Or if perhaps Phi supports OpenCL?

I also wonder if having "only" 8 GB of RAM on the Phi PCIe card for all those cores will be limiting (8 GB holds about 260 frames of 1920x1080 32bpc RGBA), or if PCIe is fast enough to allow practical swapping to system RAM as necessary without slowing all those sizzle cores down too much.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 13, 2012 at 6:45:06 pm

[Walter Soyka] "Anybody know if this processing power is available automatically to the system, or if applications need to be written with Phi in mind (as with NVIDIA/CUDA or OpenCL)? Or if perhaps Phi supports OpenCL?"

I don't know. Perhaps Phi means nothing to us video peoples.

Technology is simply flying off the rails:

http://www.barefeats.com/hard159.html

Jeremy


Return to posts index


Walter Soyka
Re: Phi
on Nov 13, 2012 at 11:30:19 pm

[Jeremy Garchow] "I don't know. Perhaps Phi means nothing to us video peoples."

Did you post that just to get me all riled up about sizzle core performance again? :)

Even if it doesn't matter to video editorial (which is primarily throughput and secondarily processing), every advance in computing affects adjacent disciplines like animation, compositing, 3D rendering, etc. (which are primarily processing and secondarily throughput).

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 14, 2012 at 12:12:52 am

[Walter Soyka] "Did you post that just to get me all riled up about sizzle core performance again? :)"

I'm saying I don't have a clue what Phi will do except help Stephen Hawking wrap his genius around the cosmos faster than what is possible today.

I don't know what this means for the impending MacPro thingy, other than the rough timeline given by Tim Cook seems to roughly match up with this development.

I can bring FCPX to it's knees, yet the software allows me to run all these process at once.

FCP7 prevented me from running all these processes at once. These are both by design.

This seems to mean to me that Apple is going to be building a decently fast machine, something that is faster than the current offerings, that will allow some sort of parallel encoding to be offloaded from the main CPU.

I don't think this type of thing will be done with an ARM processor anytime in the near distant future.

But there are times when I am over optimistic which is why I should probably stop hanging around here so much. :-D


Return to posts index

Michael Gissing
Re: Phi
on Nov 14, 2012 at 12:30:54 am

Having a multi core co processor will require software specifically tailored to take advantage of this power. I can see that having such power might make a difference to WETA and other big VFX companies. Can't see the advantage over GPU cores like CUDA for the average editor & post house but there may be. If the card is sub $2K then sure it may be worth a consideration.

Many years ago Fairlight introduced a FPGA card (field programmable gate array) card to do all audio mix & effects processing. The power in this card was such that even Apple showed some interest and had some preliminary informal talks with Fairlight. Basically it is another way of having a multi co processing unit doing dedicated tasks rather than relying on fixed 32bit DSP chips or using the main processor.

If I was AutoDesk, I might be tempted to write software that could leverage this processing but again why go there and not just use GPU as graphics cards have become cheap powerful co processors.


Return to posts index


Walter Soyka
Re: Phi
on Nov 14, 2012 at 1:17:35 am

[Michael Gissing] "Having a multi core co processor will require software specifically tailored to take advantage of this power. I can see that having such power might make a difference to WETA and other big VFX companies. Can't see the advantage over GPU cores like CUDA for the average editor & post house but there may be. If the card is sub $2K then sure it may be worth a consideration."

This is why OpenCL is so important. OpenCL isn't just about parallel computing on graphics cards; it's about parallel computing across all kinds of heterogeneous computational devices like CPUs, GPUs, FPGAs and DSPs.

With technologies like GPGPU (CUDA on NVIDIA, OpenCL on AMD/NVIDIA) and soon maybe Phi, you can fit what was a render farm a decade ago (making that up, but it might not be that far off) onto a PCIe card today.

That's just plain cool. Ever increasing computational power translates into ever increasing creative opportunities.


[Michael Gissing] "If I was AutoDesk, I might be tempted to write software that could leverage this processing but again why go there and not just use GPU as graphics cards have become cheap powerful co processors."

Current Autodesk software is pretty reliant on the GPU for actual graphics processing (OpenGL), not general purpose computing on the the GPU (CUDA/OpenCL).

There's actually a very interesting battle on the desktop going on now between technologies like OpenGL (Smoke, or Ae effects like Element/Magic Bullet Looks/Mir/FreeForm/ShapeShifter) and technologies like CUDA (Ae's 3D ray-tracing renderer, GenArts Sapphire, DaVinci Resolve). I know Motion used to be OpenGL-reliant, but I'm not sure if or to what degree that has changed with FCPX/M5/OpenCL.

These are two different (but valid!) approaches to computer graphics, each with unique strengths and weaknesses. Exciting times to be a designer!

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Michael Gissing
Re: Phi
on Nov 14, 2012 at 1:25:30 am

Thanks for the info Walter. Yes we live in times where processing power is so cheap it is easy to forget that what was gee whiz only a few years ago is ho hum now.


Return to posts index

Walter Soyka
Re: Phi
on Nov 14, 2012 at 1:30:23 am

[Michael Gissing] "Thanks for the info Walter."

My pleasure -- thanks to you and Jeremy for giving me the opportunity to spout off on one of my favorite topics!



[Michael Gissing] "Yes we live in times where processing power is so cheap it is easy to forget that what was gee whiz only a few years ago is ho hum now."

This statement has a twin. We also live in a time where processing power is so enormous today that it's easy to forget that what is gee wiz now will be ho hum in a few years.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index


Richard Herd
Re: Phi
on Nov 14, 2012 at 5:51:38 am

[Walter Soyka] "ho hum"

all the processing power in the world won't make up for the pedantic PIOP debate.


Return to posts index

Walter Soyka
Re: Phi (now PIOPs -- really???)
on Nov 14, 2012 at 6:24:24 am

[Richard Herd] "all the processing power in the world won't make up for the pedantic PIOP debate."

Speaking of topics I spout off on...

I asked for PIOPs, I got them, and I was wrong. PIOPs in FCPX, as implemented, are the worst.

What I really wanted was a method for addressing the underlying problem with FCPX vis-a-vis PIOPs, which I raised over and over in that so-called pedantic debate: setting IOPs is intentional and that data shouldn't be nuked by the app with a simple focus-shifting mouse click. The editor should never, ever have to manually relocate range boundaries in a clip-ish object that he or she did not manually clear.

That's why I first proposed pushing selection onto the undo stack, and why I most recently proposed what I'll call SIOPs (stored in and out points) [link], which would not preserve the last un-favorited range as such, but would allow the last un-favorited range in a clip to be recalled with a keystroke after said range's deselection (and before selecting a new range, which would have cleared the old range anyway). I think that SIOPs are a worthwhile compromise in that they don't mess with FCPX's range paradigm, thereby preserving something legitimately good about FCPX, but they also don't wantonly obliterate user data, thereby preserving something legitimately good about FCP Legend.

I'm happy to debate their merits in any thread or sub-thread, even when they come up apropos of nothing like this -- and that's saying something in this forum!

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Michael Gissing
Re: Phi (now PIOPs -- really???)
on Nov 14, 2012 at 6:44:45 am

PIOPs
NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOoooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo...................................................................................


Return to posts index


Walter Soyka
Re: Phi (now PIOPs -- really???)
on Nov 14, 2012 at 6:51:31 am

[Michael Gissing] "PIOPs NOOOOOOO......."

Just when you thought threads on general computing trends were safe!

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

David Lawrence
Re: Phi (now PIOPs -- really???)
on Nov 14, 2012 at 8:04:58 am

[Walter Soyka] "Just when you thought threads on general computing trends were safe!"

Must. Not. Post....................................................................................................................................

_______________________
David Lawrence
art~media~design~research
propaganda.com
publicmattersgroup.com
facebook.com/dlawrence
twitter.com/dhl


Return to posts index

Aindreas Gallagher
Re: Phi (now PIOPs -- really???)
on Nov 15, 2012 at 10:08:24 pm

hang on - I'll think of something generically inflammatory to say.

then we can start it.

All.

Over.

AGAIN.

thunderclap:







http://vimeo.com/user1590967/videos http://www.ogallchoir.net promo producer/editor.grading/motion graphics


Return to posts index


Jeremy Garchow
Re: Phi
on Nov 14, 2012 at 1:33:28 am

I don't understand.

You are saying this only makes sense to Weta?

You must never transcode anything ever, or very little.


Return to posts index

Michael Gissing
Re: Phi
on Nov 14, 2012 at 3:11:33 am

[Jeremy Garchow]"I don't understand.
You are saying this only makes sense to Weta?
You must never transcode anything ever, or very little."

I was implying that this sort of co processor would make more sense to a company that was doing huge rendering and number crunching manipulation and may well have custom software that can be tweaked to take advantage of this sort of multi core processor.

I don't get that you see "makes more sense to X" as "makes no sense to the rest". Also I said nothing about transcoding??


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 14, 2012 at 3:21:53 am

[Michael Gissing] "I don't get that you see "makes more sense to X" as "makes no sense to the rest". Also I said nothing about transcoding??"

That's not what I see.

I see this as making more sense to much more than the highest of high end, i.e. Weta


Return to posts index

Michael Gissing
Re: Phi
on Nov 14, 2012 at 3:30:33 am

"That's not what I see.
I see this as making more sense to much more than the highest of high end, i.e. Weta"

So we basically agree then. In the short term big players that have custom software can make best use. We can all potentially benefit but my question was whether software companies would see sense (ie $$s) in re writing software for this sort of co-processor over GPUs which they are currently doing.

If they do then we can all benefit. However when the trend is to cheap NLE software in mobile devices, I see perhaps less use for the average editor than say CUDA or OpenCL.


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 14, 2012 at 11:09:17 am

[Michael Gissing] "So we basically agree then. In the short term big players that have custom software can make best use. We can all potentially benefit but my question was whether software companies would see sense (ie $$s) in re writing software for this sort of co-processor over GPUs which they are currently doing.

If they do then we can all benefit. However when the trend is to cheap NLE software in mobile devices, I see perhaps less use for the average editor than say CUDA or OpenCL."


None of this could be handed to grand central dispatch?


Return to posts index

Richard Herd
Re: Phi
on Nov 14, 2012 at 5:50:07 am

[Jeremy Garchow] "I don't know"

To hell with Persistent In and out points.


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 1:54:11 am

[Richard Herd] "To hell with Persistent In and out points."

Did you know persistent in and out points spelled backwards is stniop tuo dna ni tnetsisrep?


Return to posts index

Richard Herd
Re: Phi
on Nov 15, 2012 at 4:54:17 pm

[Jeremy Garchow] "stniop tuo dna ni tnetsisrep"

Ah ha! We now know what he meant in Temple of Doom while pulling the heart out of the poor victim.







Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 5:06:32 pm

I see your youtube and raise you a youtube







Return to posts index

Richard Herd
Re: Phi
on Nov 16, 2012 at 12:28:08 am

Doubling down







Return to posts index

Jakub Vomacka
Re: Phi
on Nov 14, 2012 at 8:44:05 am

"... it is a pretty standard accelerator card for HPC and generic number crunching. It looks like a cluster of x86 servers that are on a TCP/IP network, and takes almost no programming expertise to port to. Optimization is very similar to a standard x86 CPU, and you can use the Intel tools that are very well established in the HPC world."

basicly takes x86 code and after recompilation in intel software its able to run on many more parallel x86 CPU cores

more to read: http://www.theregister.co.uk/2012/11/12/intel_xeon_phi_coprocessor_launch/

http://newsroom.intel.com/community/intel_newsroom/blog/2012/11/12/intel-de...
http://download.intel.com/newsroom/kits/xeon/phi/pdfs/Intel-Xeon-Phi_Factsh...
http://download.intel.com/newsroom/kits/xeon/phi/pdfs/Intel-Xeon-Phi-Coproc...


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 2:15:29 am

[Jakub Vomacka] "basicly takes x86 code and after recompilation in intel software its able to run on many more parallel x86 CPU cores"

Thanks Jakub.

In my limited understand of the inner workings of a computer, doesn't this mean that, with a fairly simple exchange this type of processing is simply added on? It goes from the bigger Xeons to the smaller 60(!) cores?

Can't something like Grand Central Dispatch manage the traffic efficiently?

Or am I dreaming too much?


Return to posts index

Walter Soyka
Re: Phi
on Nov 15, 2012 at 5:47:44 pm

[Jeremy Garchow] "In my limited understand of the inner workings of a computer, doesn't this mean that, with a fairly simple exchange this type of processing is simply added on? It goes from the bigger Xeons to the smaller 60(!) cores?"

Well, memory access is still an issue -- it's not clear to me that Phi has direct access to main system memory (and even if it does, it will be relatively slow over the PCIe bus). Hearing that it's addressable via TCP/IP suggests it runs almost as a separate computer within its host.


[Jeremy Garchow] "Can't something like Grand Central Dispatch manage the traffic efficiently?"

I don't think this would work, as GCD currently stands. Wikipedia says GCD is built for symmetric multiprocessing [link], which means multiple identical CPUs (or cores) attached to the same pool of memory.

I could be mistaken, but it sounds more like Phi cores are different from mainboard Xeon E5s, and it sounds like the Phi system has separate memory. That's why I had wondered about OpenCL, which allows execution of parallel tasks on different hardware.

This is all still pretty speculative on my part, so hopefully as more details emerge on Phi we'll get a clearer picture of what's possible (assuming the next Mac Pro has a PCIe slot to stick this bad boy in).

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 6:47:23 pm

[Walter Soyka] "Well, memory access is still an issue -- it's not clear to me that Phi has direct access to main system memory (and even if it does, it will be relatively slow over the PCIe bus). Hearing that it's addressable via TCP/IP suggests it runs almost as a separate computer within its host."

It's still faster than not having it, right?



[Walter Soyka] "I don't think this would work, as GCD currently stands. Wikipedia says GCD is built for symmetric multiprocessing [link], which means multiple identical CPUs (or cores) attached to the same pool of memory. "

As it currently stands, sure. What I don't know is that it seems this truly is an extension of the Xeon code, if that's true, then perhaps GCD could be updated fairly easily since it already seems to be addressing intel's code.

[Walter Soyka] "This is all still pretty speculative on my part, so hopefully as more details emerge on Phi we'll get a clearer picture of what's possible (assuming the next Mac Pro has a PCIe slot to stick this bad boy in)."

We'll see, but that's why I found the release date to these and Tim Cook's proposed "something special" date to be a convenient coincidence.

Who knows.


Return to posts index

Walter Soyka
Re: Phi
on Nov 15, 2012 at 7:29:05 pm

[Jeremy Garchow] "It's still faster than not having it, right?"

Well, sure, but I don't think popping a Phi in your computer instantly supercharges it as if the main computer were a 64-core monster.

This would certainly be awesome for throughput-light, computation-heavy work like 3D rendering where you can load up the onboard RAM and let the CPUs have at it, and it should fare at least as well as GPUs at throughput-heavy work.


[Jeremy Garchow] "As it currently stands, sure. What I don't know is that it seems this truly is an extension of the Xeon code, if that's true, then perhaps GCD could be updated fairly easily since it already seems to be addressing intel's code."

It's not just an issue of different instruction sets; it's an issue of physically different architectures. I don't think you can't arbitrarily shuffle instructions around from the mainboard to the Phi without also shuffling the data it needs to operate on from main memory to Phi memory. If I'm understanding all of this correctly (a decent-sized if!), that's creating a lot of overhead just to move thread execution around.

Phi sounds more like clustering, which GCD is not built for. Apple's Advanced Computation Group used to be active in clustering research (XGrid), but their interest seems to have waned in recent years.

But of course I could be wrong.


[Jeremy Garchow] "We'll see, but that's why I found the release date to these and Tim Cook's proposed "something special" date to be a convenient coincidence."

This would be very cool, but I'm not sure I see the business case for Apple pushing the envelope in performance computing anymore. I'd love to be wrong.

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 8:02:53 pm

[Walter Soyka] "This would be very cool, but I'm not sure I see the business case for Apple pushing the envelope in performance computing anymore. I'd love to be wrong."

If you spend some serious time with FCPX, and watch all the things it does most of it seemingly when you aren't watching, a case could be made.

There's a reason they kept the MacPro channel open. I'm dreaming big, I guess.

Jeremy


Return to posts index

Walter Soyka
Re: Phi
on Nov 15, 2012 at 8:19:10 pm

[Jeremy Garchow] "If you spend some serious time with FCPX, and watch all the things it does most of it seemingly when you aren't watching, a case could be made."

I do get that -- there's a dizzying amount of work that can be done in the background. I'm just not sure how many people will want to pay $8,000 for a workstation to do it "super-fast" (as my four-year-old would say) when they could pay $2,000 for an iMac that does it "regular-fast" (again, as my four-year-old would say).


[Jeremy Garchow] "There's a reason they kept the MacPro channel open. I'm dreaming big, I guess."

I do sincerely hope I will be impressed by the 2013 Mac Pro, whatever it will be. Competition is a good thing, and I like to keep my options open!

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

Jeremy Garchow
Re: Phi
on Nov 15, 2012 at 8:28:34 pm

[Walter Soyka] "I do get that -- there's a dizzying amount of work that can be done in the background. I'm just not sure how many people will want to pay $8,000 for a workstation to do it "super-fast" (as my four-year-old would say) when they could pay $2,000 for an iMac that does it "regular-fast" (again, as my four-year-old would say)."

I could go on about this.

I think FCPX is poised for much more than a few iMacs.

It is a whole new take on editing and transcoding. My beliefs are unpopular, and people can only see it for the magnetic timeline. It is much more than that.

Jeremy


Return to posts index

Walter Soyka
Re: Phi
on Nov 15, 2012 at 8:47:54 pm

[Jeremy Garchow] "I could go on about this. I think FCPX is poised for much more than a few iMacs. It is a whole new take on editing and transcoding. My beliefs are unpopular, and people can only see it for the magnetic timeline. It is much more than that."

Please do. I'd love to hear your thoughts.

I think we'll find a lot of common ground. FCPX is built on a set of very modern and scalable core OS X libraries and technologies, and as such, I think it's well-positioned to take advantage of whatever hardware improvements Apple chooses to implement.

Coming at this from my effects perspective, the performance of its 32b floating point processing pipeline on these "little" machines is really, really impressive. The sky is the limit on bigger machines.

I'd be really interested in what you would have to say from a more editorial perspective.

Thanks,

Walter Soyka
Principal & Designer at Keen Live
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
RenderBreak Blog - What I'm thinking when my workstation's thinking
Creative Cow Forum Host: Live & Stage Events


Return to posts index

David Lawrence
Re: Phi
on Nov 15, 2012 at 9:19:47 pm

[Walter Soyka] "I'd be really interested in what you would have to say from a more editorial perspective. "

I would too.

_______________________
David Lawrence
art~media~design~research
propaganda.com
publicmattersgroup.com
facebook.com/dlawrence
twitter.com/dhl


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]