If I'm not mistaken, Thunderbolt is the equivalent of one 16x PCIe 2 lane. Really hasn't been an issue for me and my setup. This is going to be a bottleneck for high end GPU plus Kona video card, plus say a Red Rocket, no?
It wouldn't be the norm to run an extra GPU card externally, the GTX 680 class of card is among the best and while you can get marginally better performance from other desktop cards vs the 2GB GTX 680M, it's not going to be worth investing in that. If you ran a whole load of things at once over Thunderbolt, you might hit a bottleneck somewhere but it's all bi-directional channels and out of all the tests done in various setups, they don't seem to have run into any show-stopping bottlenecks.
One good demo showing the power of it is the following where they are using 4K RED footage coming off a Promise RAID and a BlackMagic Ultrastudio to record the video itself and recording the screen session out to the RAID at the same time and still real-time 4K on a Macbook Air with a single Thunderbolt port:
[VIDEO]
It's not a setup anyone should adopt for that kind of work by any means but it's cool that it can do that and with dual Thunderbolt ports and Thunderbolt going to 20Gbps next year, the options are just going to get better.
Thunderbolt is only nearly 2 years old, had a 1 year Apple exclusivity and is proprietary to Intel and has a certification process. New technology tends to get an immediate resistance, especially when it conflicts with a 20 year-old status quo but it's enabling a lot of nice, affordable solutions so there's little reason to be against it.
Using a Magma external chassis. As far as I know they only have 2 slots available. A workstation, such as the HP (or a new Mac Pro) can have many multiple lane slots. Plus plenty of power and cooling.
The Mac Pro has a GPU in one of the x16 slots and uses up to 200W of power so that leaves about 100W and 1 x16 and 2 x4s for expansion. Dual TB on the iMac is like the 2 x4 slots so the MP offers the extra x16 but one good thing TB offers is plug and play so you can have a setup of peripherals for video editing and switch them for different peripherals for audio editing just by changing a plug.
Also, if you get involved in threads on Adobe Premiere and After Effects, people are switching primarily to PC workstations, not iMacs. They are now going after the people on Macs alienated by Final Cut X, which are many. Without a new Mac Pro, I only see that migration continuing.
It's not a good thing for anyone to have a reason to migrate away from the platform, even if it is less than 5% of Mac users and a new Mac Pro will give some of those people a reason not to move. I think some people will move because of FCPX alone because FCP would have been their only reason to use the platform in the first place.
Hardware requirements for Premiere/After Effects are high when you get into complex work, which many shops do, even the one man facilities.
Faster hardware will always give faster results but they always stop somewhere. No matter how many slots they put in, you'd still only be able to run 2 GPUs at most and it'll fall far short of systems that can run 8 GPUs:
[VIDEO]
Apple can't build a machine based on the needs of the highest imaginable task someone can use a computer for because there's no limit to that. They have to build a machine that hits a certain price range, has a sizeable market and has a future. They will never offer Intel's fastest CPUs for example because they are too expensive.
I doubt the Mac Pro they will introduce this Summer will be exactly the same as before, otherwise they'd have put in Sandy Bridge. The only question now is what compromises they've made.
Well, if indeed there is Thunderbolt in the next Mac Pro, and why not, I would hope that would be in addition to slots, and also multiple easily removable HD's and/or SSDs. So fast external peripherals should be possible there as well. Perhaps more power (watts), but who knows. Not to nitpick, but the Quadro K5000, the most powerful GPU for the Mac Pro draws 122 watts, not 200. Also 32GB of RAM is not enough for a lot of Photoshop work unless you want to be swapping data out of RAM on very large compositions. The current Mac Pro will go to 128GB of OWC memory. With the advent of cameras like the Nikon D800 for $2995, producing 75MB RAW images, layered compositions in PS can easily reach 500MB and more. That's a lot of swapping. Mac Performance Guide has great documentation on this.
I'm not saying that 27" iMac tricked out is not a good choice for many people, perhaps for most. It is over $3000, however, cheaper, but not by much over current models of MacPro like the Hexacore 3.33 Ghz with a decent GPU and the ability to upgrade to the best, internal HDs and lots of RAM. Most people in the business already have a computer monitor that is suitable. That being said there is a real need for a workstation class computer in the Mac lineup. Just hope we get one and not a box with a lot of Thunderbolt ports and little internal expansion capability. But anything would show that Apple still cares about the video market. That is important to software developers as well as customers.
But do those things actually happen? You seem to decide these problems exist without seeing that they exist. The compute tests that have been done with the CS suite in OpenCL don't really back up what you're saying. You seem to be happy to keep price out of the equation too. The starting price for the Quadro K5000 is said to be $2249. That's more than a whole iMac with a GTX 680M.
Actually the only times I've looked at Quadro cards have been a bit after their release. They often drop in price fairly quick, and I compare first to what people are getting from the gaming cards. I've tested Mudbox on a couple different configurations (not all of them owned by me), including a macbook pro. Keep in mind I was talking about options. Note my prior comment about the Quadro K5000. It's just what they've brought out in the current generation so far. If you read my post below from another thread. I mentioned it as being one that is probably more commonly used in hero suites. They might be much more widespread at top end facilities, but I wasn't referring to those. A lot of this stuff just wasn't feasible on almost anything a couple years ago. In my 3d paint example, that's mostly driven by video memory. I don't think I would have tried to load up that kind of resolution in the past in a program that loads it all to video memory. As far as freelancers are concerned, Macs are reasonably popular but extremely inflexible as you just noted. Also you didn't note with NEC and Eizo, you do get more out of them. Eizo bundles the calibration software. Color Navigator has a lot of tracking functions. They also have a lot of compensation for panel uniformity. NEC does the same with colorcomp and spectraview. In the case of NEC you buy their colorimeter + software package for $400 or so. They're both good. If you were talking about graphic design or compositing and grading, I'd say they're worth it. Even the mini would have a lot of potential with a better gpu.
Quote:
Originally Posted by hmm
1GB wasn't even high when the mac pro 2010 revision debuted. That would be the area where it really irritates me. The common "average user just checks facebook" argument isn't exactly valid when we restrict the comparison to workstation level hardware. Check out the P90X of graphics cards. It's a pretty extreme comparison, but I'm not sure any of the current mac gpus including the 680mx could load even half of that without crashing. Companies probably buy hardware like that for hero suites so that it can be displayed for clients at full resolution without exporting to an offline renderer when a change is made. That's my guess.
Quote:
This is before you add a single PCI expansion card and the performance of that whole machine is still comparable to a $2200 iMac. The Mac Pro is better no question but not so much better to be worth a one man shop paying $7300 more.
I always ended up upgrading in bits and pieces. The display thing was due to consistent irritations with cheaper displays. I don't recall ever suggesting the dual ssd cards there, although I own a couple ssds, one of them being a 512. I haven't personally suggested 64GB of ram. Adobe suggests that for After Effects rendering with a 16 core machine to render on all cores. This is due to hyperthreading. 16*2 = 32 logical cores * 2GB/core. If you're going with a 16 core machine, the extra cost is probably doable. 32GB isn't uncommon for a new machine today. You can achieve that on the imac. Prior to 2012, if you really had high ram requirements, it was definitely in favor of the mac pro due to the price disparity between 4GB and 8GB dimms. A mac pro with 32GB was cheaper than the same in an imac.
Quote:
So are you saying the 2GB Quadro 4000 you previously decided would be your choice is no longer good enough?
You know I've been waiting for video memory to increase for a while. Some applications hitting the market in recent years have relied more on the gpu where they're loading huge files onto it. Once you go past a certain size, you notice a huge drop in frame rate where orbiting around something becomes really choppy. I wouldn't want to work like that for hours at a time. The workstation cards have different driver tuning. I'm not sure how they compensated, but it doesn't seem to get as bad. You can hit a wall either way. Note the link above. You'll see why I said it seemed to be designed for hero stations. You can work with a ton of resolution without it bogging down. Several years from now, that will probably be doable in desktop graphics.
Anyway back to the question regarding one man shops. Typically a slightly customized solution is ideal. Some of them need a lot of power to support their workflow. Otherwise low end probably something like a 3770k. Otherwise go for one of the more cost effective solutions in a 6 or 12 core LGA2011 package depending on available configurations. If you use anything that is heavy on the gpu, in gaming cards I'd go with something like this assuming drivers are stable in the applications in question. Ram is a much bigger help for these things than the clock speed of the card once you're using at least a mid range card. Driver stability is often an issue in 3d apps, even under OSX. There are obviously other things. An example would be if you use After Effects or Premiere, you're reliant on CUDA. This means if Apple releases all AMD this year (as they seem to alternate) you're waiting that one out. If the gpu drivers are flakey or don't support a specific hardware feature you want, it's whatever is available in the Quadros. Unfortunately there is no Quadro K4000 at this point. It might show up later. On ram I'd suggest using 8GB dimms. Install 2. Memtest them prior to placing the machine back in service. If the system still hits the disk frequently, install 2 more bringing it to 32GB. An SSD boot drive is only critical if it has to frequently read information on that drive or it needs to double as a scratch disk for whatever app (if 32GB isn't enough). If you're doing anything that involves a lot of color, I'd at least go with NEC. It's easier to keep them fairly consistent over time with their LUT system. They perform better over their brightness range than Apple, which always seems to tune for max brightness. If you ever have to bring images up side by side to check changes, you learn the value of uniformity pretty quickly. It's annoying when two can be identical yet look a bit different. In terms of price, you can get a PA271w around the price of a thunderbolt display and just use a mini displayport --> displayport cable.
Does that clarify my opinions a bit better? They're fairly consistent, although re-reading I can see where you might have perceived a conflict.
Quote:
Yes, it's very annoying. Having to type purge into the terminal regularly shouldn't be necessary.
I hate that. I hate it when finder corrupts. I hate HFS+ and the continued need for disk warrior with optical drives phasing out. I hate the lack of full displayport 1.2 support. There are many things about OSX that genuinely piss me off. I hate spotlight conflicts with hidden scratch disk files. It makes me mad that Apple keeps adding features rather than fixing old problems.
I'm just going to add what I dislike about thunderbolt. It's proprietary rather than something that is easy to implement across various classes of devices. It's more limiting than you would suggest. It's not the same as PCI. Thunderbolt is its own protocol. I'm amazed people are this surprised that it hasn't caught on. Apple claimed they removed the express card slot because hardly anyone used it. What would make thunderbolt any different outside of available Apple's dedicated peripherals? I would have preferred to see a higher bandwidth PCI spec dedicated to notebooks "and" mobile devices that can be implemented across various designs rather than tied to intel products.
But but but the heavy lifting gets done on AWS because the iMac can't do it and they didn't want to invest 1M in the required hardware.
I know but the Mac Pro in a workstation setup can't do that heavy lifting either. It requires too much hardware. The point is that for the workstation part, loads of computer models are capable of it. There doesn't need to be some platinum members only club mentality.
I never believed they would kill the Mac Pro anyhow
This update will be nearly 3 years out. That's got to be the longest product refresh gap in the company history. The Mini only went 1.5 years without an update and people were saying it was dead before that. I think they must have seriously considered it.
Valid points.
A different thing, but still, is Aperture. In 4 days it's 3.0 version will be 3 years old. Many dot releases have come since, but there are many feature requests being ignored. Good to get updates for free these past years, but I'd rather pay for a 'catchup to the competition'
Apple claimed they removed the express card slot because hardly anyone used it. What would make thunderbolt any different outside of available Apple's dedicated peripherals? I would have preferred to see a higher bandwidth PCI spec dedicated to notebooks "and" mobile devices that can be implemented across various designs rather than tied to intel products.
That might come about but TB has already gained a fair amount of traction in hardware:
They can't put in TB; that requires a new CPU that Intel simply doesn't make.
Not necessarily a custom CPU but it would likely need some form of GPU soldered onto the motherboard in addition to any PCI cards. It's not essential for them to include Thunderbolt on the Mac Pro but it would allow easily sharing peripherals between it and other machines.
That is very true, many people use way more than that. See here
They said they were doing a big photo comparison using loads of layers, which sounds like loads of stills next to each other to compare them. If you set your render cache in After Effects to 90% of say 96GB RAM and hit play on a long movie, it'll fill it up too . You can even make a RAM disk and put your whole filesystem on it but being able to use it doesn't make it a requirement. It would be a better use of RAM not to put photos for comparison into a single comp but to use an app meant for photo comparisons.
There are massive image comps like the gigapixel images you see. There's one here done on a machine with 8GB RAM:
192,000 by 96,000 pixels, Photoshop file was 120GB, built from 600 photos. 8GB shouldn't be the aim - that was done in 2009 - but it's possible to do these things with much lower resources. The higher resources just make them easier to do.
Once you do things like that, it's better having a database-like setup. Like how FCPX uses a database instead of loading an entire file into memory or how running Google Maps doesn't load the entire world into your web browser. Photoshop should allow you to open a folder of images like a .bundle and it would just load the tiles you were looking at. It would be quicker to dynamically load things like that than put everything in RAM.
Well, if indeed there is Thunderbolt in the next Mac Pro, and why not, I would hope that would be in addition to slots, and also multiple easily removable HD's and/or SSDs. So fast external peripherals should be possible there as well. Perhaps more power (watts), but who knows. Not to nitpick, but the Quadro K5000, the most powerful GPU for the Mac Pro draws 122 watts, not 200. Also 32GB of RAM is not enough for a lot of Photoshop work unless you want to be swapping data out of RAM on very large compositions. The current Mac Pro will go to 128GB of OWC memory. With the advent of cameras like the Nikon D800 for $2995, producing 75MB RAW images, layered compositions in PS can easily reach 500MB and more. That's a lot of swapping. Mac Performance Guide has great documentation on this.
It varies. OSX is absolute shit at memory management, but even at 16 bpc with a large comp that saves out a 2-3GB compressed psb, 32GB should not generate slow results. I've tested this with 10-15k comps, spherical panos, and layered texture maps. You just max out ram and add scratch disks. You should not experience brush lag or find it impossible to apply adjustments due to memory. If the objective is photoshop filter drag racing, that doesn't align well with any typical workload I've ever seen. It was a 32 bit app on OSX until CS5 anyway where scratch disks were even more crucial with inflated file sizes. Now you just push as much of that as possible onto ram, but the mechanic is still present in its programming. Further I can attest to the need for a lot of memory at 32bpc, but support for floating point formats in photoshop is terrible. They don't prioritize it at all, so you're left with a very limited toolset. It's a very stagnant program compared to those used in video compositing (nuke, fusion, after effects). They keep adding new poorly implemented features. Further D800 files aren't that large if you had to deal with P45 files at the end of the PowerPC era. These were around 110MB as 8 bit tiffs or twice that at 16. Obviously add layers and the sizes ballooned quickly. I will say this again. 500MB is nothing. People dealt with that on G4 towers, even if it wasn't pleasant.
Quote:
Originally Posted by PhilBoogie
They can't put in TB; that requires a new CPU that Intel simply doesn't make.
That isn't true. What is true is that they can't route it through integrated graphics. There are certain placement requirements in that it demands specific logic board space, but anyone in charge of engineering might present several possible solutions to how this might be routed. It might require some form of embedded graphics if it is seen as a priority. Frankly none of the thunderbolt peripherals were designed with the mac pro in mind. Note the short display cord.
Once Intel takes more control of the motherboard market with Broadwell, it'll be adopted a lot more.
Don't worry, Facebook has you covered:
I'd like to see something that could be implemented in future tablets and phones as well regardless of their reliance on intel and not encumbered by Apple's licensing clauses for the connector. Phones are still pretty slow in terms of transfer rates, but they have been improving. I dislike specialized solutions until you're at a level where only a smaller number of individuals really needs that capability. I've mentioned SAS. It serves a purpose, yet I don't think every computer requires a SAS or mini-SAS HBA.
Quote:
There are massive image comps like the gigapixel images you see. There's one here done on a machine with 8GB RAM:
192,000 by 96,000 pixels, Photoshop file was 120GB, built from 600 photos. 8GB shouldn't be the aim - that was done in 2009 - but it's possible to do these things with much lower resources. The higher resources just make them easier to do.
Once you do things like that, it's better having a database-like setup. Like how FCPX uses a database instead of loading an entire file into memory or how running Google Maps doesn't load the entire world into your web browser. Photoshop should allow you to open a folder of images like a .bundle and it would just load the tiles you were looking at. It would be quicker to dynamically load things like that than put everything in RAM.
I'll admit I've never done anything near that kind of size. This isn't uncommon though. As for databases, it's also possible to make formats that are memory mappable. In other words readable directly to virtual memory rather than loaded into ram and immediately swapped due to size. CS6 actually recommends at least 8GB of ram, but it can sometimes require significantly more to behave properly. I'd usually say add it until things no longer lag. The performance per dollar up to a certain point is much lower than the cost of ssds. If it's maxed add an ssd. Part of the problem is OSX and its terrible memory management. It's necessary not to drift off into hyperbole here. Still photography hasn't evolved much in recent years. Pretty much anything you can do today was done on PowerPC era hardware and graphics workstations prior to that. It was a lot slower on a G4 or G5, but you could do many of those things. You'd keep active indexing things like spotlight off, limit history states to something sane, use fast scratch disks, and be as efficient as possible with layer sets (they weren't called groups back then). What you're describing with databases is something like a mipmapped setup. Google maps has things cached at different levels. When you're zoomed out it's can load different files.
I really don't think they have anything resembling traction. I think they should have already been at the "hundreds of devices" by now, it's been nearly two years already.
I really don't think they have anything resembling traction. I think they should have already been at the "hundreds of devices" by now, it's been nearly two years already.
USB 3 was quicker to get there at around 9 months - USB 3 only arrived January 2010 so I guess Intel wasn't all that far behind in supporting it:
They're including computers, motherboards and even PCI cards there but USB has a much wider audience than Thunderbolt because it's backwards compatible with USB 2.
Apple's 1 year exclusivity with Thunderbolt was quite damaging because non-Apple people started to hate the standard just like they did with Firewire. It really should have been a much shorter exclusivity period.
It's more comparable to expressCard though. People are running GPUs over it in Windows without modifying the drivers by making Windows think it's running internally so it's PCIe but wrapped up in a way to be plug and play and put down a cable with displayport. There aren't as many devices need to be done over PCIe.
Intel invented PCI so they know what they're doing here. The first Intel certified motherboard (besides Apple's) only arrived May last year so the PC market hasn't even had it for a year. There's also not much point in implementing it just before Redwood Ridge in Q2 as it has displayport 1.2 support.
Some people want to see Thunderbolt fail because PCI slots are the one thing that always justifies a tower form factor. Some people want it to fail because it's Intel's product (despite PCI being Intel's too originally). If it was left up to a committee, we'd still be here in 2013 with expressCard.
Those standards are eventually passed off to 3rd party groups so maybe that's what would need to happen for TB to take off but they won't drive it forward quickly enough. If it hits 50Gbps in 2015, Intel can hand off the spec to the PCI-SIG to handle certification. They obviously have objected to the standard as it stands because the port is too big for tablets so there could well be a successor format but it doesn't make much difference because they have to allow it to be plug and play too so there's no reason the devices can't easily work on both like how Firewire 800 and 400 can go over the same cable.
They're including computers, motherboards and even PCI cards there but USB has a much wider audience than Thunderbolt because it's backwards compatible with USB 2.
The Thunderbolt promo site also included cables, motherboards too.
Third party chip makers supported USB 3 much sooner, so computers could have it a year before Intel bothered to offer it. Not offering timely support in their chipsets isn't helping TB, but because Intel couldn't stop it, USB 3 simply got off to a better start.
Even if you subtract Apple's one year exclusivity, USB compares better in my opinion in device support.
It's more comparable to expressCard though. People are running GPUs over it in Windows without modifying the drivers by making Windows think it's running internally so it's PCIe but wrapped up in a way to be plug and play and put down a cable with displayport. There aren't as many devices need to be done over PCIe.
If you weed out the drives with just one port the list diminishes a lot more than it should. Given the experiences I've seen users have with LaCie, I'd leave them off the list too. A premium brand with off-brand reliability and customer service is not something I care to tangle with.
Some people want to see Thunderbolt fail because PCI slots are the one thing that always justifies a tower form factor. Some people want it to fail because it's Intel's product (despite PCI being Intel's too originally). If it was left up to a committee, we'd still be here in 2013 with expressCard.
Also, some people associate it with being Apple's product, despite it not being true, and want it to fail for that reason.
Just to be clear, in case you don't know what I've said on the topic before, I do want Thunderbolt to succeed. I'm also over towers too. But TB's minimal uptake is disheartening.
"Last July, Ars Technica took a look at the state of the Thunderbolt standard developed by Apple and Intel, noting that adoption remained slow amid high pricing some eighteen months after the standard's introduction. The report did, however, suggest that pricing could begin to improve in [Q2] 2013 as the next generation of Thunderbolt chips hit the market.
Ars Technica has now published a follow-up report looking at how things have changed over the past six months, pointing to a number of improvements such as slightly lower pricing on Thunderbolt cables from Apple, the introduction of the first wave of optical cables supporting the standard, and the launch of new docking stations and other peripherals taking advantage of Thunderbolt.
The report indicates that the biggest holdup to wider adoption of Thunderbolt appears to be Intel's licensing and certification process, with the company dedicating only limited resources to helping third-party vendors bring their Thunderbolt products to market.
Several vendors we have spoke to over the past year have claimed that Intel was holding up the process, cherry picking which vendors it worked with.
Though Intel had effectively denied this characterization in the past, the company explained the situation a bit differently when we spoke at CES last week. Jason Ziller, Director of Thunderbolt Marketing & Planning at Intel, told Ars that Intel has "worked closely" with vendors it felt could "offer the best products" and could meet its stringent "certification requirements." The subtext seemed to be that Intel had limited resources to support and certify new products, and so it gave priority to devices that were perhaps more novel than those proposed by other makers.
Ziller indicated that Intel would be broadening its efforts this year, suggesting that more Thunderbolt products may be able to make their way into consumers' hands. Combining those efforts with continued cost drops and the first moves into Windows machines, Apple and Intel are undoubtedly hoping that 2013 will finally see Thunderbolt turn the corner."
With USB 3, you get a broad range of much of the same types of device - mostly hard drives. I can understand if Intel wanted to prioritise more unique applications of Thunderbolt rather than just end up with a batch of hard drives that perform the same over USB 3 and are cheaper. The Blackmagic 2.5k camera is one example:
[VIDEO]
They have HD-SDI output for 1080p but you get the full live 2.5k output over Thunderbolt (1:12 in the video).
These are things you can't do with USB 3. They have USB 3 displays but you can't run FW800 over USB 3.
I see them as complimentary interfaces: USB 3 for everything unless you need faster data transfer and advanced functionality. I wouldn't buy a Thunderbolt storage pen for example if the speed was covered by USB 3 and cheaper and even if it wasn't, I can use it on USB 2 machines.
I can see people having 1 Thunderbolt device vs 3-4 USB 3 devices just like with FW/USB2 but that's not a big problem. I don't think I'll ever have a TB kb/mouse or storage pen. Maybe a hard drive but there's not much point if it's fast enough on USB 3. It's really beneficial for RAID storage, docking solutions and high-end audio/video.
With USB 3, you get a broad range of much of the same types of device - mostly hard drives. I can understand if Intel wanted to prioritise more unique applications of Thunderbolt rather than just end up with a batch of hard drives that perform the same over USB 3 and are cheaper. The Blackmagic 2.5k camera is one example:
Unfortunately, most of the unique applications are also niche applications. It seems that a company of Intel's resources should have handled it better, sooner.
I see them as complimentary interfaces: USB 3 for everything unless you need faster data transfer and advanced functionality. I wouldn't buy a Thunderbolt storage pen for example if the speed was covered by USB 3 and cheaper and even if it wasn't, I can use it on USB 2 machines.
I can see people having 1 Thunderbolt device vs 3-4 USB 3 devices just like with FW/USB2 but that's not a big problem. I don't think I'll ever have a TB kb/mouse or storage pen. Maybe a hard drive but there's not much point if it's fast enough on USB 3. It's really beneficial for RAID storage, docking solutions and high-end audio/video.
They offer complimentary capabilities, I wasn't expecting USB 3 to be obsoleted in the least bit.
This story made BBC national news today. Why the delay I don't know, they seemed to think it had only just been announced.
They said it may be a problem for TV and film production companies and graphics art professionals. Maybe the Beeb are worrying that they won't be able to edit any TV programmes after March 1st.
The fan problem was mentioned as the problem.
When pressed the technology correspondent said Apple Pro users would probably stockpile machines until Apple bring out something new as there was no alternative for many users.
This story made BBC national news today. Why the delay I don't know, they seemed to think it had only just been announced.
They said it may be a problem for TV and film production companies and graphics art professionals. Maybe the Beeb are worrying that they won't be able to edit any TV programmes after March 1st.
The fan problem was mentioned as the problem.
When pressed the technology correspondent said Apple Pro users would probably stockpile machines until Apple bring out something new as there was no alternative for many users.
I read somewhere that the beeb was using a mixed environment but at some point completely switched to Mac only. With FCP as their main software editor. I wonder if there's anymore info to be found on what the beeb uses. I'd like to see numbers and editorial articles but am quite poor when it comes to using Google. For some reason they give me links to pages that haven't been updated for many many years. Yes, yes, I set the time, country and whatnot filters. It just doesn't seem to play nice with me.
I read somewhere that the beeb was using a mixed environment but at some point completely switched to Mac only. With FCP as their main software editor. I wonder if there's anymore info to be found on what the beeb uses.
Ah Marvin, you always come through! Thanks for the links. Good read up. I have actually read, and still read, all the Pro articles on Apple.com. Going to finish the video now. The guy is funny, telling the picture quality on the beamer is crap.
This story made BBC national news today. Why the delay I don't know, they seemed to think it had only just been announced.
They said it may be a problem for TV and film production companies and graphics art professionals. Maybe the Beeb are worrying that they won't be able to edit any TV programmes after March 1st.
Because all currently installed Mac Pros suddenly disappear like Cinderella's carriage at midnight?
Comments
deleted
It wouldn't be the norm to run an extra GPU card externally, the GTX 680 class of card is among the best and while you can get marginally better performance from other desktop cards vs the 2GB GTX 680M, it's not going to be worth investing in that. If you ran a whole load of things at once over Thunderbolt, you might hit a bottleneck somewhere but it's all bi-directional channels and out of all the tests done in various setups, they don't seem to have run into any show-stopping bottlenecks.
One good demo showing the power of it is the following where they are using 4K RED footage coming off a Promise RAID and a BlackMagic Ultrastudio to record the video itself and recording the screen session out to the RAID at the same time and still real-time 4K on a Macbook Air with a single Thunderbolt port:
[VIDEO]
It's not a setup anyone should adopt for that kind of work by any means but it's cool that it can do that and with dual Thunderbolt ports and Thunderbolt going to 20Gbps next year, the options are just going to get better.
Thunderbolt is only nearly 2 years old, had a 1 year Apple exclusivity and is proprietary to Intel and has a certification process. New technology tends to get an immediate resistance, especially when it conflicts with a 20 year-old status quo but it's enabling a lot of nice, affordable solutions so there's little reason to be against it.
The Mac Pro has a GPU in one of the x16 slots and uses up to 200W of power so that leaves about 100W and 1 x16 and 2 x4s for expansion. Dual TB on the iMac is like the 2 x4 slots so the MP offers the extra x16 but one good thing TB offers is plug and play so you can have a setup of peripherals for video editing and switch them for different peripherals for audio editing just by changing a plug.
It's not a good thing for anyone to have a reason to migrate away from the platform, even if it is less than 5% of Mac users and a new Mac Pro will give some of those people a reason not to move. I think some people will move because of FCPX alone because FCP would have been their only reason to use the platform in the first place.
Faster hardware will always give faster results but they always stop somewhere. No matter how many slots they put in, you'd still only be able to run 2 GPUs at most and it'll fall far short of systems that can run 8 GPUs:
[VIDEO]
Apple can't build a machine based on the needs of the highest imaginable task someone can use a computer for because there's no limit to that. They have to build a machine that hits a certain price range, has a sizeable market and has a future. They will never offer Intel's fastest CPUs for example because they are too expensive.
I doubt the Mac Pro they will introduce this Summer will be exactly the same as before, otherwise they'd have put in Sandy Bridge. The only question now is what compromises they've made.
Well, if indeed there is Thunderbolt in the next Mac Pro, and why not, I would hope that would be in addition to slots, and also multiple easily removable HD's and/or SSDs. So fast external peripherals should be possible there as well. Perhaps more power (watts), but who knows. Not to nitpick, but the Quadro K5000, the most powerful GPU for the Mac Pro draws 122 watts, not 200. Also 32GB of RAM is not enough for a lot of Photoshop work unless you want to be swapping data out of RAM on very large compositions. The current Mac Pro will go to 128GB of OWC memory. With the advent of cameras like the Nikon D800 for $2995, producing 75MB RAW images, layered compositions in PS can easily reach 500MB and more. That's a lot of swapping. Mac Performance Guide has great documentation on this.
I'm not saying that 27" iMac tricked out is not a good choice for many people, perhaps for most. It is over $3000, however, cheaper, but not by much over current models of MacPro like the Hexacore 3.33 Ghz with a decent GPU and the ability to upgrade to the best, internal HDs and lots of RAM. Most people in the business already have a computer monitor that is suitable. That being said there is a real need for a workstation class computer in the Mac lineup. Just hope we get one and not a box with a lot of Thunderbolt ports and little internal expansion capability. But anything would show that Apple still cares about the video market. That is important to software developers as well as customers.
Quote:
Originally Posted by Marvin
But do those things actually happen? You seem to decide these problems exist without seeing that they exist. The compute tests that have been done with the CS suite in OpenCL don't really back up what you're saying. You seem to be happy to keep price out of the equation too. The starting price for the Quadro K5000 is said to be $2249. That's more than a whole iMac with a GTX 680M.
Actually the only times I've looked at Quadro cards have been a bit after their release. They often drop in price fairly quick, and I compare first to what people are getting from the gaming cards. I've tested Mudbox on a couple different configurations (not all of them owned by me), including a macbook pro. Keep in mind I was talking about options. Note my prior comment about the Quadro K5000. It's just what they've brought out in the current generation so far. If you read my post below from another thread. I mentioned it as being one that is probably more commonly used in hero suites. They might be much more widespread at top end facilities, but I wasn't referring to those. A lot of this stuff just wasn't feasible on almost anything a couple years ago. In my 3d paint example, that's mostly driven by video memory. I don't think I would have tried to load up that kind of resolution in the past in a program that loads it all to video memory. As far as freelancers are concerned, Macs are reasonably popular but extremely inflexible as you just noted. Also you didn't note with NEC and Eizo, you do get more out of them. Eizo bundles the calibration software. Color Navigator has a lot of tracking functions. They also have a lot of compensation for panel uniformity. NEC does the same with colorcomp and spectraview. In the case of NEC you buy their colorimeter + software package for $400 or so. They're both good. If you were talking about graphic design or compositing and grading, I'd say they're worth it. Even the mini would have a lot of potential with a better gpu.
Quote:
Originally Posted by hmm
1GB wasn't even high when the mac pro 2010 revision debuted. That would be the area where it really irritates me. The common "average user just checks facebook" argument isn't exactly valid when we restrict the comparison to workstation level hardware. Check out the P90X of graphics cards. It's a pretty extreme comparison, but I'm not sure any of the current mac gpus including the 680mx could load even half of that without crashing. Companies probably buy hardware like that for hero suites so that it can be displayed for clients at full resolution without exporting to an offline renderer when a change is made. That's my guess.
Quote:
This is before you add a single PCI expansion card and the performance of that whole machine is still comparable to a $2200 iMac. The Mac Pro is better no question but not so much better to be worth a one man shop paying $7300 more.
I always ended up upgrading in bits and pieces. The display thing was due to consistent irritations with cheaper displays. I don't recall ever suggesting the dual ssd cards there, although I own a couple ssds, one of them being a 512. I haven't personally suggested 64GB of ram. Adobe suggests that for After Effects rendering with a 16 core machine to render on all cores. This is due to hyperthreading. 16*2 = 32 logical cores * 2GB/core. If you're going with a 16 core machine, the extra cost is probably doable. 32GB isn't uncommon for a new machine today. You can achieve that on the imac. Prior to 2012, if you really had high ram requirements, it was definitely in favor of the mac pro due to the price disparity between 4GB and 8GB dimms. A mac pro with 32GB was cheaper than the same in an imac.
Quote:
So are you saying the 2GB Quadro 4000 you previously decided would be your choice is no longer good enough?
You know I've been waiting for video memory to increase for a while. Some applications hitting the market in recent years have relied more on the gpu where they're loading huge files onto it. Once you go past a certain size, you notice a huge drop in frame rate where orbiting around something becomes really choppy. I wouldn't want to work like that for hours at a time. The workstation cards have different driver tuning. I'm not sure how they compensated, but it doesn't seem to get as bad. You can hit a wall either way. Note the link above. You'll see why I said it seemed to be designed for hero stations. You can work with a ton of resolution without it bogging down. Several years from now, that will probably be doable in desktop graphics.
Anyway back to the question regarding one man shops. Typically a slightly customized solution is ideal. Some of them need a lot of power to support their workflow. Otherwise low end probably something like a 3770k. Otherwise go for one of the more cost effective solutions in a 6 or 12 core LGA2011 package depending on available configurations. If you use anything that is heavy on the gpu, in gaming cards I'd go with something like this assuming drivers are stable in the applications in question. Ram is a much bigger help for these things than the clock speed of the card once you're using at least a mid range card. Driver stability is often an issue in 3d apps, even under OSX. There are obviously other things. An example would be if you use After Effects or Premiere, you're reliant on CUDA. This means if Apple releases all AMD this year (as they seem to alternate) you're waiting that one out. If the gpu drivers are flakey or don't support a specific hardware feature you want, it's whatever is available in the Quadros. Unfortunately there is no Quadro K4000 at this point. It might show up later. On ram I'd suggest using 8GB dimms. Install 2. Memtest them prior to placing the machine back in service. If the system still hits the disk frequently, install 2 more bringing it to 32GB. An SSD boot drive is only critical if it has to frequently read information on that drive or it needs to double as a scratch disk for whatever app (if 32GB isn't enough). If you're doing anything that involves a lot of color, I'd at least go with NEC. It's easier to keep them fairly consistent over time with their LUT system. They perform better over their brightness range than Apple, which always seems to tune for max brightness. If you ever have to bring images up side by side to check changes, you learn the value of uniformity pretty quickly. It's annoying when two can be identical yet look a bit different. In terms of price, you can get a PA271w around the price of a thunderbolt display and just use a mini displayport --> displayport cable.
Does that clarify my opinions a bit better? They're fairly consistent, although re-reading I can see where you might have perceived a conflict.
Quote:
Yes, it's very annoying. Having to type purge into the terminal regularly shouldn't be necessary.
I hate that. I hate it when finder corrupts. I hate HFS+ and the continued need for disk warrior with optical drives phasing out. I hate the lack of full displayport 1.2 support. There are many things about OSX that genuinely piss me off. I hate spotlight conflicts with hidden scratch disk files. It makes me mad that Apple keeps adding features rather than fixing old problems.
I'm just going to add what I dislike about thunderbolt. It's proprietary rather than something that is easy to implement across various classes of devices. It's more limiting than you would suggest. It's not the same as PCI. Thunderbolt is its own protocol. I'm amazed people are this surprised that it hasn't caught on. Apple claimed they removed the express card slot because hardly anyone used it. What would make thunderbolt any different outside of available Apple's dedicated peripherals? I would have preferred to see a higher bandwidth PCI spec dedicated to notebooks "and" mobile devices that can be implemented across various designs rather than tied to intel products.
Valid points.
A different thing, but still, is Aperture. In 4 days it's 3.0 version will be 3 years old. Many dot releases have come since, but there are many feature requests being ignored. Good to get updates for free these past years, but I'd rather pay for a 'catchup to the competition'
They can't put in TB; that requires a new CPU that Intel simply doesn't make.
That is very true, many people use way more than that. See here
That's not true; OSX only uses 96GB. Take it from someone that knows this stuff:
http://macperformanceguide.com/blog/2012/20120106_1-OWC-16GB-modules.html
A MP does take 128GB, and Windows sees all of it. Just not OSX.
That might come about but TB has already gained a fair amount of traction in hardware:
http://www.pcauthority.com.au/News/296786,expect-hundreds-of-thunderbolt-devices-says-intel.aspx
Once Intel takes more control of the motherboard market with Broadwell, it'll be adopted a lot more.
Don't worry, Facebook has you covered:
https://www.facebook.com/WeWantAperture4
Not necessarily a custom CPU but it would likely need some form of GPU soldered onto the motherboard in addition to any PCI cards. It's not essential for them to include Thunderbolt on the Mac Pro but it would allow easily sharing peripherals between it and other machines.
They said they were doing a big photo comparison using loads of layers, which sounds like loads of stills next to each other to compare them. If you set your render cache in After Effects to 90% of say 96GB RAM and hit play on a long movie, it'll fill it up too . You can even make a RAM disk and put your whole filesystem on it but being able to use it doesn't make it a requirement. It would be a better use of RAM not to put photos for comparison into a single comp but to use an app meant for photo comparisons.
There are massive image comps like the gigapixel images you see. There's one here done on a machine with 8GB RAM:
http://www.wired.co.uk/news/archive/2009-12/21/18-gigapixel-panorama-offers-breathtaking-view-of-prague
192,000 by 96,000 pixels, Photoshop file was 120GB, built from 600 photos. 8GB shouldn't be the aim - that was done in 2009 - but it's possible to do these things with much lower resources. The higher resources just make them easier to do.
Once you do things like that, it's better having a database-like setup. Like how FCPX uses a database instead of loading an entire file into memory or how running Google Maps doesn't load the entire world into your web browser. Photoshop should allow you to open a folder of images like a .bundle and it would just load the tiles you were looking at. It would be quicker to dynamically load things like that than put everything in RAM.
Quote:
Originally Posted by Jim W
Well, if indeed there is Thunderbolt in the next Mac Pro, and why not, I would hope that would be in addition to slots, and also multiple easily removable HD's and/or SSDs. So fast external peripherals should be possible there as well. Perhaps more power (watts), but who knows. Not to nitpick, but the Quadro K5000, the most powerful GPU for the Mac Pro draws 122 watts, not 200. Also 32GB of RAM is not enough for a lot of Photoshop work unless you want to be swapping data out of RAM on very large compositions. The current Mac Pro will go to 128GB of OWC memory. With the advent of cameras like the Nikon D800 for $2995, producing 75MB RAW images, layered compositions in PS can easily reach 500MB and more. That's a lot of swapping. Mac Performance Guide has great documentation on this.
It varies. OSX is absolute shit at memory management, but even at 16 bpc with a large comp that saves out a 2-3GB compressed psb, 32GB should not generate slow results. I've tested this with 10-15k comps, spherical panos, and layered texture maps. You just max out ram and add scratch disks. You should not experience brush lag or find it impossible to apply adjustments due to memory. If the objective is photoshop filter drag racing, that doesn't align well with any typical workload I've ever seen. It was a 32 bit app on OSX until CS5 anyway where scratch disks were even more crucial with inflated file sizes. Now you just push as much of that as possible onto ram, but the mechanic is still present in its programming. Further I can attest to the need for a lot of memory at 32bpc, but support for floating point formats in photoshop is terrible. They don't prioritize it at all, so you're left with a very limited toolset. It's a very stagnant program compared to those used in video compositing (nuke, fusion, after effects). They keep adding new poorly implemented features. Further D800 files aren't that large if you had to deal with P45 files at the end of the PowerPC era. These were around 110MB as 8 bit tiffs or twice that at 16. Obviously add layers and the sizes ballooned quickly. I will say this again. 500MB is nothing. People dealt with that on G4 towers, even if it wasn't pleasant.
Quote:
Originally Posted by PhilBoogie
They can't put in TB; that requires a new CPU that Intel simply doesn't make.
That isn't true. What is true is that they can't route it through integrated graphics. There are certain placement requirements in that it demands specific logic board space, but anyone in charge of engineering might present several possible solutions to how this might be routed. It might require some form of embedded graphics if it is seen as a priority. Frankly none of the thunderbolt peripherals were designed with the mac pro in mind. Note the short display cord.
Quote:
Originally Posted by Marvin
That might come about but TB has already gained a fair amount of traction in hardware:
http://www.pcauthority.com.au/News/296786,expect-hundreds-of-thunderbolt-devices-says-intel.aspx
Once Intel takes more control of the motherboard market with Broadwell, it'll be adopted a lot more.
Don't worry, Facebook has you covered:
I'd like to see something that could be implemented in future tablets and phones as well regardless of their reliance on intel and not encumbered by Apple's licensing clauses for the connector. Phones are still pretty slow in terms of transfer rates, but they have been improving. I dislike specialized solutions until you're at a level where only a smaller number of individuals really needs that capability. I've mentioned SAS. It serves a purpose, yet I don't think every computer requires a SAS or mini-SAS HBA.
Quote:
There are massive image comps like the gigapixel images you see. There's one here done on a machine with 8GB RAM:
http://www.wired.co.uk/news/archive/2009-12/21/18-gigapixel-panorama-offers-breathtaking-view-of-prague
192,000 by 96,000 pixels, Photoshop file was 120GB, built from 600 photos. 8GB shouldn't be the aim - that was done in 2009 - but it's possible to do these things with much lower resources. The higher resources just make them easier to do.
Once you do things like that, it's better having a database-like setup. Like how FCPX uses a database instead of loading an entire file into memory or how running Google Maps doesn't load the entire world into your web browser. Photoshop should allow you to open a folder of images like a .bundle and it would just load the tiles you were looking at. It would be quicker to dynamically load things like that than put everything in RAM.
I'll admit I've never done anything near that kind of size. This isn't uncommon though. As for databases, it's also possible to make formats that are memory mappable. In other words readable directly to virtual memory rather than loaded into ram and immediately swapped due to size. CS6 actually recommends at least 8GB of ram, but it can sometimes require significantly more to behave properly. I'd usually say add it until things no longer lag. The performance per dollar up to a certain point is much lower than the cost of ssds. If it's maxed add an ssd. Part of the problem is OSX and its terrible memory management. It's necessary not to drift off into hyperbole here. Still photography hasn't evolved much in recent years. Pretty much anything you can do today was done on PowerPC era hardware and graphics workstations prior to that. It was a lot slower on a G4 or G5, but you could do many of those things. You'd keep active indexing things like spotlight off, limit history states to something sane, use fast scratch disks, and be as efficient as possible with layer sets (they weren't called groups back then). What you're describing with databases is something like a mipmapped setup. Google maps has things cached at different levels. When you're zoomed out it's can load different files.
I really don't think they have anything resembling traction. I think they should have already been at the "hundreds of devices" by now, it's been nearly two years already.
USB 3 was quicker to get there at around 9 months - USB 3 only arrived January 2010 so I guess Intel wasn't all that far behind in supporting it:
http://hothardware.com/News/USB-30-Sees-Certified-Product-List-Soar-Beyond-100/
They're including computers, motherboards and even PCI cards there but USB has a much wider audience than Thunderbolt because it's backwards compatible with USB 2.
Apple's 1 year exclusivity with Thunderbolt was quite damaging because non-Apple people started to hate the standard just like they did with Firewire. It really should have been a much shorter exclusivity period.
It's more comparable to expressCard though. People are running GPUs over it in Windows without modifying the drivers by making Windows think it's running internally so it's PCIe but wrapped up in a way to be plug and play and put down a cable with displayport. There aren't as many devices need to be done over PCIe.
There are a few storage devices:
http://www.tomshardware.com/reviews/my-book-thunderbolt-duo-pegasus-r4-2big,3222.html
other products are here:
https://thunderbolttechnology.net/products
Products like the Magma 3T let you use all the cards listed here:
http://www.magma.com/thunderbolt-compatibility
Intel invented PCI so they know what they're doing here. The first Intel certified motherboard (besides Apple's) only arrived May last year so the PC market hasn't even had it for a year. There's also not much point in implementing it just before Redwood Ridge in Q2 as it has displayport 1.2 support.
Some people want to see Thunderbolt fail because PCI slots are the one thing that always justifies a tower form factor. Some people want it to fail because it's Intel's product (despite PCI being Intel's too originally). If it was left up to a committee, we'd still be here in 2013 with expressCard.
Those standards are eventually passed off to 3rd party groups so maybe that's what would need to happen for TB to take off but they won't drive it forward quickly enough. If it hits 50Gbps in 2015, Intel can hand off the spec to the PCI-SIG to handle certification. They obviously have objected to the standard as it stands because the port is too big for tablets so there could well be a successor format but it doesn't make much difference because they have to allow it to be plug and play too so there's no reason the devices can't easily work on both like how Firewire 800 and 400 can go over the same cable.
The Thunderbolt promo site also included cables, motherboards too.
Third party chip makers supported USB 3 much sooner, so computers could have it a year before Intel bothered to offer it. Not offering timely support in their chipsets isn't helping TB, but because Intel couldn't stop it, USB 3 simply got off to a better start.
Even if you subtract Apple's one year exclusivity, USB compares better in my opinion in device support.
I agree that the uptake is comparable to or maybe better than expressCard. It's not a flattering comparison though.
There are devices, but that's not a great variety.
The official promo site listing all the products:
https://thunderbolttechnology.net/products
Once you take out the computers and cables, it's not all that much.
The broadest category is storage, here is that list:
https://thunderbolttechnology.net/products?tid=1&field_company_nid=All&field_prod_os_value_many_to_one=All
If you weed out the drives with just one port the list diminishes a lot more than it should. Given the experiences I've seen users have with LaCie, I'd leave them off the list too. A premium brand with off-brand reliability and customer service is not something I care to tangle with.
Also, some people associate it with being Apple's product, despite it not being true, and want it to fail for that reason.
Just to be clear, in case you don't know what I've said on the topic before, I do want Thunderbolt to succeed. I'm also over towers too. But TB's minimal uptake is disheartening.
There have been articles recently suggesting that Intel will try to improve this:
http://www.macrumors.com/2013/01/15/thunderbolt-adoption-reportedly-slowed-by-intels-licensing-and-certification-but-improvements-coming/
"Last July, Ars Technica took a look at the state of the Thunderbolt standard developed by Apple and Intel, noting that adoption remained slow amid high pricing some eighteen months after the standard's introduction. The report did, however, suggest that pricing could begin to improve in [Q2] 2013 as the next generation of Thunderbolt chips hit the market.
Ars Technica has now published a follow-up report looking at how things have changed over the past six months, pointing to a number of improvements such as slightly lower pricing on Thunderbolt cables from Apple, the introduction of the first wave of optical cables supporting the standard, and the launch of new docking stations and other peripherals taking advantage of Thunderbolt.
The report indicates that the biggest holdup to wider adoption of Thunderbolt appears to be Intel's licensing and certification process, with the company dedicating only limited resources to helping third-party vendors bring their Thunderbolt products to market.
Several vendors we have spoke to over the past year have claimed that Intel was holding up the process, cherry picking which vendors it worked with.
Though Intel had effectively denied this characterization in the past, the company explained the situation a bit differently when we spoke at CES last week. Jason Ziller, Director of Thunderbolt Marketing & Planning at Intel, told Ars that Intel has "worked closely" with vendors it felt could "offer the best products" and could meet its stringent "certification requirements." The subtext seemed to be that Intel had limited resources to support and certify new products, and so it gave priority to devices that were perhaps more novel than those proposed by other makers.
Ziller indicated that Intel would be broadening its efforts this year, suggesting that more Thunderbolt products may be able to make their way into consumers' hands. Combining those efforts with continued cost drops and the first moves into Windows machines, Apple and Intel are undoubtedly hoping that 2013 will finally see Thunderbolt turn the corner."
With USB 3, you get a broad range of much of the same types of device - mostly hard drives. I can understand if Intel wanted to prioritise more unique applications of Thunderbolt rather than just end up with a batch of hard drives that perform the same over USB 3 and are cheaper. The Blackmagic 2.5k camera is one example:
[VIDEO]
They have HD-SDI output for 1080p but you get the full live 2.5k output over Thunderbolt (1:12 in the video).
These are things you can't do with USB 3. They have USB 3 displays but you can't run FW800 over USB 3.
I see them as complimentary interfaces: USB 3 for everything unless you need faster data transfer and advanced functionality. I wouldn't buy a Thunderbolt storage pen for example if the speed was covered by USB 3 and cheaper and even if it wasn't, I can use it on USB 2 machines.
I can see people having 1 Thunderbolt device vs 3-4 USB 3 devices just like with FW/USB2 but that's not a big problem. I don't think I'll ever have a TB kb/mouse or storage pen. Maybe a hard drive but there's not much point if it's fast enough on USB 3. It's really beneficial for RAID storage, docking solutions and high-end audio/video.
Unfortunately, most of the unique applications are also niche applications. It seems that a company of Intel's resources should have handled it better, sooner.
They offer complimentary capabilities, I wasn't expecting USB 3 to be obsoleted in the least bit.
This story made BBC national news today. Why the delay I don't know, they seemed to think it had only just been announced.
They said it may be a problem for TV and film production companies and graphics art professionals. Maybe the Beeb are worrying that they won't be able to edit any TV programmes after March 1st.
The fan problem was mentioned as the problem.
When pressed the technology correspondent said Apple Pro users would probably stockpile machines until Apple bring out something new as there was no alternative for many users.
I read somewhere that the beeb was using a mixed environment but at some point completely switched to Mac only. With FCP as their main software editor. I wonder if there's anymore info to be found on what the beeb uses. I'd like to see numbers and editorial articles but am quite poor when it comes to using Google. For some reason they give me links to pages that haven't been updated for many many years. Yes, yes, I set the time, country and whatnot filters. It just doesn't seem to play nice with me.
/rant
There are older links about their workflows here:
http://alex4d.wordpress.com/2009/02/17/broadcast-video-expo-bbc-on-fcs/
http://www.apple.com/uk/pro/profiles/fullonfood/
The following video from 2 years ago at 10:55 details some interesting numbers:
[VIDEO]
2500 seats of FCP and more than that of Avid and they use Premiere as the desktop systems. I doubt they are all Mac but they have a fair amount:
http://www.dailymail.co.uk/news/article-2058665/BBC-splashes-500-000-worth-iPhones-iPads-Apple-Macs--2000-staff-face-axe.html
Obviously they need at least 2500 Macs for the FCP seats but they show a MBP on the slide in the video.
The news will warm up as we get closer to March 1st. OMG it's just two weeks away. #panicbuy4kMacPro
I doubt there will be much uproar about this when March 1st arrives. People know a new one is coming anyway.
Ah Marvin, you always come through! Thanks for the links. Good read up. I have actually read, and still read, all the Pro articles on Apple.com. Going to finish the video now. The guy is funny, telling the picture quality on the beamer is crap.
I like all these Mac Pro threads, and articles.
Thanks,
Phil.
Because all currently installed Mac Pros suddenly disappear like Cinderella's carriage at midnight?
Too funny! I don't understand all this talk either. Happened with other products as well. Hardware and software. I'm left in a state of oblivion ¡
Quote:
Originally Posted by tylerk36
Maybe a power consumption issue?
Yeah, that stupid UK power consumption law that prevents Apple from selling the existing Mac Pro into the UK. Freaking UK, what's their problem?
I wonder what the PC mfg are gonna do.. Stop selling towers PCs into the UK?