Adobe Photoshop CS4 Will Be 64-Bit For Windows Only

HighWizard notes that Adobe Systems has shared the first scrap of information about its next version of Photoshop, CS4, and it's a doozy: there will be a 64-bit version of the photo-editing software, but only for Windows Vista and not for Mac OS X. Ars explains the history of how this conundrum came to pass — blame Apple and/or Adobe as you will.

Comments

Re:64 bit is no panacea

[joaommp]

Well, it will run faster if you have a large pool of physical memory and do some heavy Photoshop editing, because Photoshop will be able to access more than 3GB of memory (remember that 1GB of the 4GB address space is already reserved for system code sharing) and not resort to it's own swap/disk cache system as much.

bad summary - there will be a 32-bit version

[Anonymous Coward]

You misread the article:

The Lightroom news naturally raises the question: What's Adobe doing with Photoshop? In the interest of giving customers guidance as early as possible, we have some news to share on this point: in addition to offering 32-bit-native versions for Mac OS X and 32-bit Windows, just as we do today, we plan to ship the next version of Photoshop as 64-bit-native for Windows 64-bit OSes only.

Re:bad summary - there will be a 32-bit version

[langelgjm]

You misread the summary (so did I, at first). It's not "Adobe Photoshop CS4 Will Be For 64-Bit Windows Only," it's "Adobe Photoshop CS4 Will Be 64-Bit For Windows Only."

But AMD64 could be...

[mrchaotica]

Remember, going to 64-bit on x86 can make programs faster, but not because of the extra bits. The speedup comes from the fact that, in addition to increasing the bits, AMD also added a bunch of extra registers to the spec.

Re:What will happen?

[john82]

As a matter of fact, Slashdot once again misleads with the choice of headline and half-the-story lead-in. Just a bit of reading reveals:

On the other hand, we work very hard at maintaining parity across platforms, and it's a drag that the Mac x64 revision will take longer to deliver. We will get there, but not in CS4. (Our goal is to ship a 64-bit Mac version with Photoshop CS5, but we'll be better able to assess that goal as we get farther along in the development process.) Hmmm. Not the end of the world after all.

Re:64 bit is no panacea

[Kjella]

Yep, and memory prices have dropped *extremely* over the last year. If I was working with many and large photoshop images, getting 4x4GB memory wouldn't be out of the question. Honestly I don't need it, but if you're working with high-quality print images I can easily see why you might need that...

Use QT, like the rest of the world

[SwiftX]

Trolltech(/ nokia) is working with Apple to get QT on MacOSX using Cocoa.
Problem solved!
SwiftX

Re:64 bit is no panacea

[alta]

Unless I'm mistaken the only thing 64bit color has to do with 64bit processing is that they both start with the number 64.

64Bit will allow the computer to deal with more data at a time, no matter what the color depth of the file is... It'll let the program have more memory. That will help a 64bit image if it's BIG, but just because it's BIG, not because it's 64 bit.

Re:I vote Apple

[jcupitt65]

You're thinking of Tiger, I think. Leopard is fully 64-bit. http://www.apple.com/macosx/technology/64bit.html

Re:The blame falls solely on Apple

[Teese]

Apple have never promised 64-bit Carbon. They did promise 64bit carbon, during the 2006 WWDC. It wasn't until the 2007 WWDC that they rescinded the promise. Before the 2007 WWDC, they backed up the promise with seeds with 64bit carbon support in. They removed that 64bit carbon support in the 2007 WWDC seed. Of course they also slightly redefined what carbon meant. It now means the GUI portions of what used to be called carbon. So there are parts of "carbon" that are 64bit. They just aren't called carbon anymore.

Re:What will happen?

[bhima]

Just want to correct one thing: Aperture is not a replacement for photoshop it is a competitor of Adobe Lightroom. Apple doesn't have a direct replacement for photoshop.

You are still wrong...

[hummassa]

Okay, stipulated. Regardless, it's the memory address space that allows for the speed improvement when working with large files, not new JMP routines. JMP routines? Actually, the extra register things is capable of optimizing out _many_, _many_ memory accesses... leaving the path clear for the SIMD instructions to fetch repeatedly only the data that your extended addressing is capable of. Imagine (simplifying a little) some transform being done to an image, that alters some data:

for( all pixels in the image ) { x1 = red(pixel); x2 = (x2 + x1 *2 + 3) % MAX; blue(pixel) = x2 }

if x1, x2 are put in registers then your transform will fetch only the pages where the pixel values are; if x2 is in memory, then _each_ fetch of a page where a pixel are is interleaved with one fetch and one write of the page where x2 is. This means that the operation becomes probably three to four times slower.

Re:You are still wrong...

[fitten]

To add to your post, the folks at Cinebench were very happy transitioning from 32-bit land to 64-bit land (x86-64) because they were able to get rid of a lot of cruft and make use of the registers and such to achieve a significant speedup with their 64-bit version over their 32-bit version.

Also, in 32-bit land, you can use blocking algorithms to get by memory limitations. Not all operations must be done over the entire file, requiring all the data be in memory at the same time so it isn't like 32-bit can't do what 64-bit can memory-wise.

Re:64 bit is no panacea

[baadger]

x86_64 is more of a cleanup to the aging x86 ISA. Not only does it future proof the architecture against big memory requirements but it also does away with ancient segmented addresses, provides more CPU registers (leading the the possibility of more specialized register operations, I assume) and generally allows people to break ABI *as a matter of course* which is great because on AMD64 arch's you can *assume* the presence of MMX, SSE, and SSE2 instruction sets. Even Microsoft, anally retentive back-compat evangelists, took the opportunity in Windows XP x64 and Windows 2003 x64 Server to introduce further kernel mode memory protections ('PatchGuard')

No x86 software, *including drivers*, should be shipping in both 64bit and 32bit binary form, all of the problems you mention with 64bit are essentially proprietary software exclusive btw, and just highlight the highly broken software ecosystem Microsoft Windows has fostered.

Re:64 bit is no panacea

[jocknerd]

Yep, Adobe was told that Cocoa was the future for OS X development. And yet, they chose to stick with Carbon. Can't blame Apple for that one. Simple fact is Photoshop is designed for Windows first and then ported. So its not a native Mac app and doesn't take advantage of all the technology in OS X. If you want to see what I'm talking about, take a look at Pixelmator. While its not on par with Photoshop, it combines the power of ImageMagick with the underlying technology from OS X.

Re:64 bit is no panacea

[MidnightBrewer]

Photoshop's legacy is all Apple. If you've ever tried to use Photoshop in Windows from, say 4.0 on (about the time I tried using it on Windows, instead of my Mac, where I started at 2.5), you'd know how horribly awkward it was compared to its smooth operation on the Mac OS. Adobe has been criticized on the Windows side for having a Mac-like interface as well, which Windows users have been forced to just get used to.

Adobe has a long legacy of making sure their application is rock-solid and reliable before releasing. They, of any company, were the ones to set the bar for what it's like to release incredibly stable, bug-free software without any major point release dramas to fix glaring mistakes.

Yes, ever since they decided it was a good idea to put a salesman at the head of the company who believes that it's not what you sell, but how you sell it that's important, they've started going down a bad road. Their subsequent attempts to lock the system down with draconian DRM etc. has not improved their image much. However, there is no way to compare Pixelmator to Photoshop. Pixelmator can afford to be nimble because it has no expectations to live up to and is a relatively limited application with little utility in comparison.

Photoshop is a (reasonably) well-thought-out system of utilities and tools that have always done a darn good job working within the limits of the processor and memory and still manage to offer speedy, capable performance and a comprehensive 3rd-party plugin architecture. Adobe's main crutch is that they can't afford to simply throw away doing things the way they've always done in order to move forward; their own success has locked them into evolutionary, rather than revolutionary, progress.