Uh, what ? OSX was the *last* major OS to the multiprocessor game. 10.4 put it about on par with XP and Linux 2.4-series kernels.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Mac OS X 'Snow Leopard' to Focus on Performance, Quality
- Thread starter MacRumors
- Start date
- Sort by reaction score
Uh, what ? OSX was the *last* major OS to the multiprocessor game. 10.4 put it about on par with XP and Linux 2.4-series kernels.
This is also what Psystar and Apple have been working on for a while now... it's OS X for the generic PC that lack that nice multi touch display.
With Hyperthreading you can take advantage of execution units in a CPU that would normally be sitting idle. Hyperthreading doesn't give you the equivalent of another full CPU but it certainly gives you increased crunching power for very little power. Effectively you are utilizing more of the CPU.
Encoding video shows a nice improvements with HT and some games also improve. I expect that Apple will attempt to thread many parts of the OS so that OS X remains as responsive as ever under load.
You certainly aren't derailing this thread at all. Grand Central is all about managing threads for multi core CPU. It's more important than people realize.
Does Vista do better than Leopard with multiple cores? So maybe OSX needs this update to improve lacking multicore support, I don't know, but it certainly should be useful!
Yay $129 for Leopard SP1
Now that this has been announced, I would bet good money that sales of Leopard retail discs are going to drop like flies. What's the point of buying Leopard at $129 when the OS it SHOULD have been will be out in about a year?
Makes no sense. I could say the same thing about you. Next year in the June timeframe mosx will be the person he "SHOULD" have been in June of 08. That statement never makes sense. Evolution moves one a path that can't necessarily sped up at all times.
I'd be surprised if Apple charges over $80 for Snow Leopard. It would behoove them to get as many people on Snow Leopard as possible since it is the platform that will enable more future features.
The current Quicktime will do some of that. For example, the
other day I was viewing a .avi with huffyuv encoded content.
Have you tried the Perian quicktime component(s)?
Here are some pictures from the OS X State of the Union Address. Looks like 10.6 is a hot topic there-- I wonder why?!
autrefois
macrumors 65816
From your lips to Steve's ears! But I doubt it will be a free upgrade. I would be willing to pay some money for the update because I'm aware of the amount of time and money goes into programming. The past few OS upgrades have touted 100s of new features. If it's only a few new features, then I would expect a lower price because I'm (mostly) paying for what they should have gotten right the first time.
Good points. This is why, if it is mostly maintenance with a few new features, I don't think it should be a full-price upgrade. I agree that it's unlikely that it would be equivalent to a 10.5.x release. It would be more substantial than that, but it's not sounding like as substantial as going from 10.3 to 10.4, or 10.4 to 10.5
We don't know the price points, or what new feature(s) there will be, and it appears the press release from the Canadian site was an oops since it's down now. So this whole thread is mostly speculation at this point.
If they offer a number of substantial new features or substantial under the hood improvements that significantly improve the OS, then I think a full upgrade price would be warranted. If they are mostly fixing what's wrong now and tweaking it so it's faster/more efficient/etc., then I don't think people should be expected to pay full price (maybe as a couple of people have mentioned, full price for new or Tiger users, half price for Leopard, for example).
With the dollar doing how it is though, by the time Snow Leopard is released, $129 might not be a whole lot of money anyway.
With the exception of 10.1 which was free, hasn't every OSX release been $129?
Until apple announces otherwise, people are probably going to assume it will cost the same as 10.0, 2, 3, 4, and 5. And I'd understand if they didn't announce specific pricing, since it is early. But they certainly could either say it will be $129, it will be cheaper, or it will be free. Until they make a statement about it, any discussion of 10.6 will be dominated by talk of price.
And people WILL consider this a service pack, both the public and the press. Sure, it might be a better one than MS typically releases, but the arguments that it's not are mostly semantics. Really, apple should suck it up and just admit it is a service pack so people can move on.
So any word today from the session about PPC support? Honestly, that's the only part of this that interests me much.
What's the problem? Why can't the people complaining about the price of Snow not just buy it. Apple's not saying Leopard is 'bug' ridden just that they're making it faster. No new features means that if you're happy with Leopard skip Snow. I am skipping Leopard and getting Snow
Yay $129 for Leopard SP1
Now that this has been announced, I would bet good money that sales of Leopard retail discs are going to drop like flies. What's the point of buying Leopard at $129 when the OS it SHOULD have been will be out in about a year?
Sorry, I didn't see the part where they said it was $129. Please, enlighten me where they claimed that.
Interesting that you assume that dropping PPC support will improve the quality, performance and stability of the code. This seems a common meme but I can tell you as someone who has written an awful lot of cross platform C code, the best code is code which compiles and runs well on a range of platforms. Sure, some machine specific optimisations are useful but by and large you will squash more bugs and produce better quality code if you make sure it compiles and runs on as many different architectures as possible. That goes for 32/64 bit and different CPUs altogether.
Dropping support for other architectures did not help the NT kernel get more efficient did it?
You can always get it illegally if you don't feel like paying up for the speed upgrade next year.
Don't confuse SMP with Multi-Core/Thread managed from the Kernel to the Userspace Applications leveraging available Core cycles based upon Parallel Programming paradigms.
They aren't remotely the same.
Grand Central
I just read over that press release, and i think the most interesting thing said was that Grand Central will help developers CREATE apps that take advantage of multicore machines. Up until now, one of the big problems with multi-cores is how hard it is to write programs that parallelize well. Basically, we simply aren't seeing the full benefit of multiple cores because it is significantly harder to make use of all of them a once.
It would be a killer feature for developers if they have created a library or something that would make parallel programming simple. When i read about grand central, it made me wonder if apple has created a way to make PROGRAMMING for multi-core easy, or if it is something like a change in the scheduler
Edit:
has there ever been an OS X that was sort of a Developers release? Could it be possible that this comes out at a low price, mainly to grab developers and get them ready for a more consumer geared release following shortly after snow leopard?
I just read over that press release, and i think the most interesting thing said was that Grand Central will help developers CREATE apps that take advantage of multicore machines. Up until now, one of the big problems with multi-cores is how hard it is to write programs that parallelize well. Basically, we simply aren't seeing the full benefit of multiple cores because it is significantly harder to make use of all of them a once.
It would be a killer feature for developers if they have created a library or something that would make parallel programming simple. When i read about grand central, it made me wonder if apple has created a way to make PROGRAMMING for multi-core easy, or if it is something like a change in the scheduler
Edit:
has there ever been an OS X that was sort of a Developers release? Could it be possible that this comes out at a low price, mainly to grab developers and get them ready for a more consumer geared release following shortly after snow leopard?
Agreed. It's a fallacy. Badly written C is badly written C whether the IFDEF are 64 bit aware or 32 bit aware.
Gutting Carbon and leaving a 32bit VM for Carbon apps to play in while the rest of the OS moves forward with Cocoa APIs [whether 32 bit Cocoa or 64 bit] will trim the fat and allow for better QA cycle testing suites to manage the code base.
It did when I worked at NeXT and it did at Apple.
I'm thinking about this and thinking back a long ways to when new chips came out fairly quickly in the 80's. The 80286 to 80386 to 80486, each incrementally better and people went bonkers when the new machines shipped and everyone had to upgrade.
Then you tie-in the "upgrade time curve" whereas I think on the PC side it's something like every 3 years and on the Mac side it's something like 5 years (don't quote me...just surmising). Doing the math, with Y2K being a huge upgrade year (mainly for PC users and then Mac people followed suit trying to keep up with their neighbor) that says PC users will be looking to upgrade in the next year or so and Mac people aren't too far behind.
So if this Nehalem chip is as hot as it sounds, and Apple does some incredible optimizations to take advantage of it, this becomes a huge, huge selling point for new Mac sales. Maybe they're looking beyond people upgrading but more at a mass exodus from their existing hardware (especially from Vista users who've they've been targeting in ads). Just another angle...
Oh yeah, and Apple, if you do anything to the GUI please get rid of the last remaining Gel controls. They're really, really old!
Yeah I agree with most here, if it's a $130 upgrade for practically a service pack, I'll be very upset.
It should be free.
It should be free.
Yay $129 for Leopard SP1
Now that this has been announced, I would bet good money that sales of Leopard retail discs are going to drop like flies. What's the point of buying Leopard at $129 when the OS it SHOULD have been will be out in about a year?
I'm pretty sure that most people who would have upgraded to Leopard already have and everyone else gets it with a new Mac.