Just the parts you want to utilize GPGPU, luckily. Ant the results, although limited to certain fields, are quite impressive from what I've seen. A Radeon 4350 that costs 35 bucks on Newegg compressed their test video in 38 seconds, while a Nehalem 3.2 GHz quad-core CPU that costs $1000 took 100 seconds. GPU-encoded videos unfortunately had problems and glitches quite often, so the software still needs polishing, but the speed is impressive.
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.
Windows 7 to me is going to be expensive!
- Thread starter waloshin
- Start date
- Sort by reaction score
wow that is very lucky!! still would be a bitch though. i cant even program basic java/c++ applications, imagine making something useful! (i do networking not programming lol).
thanks for the chat
twas very informing yet not very technical.. was different.
Oh well, I'm a beginner too. Like everything, it gets easier with practice
tbh i HATE programming. it just turns me off so bad, luckily there are smarter people out there then me lol
Likewise, hope I will see you help people be more objective in future Snow Leopard discussions
.[/QUOTE]oh dont worry i will be. GPU calling is an old technology and apple is well behind the times, its been available for AGES. about time. as always.. apple is behind the times
Yup, It's been available for over 2 years on Windows
.More seriousely whough, I hate it when Apple proclaims things that have been around for a while new and revolutionary. It's cool they're implementing modern stuff into their OS and I give them full credit for it. I just don't like the way they market it.
Yup, It's been available for over 2 years on Windows
More seriousely whough, I hate it when Apple proclaims things that have been around for a while new and revolutionary. It's cool they're implementing modern stuff into their OS and I give them full credit for it. I just don't like the way they market it.
yea i never get that, its a marketing scam basically though. trying to get all the 'suckers' in and make more money. i was one of them towards SL until you helped me realise otherwise lol!
That's the first time in this topic I completely agree with you
hahaha so basically im one of the biggest nubs you've ever met??
lol
I'm hoping the $199 Starter is an MS trial balloon to gauge reactions. I would think $149 for a Home Basic would be more reasonable.
they can afford to make the Starter $99. that's what i would do if i were them
Yup, It's been available for over 2 years on Windows
More seriousely whough, I hate it when Apple proclaims things that have been around for a while new and revolutionary. It's cool they're implementing modern stuff into their OS and I give them full credit for it. I just don't like the way they market it.
OpenCL technology for Windows has been around for 2+ years? Wow, you'd better email every tech editor and gadget blog about that one, especially since it's blatent BS!
OpenCL in Mac should allow for system calls from any application that can benefit which will then be directed to the GPU as a multi-threaded process using the multiple cores on the GPU to do the number crunching.
Now, please demonstrate that technology existing on Windows. Ya, thought so.
He's talking about CUDA, which is the same thing, albeit CUDA is on a slightly different layer
CUDA is not part of Windows, and is not available for general applications to make calls to CUDA to execute code.
OpenCL will be PART OF SL and will be provided by the OS developer.
CUDA also has a lot of limitations. Many. You should read up more about CUDA before declaring it beat SL to the punch on leveraging GPU use in the OS.
That seems reasonable, especially considering that the OS will be supported for the next 10 years, give or take.
you raise a good point. the OS may be supported for a good 10 year, but how many service pack updates is that? microsoft charges you another >$100 for that, so over the 10 years thats a lot of money.
oh. true. that was for an "upgrade service pack" bugger
It's been in nVidia's drivers for a very long time (more than two years, in OS X since 10.5.2) so anyone who updated his box in the last 2 years shouldn't have problems. Which means that although it's not a part of Windows or OS X initially, it IS available for general applications to use on 95% of systems.
From what I've read, CUDA's only problem is the fact it is nVidia's proprietary technology and as such only works on their cards. But that aside, using C for CUDA vs. using OpenCL is very similar. I think I read somewhere that even nVidia says the way things are handled is practically the same so code will be easy to port from one API to the other. Would you mind stating the many limitations you say CUDA has that make it inferior compared to OpenCL?