This will probably be mostly rambling, apologies, but I hope to have some nice discussion on the future of open technologies now that Intel is finally showing up to the party with Ivybridge and their future products.
Irrelevant fluff, feel free to skip:
I will say now I am not an expert on these things, I am a bog standard end user and not a coder or massive power user. I also plan to buy Ivybridge when it finally comes out for my desktop, and will pair it with my HD7950.
I also plan to buy a tablet before the end of the year, using whatever has the best format and good battery life (expecting it to be Ivybridge again, you might notice a small theme).
I also have an HTPC with Sandybridge based Pentium processor, equipped with an HD6450 to give actual video playback/etc capabilities (thanks Intel for making SB mediocre at graphics).
http://www.xbitlabs.com/news/graphi...ts_Estimate_in_Q2_2011_After_Adjustments.html
It is a well known fact that Intel is #1 at graphics, and nowadays NV is #3 due to AMD's IGP/APUs.
http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/19
With Ivybridge, Intel has OpenCL 1.1 support in hardware, and we are waiting for software support.
With Intel finally showing like they might have some game, the open standards that AMD have (somewhat lazily) been "promoting" finally have more meaning.
We have had various discussions before on vendor specific vs open technologies, mostly focusing on NV (since they have more), and I have maintained that it is short sighted to want vendor specific technologies because of Intel (see: http://forums.anandtech.com/showthread.php?p=31022059&highlight=intel#post31022059, http://forums.anandtech.com/showthread.php?p=29483479&highlight=intel#post29483479).
Here we have an announcement from AMD and Adobe about OpenGL and OpenCL support in Photoshop CS6:
http://blogs.amd.com/fusion/2012/04/24/adobe-and-amd-enable-brilliant-experiences/
Then we have the bad:
http://www.anandtech.com/show/5778/amd-launches-radeon-7700m-7800m-and-7900m-mobile-gpus/2
http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/8
Intel introduced QuickSync, which accelerates transcoding videos. AMD are looking to add something similar when they get some software support, and NV already have such technologies, but the problem is that typically they produce mediocre quality output and can't be customised.
If we have three vendors shipping in 100% of computers with architectures which all support the same open standard(s), such as OpenCL, maybe finally we will see some REAL development of acceleration of basic tasks, and the decline of vendor lock in technologies.
PhysX which ran on OpenCL could be used in any PC on the iGPU (since almost all CPUs going forward will have one), while the dGPU does the graphics, meaning everyone (with a modern PC) can benefit from the performance.
Now that Intel is bringing it, is anyone else super happy that we might see open technologies and faster adoption of OpenCL acceleration for tasks which it could have been used for over the last 5 years, if people had seen enough value?
Some people seem focused on the issue of whether, for instance, vendor specific coding (e.g. CUDA plugins for Photoshop) might be faster than OpenCL implementations.
Does it matter which is faster, if one has vastly greater support? 20% vs 100%?
With IvyBridge, a new age of software development may finally be born, and Haswell and future Intel CPUs could help develop it even further. The technologies that NV has been building into its GPUs for years, and AMD has also started to focus on, could possibly finally become useful to the public at large!
Irrelevant fluff, feel free to skip:
I will say now I am not an expert on these things, I am a bog standard end user and not a coder or massive power user. I also plan to buy Ivybridge when it finally comes out for my desktop, and will pair it with my HD7950.
I also plan to buy a tablet before the end of the year, using whatever has the best format and good battery life (expecting it to be Ivybridge again, you might notice a small theme).
I also have an HTPC with Sandybridge based Pentium processor, equipped with an HD6450 to give actual video playback/etc capabilities (thanks Intel for making SB mediocre at graphics).

http://www.xbitlabs.com/news/graphi...ts_Estimate_in_Q2_2011_After_Adjustments.html
It is a well known fact that Intel is #1 at graphics, and nowadays NV is #3 due to AMD's IGP/APUs.
http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/19
With Ivybridge, Intel has OpenCL 1.1 support in hardware, and we are waiting for software support.
With Intel finally showing like they might have some game, the open standards that AMD have (somewhat lazily) been "promoting" finally have more meaning.
We have had various discussions before on vendor specific vs open technologies, mostly focusing on NV (since they have more), and I have maintained that it is short sighted to want vendor specific technologies because of Intel (see: http://forums.anandtech.com/showthread.php?p=31022059&highlight=intel#post31022059, http://forums.anandtech.com/showthread.php?p=29483479&highlight=intel#post29483479).
Here we have an announcement from AMD and Adobe about OpenGL and OpenCL support in Photoshop CS6:
http://blogs.amd.com/fusion/2012/04/24/adobe-and-amd-enable-brilliant-experiences/
Adobe is launching Adobe Premiere® Pro CS6 which now includes OpenCL accelerated features in the Mercury Playback engine as well as Adobe® Photoshop® CS6 with breakthrough performance enabled by industry-standards with OpenCL and OpenGL acceleration in the new Mercury Graphics Engine.
Then we have the bad:
http://www.anandtech.com/show/5778/amd-launches-radeon-7700m-7800m-and-7900m-mobile-gpus/2
http://www.anandtech.com/show/4083/...-core-i7-2600k-i5-2500k-core-i3-2100-tested/8
Intel introduced QuickSync, which accelerates transcoding videos. AMD are looking to add something similar when they get some software support, and NV already have such technologies, but the problem is that typically they produce mediocre quality output and can't be customised.
If we have three vendors shipping in 100% of computers with architectures which all support the same open standard(s), such as OpenCL, maybe finally we will see some REAL development of acceleration of basic tasks, and the decline of vendor lock in technologies.
PhysX which ran on OpenCL could be used in any PC on the iGPU (since almost all CPUs going forward will have one), while the dGPU does the graphics, meaning everyone (with a modern PC) can benefit from the performance.
Now that Intel is bringing it, is anyone else super happy that we might see open technologies and faster adoption of OpenCL acceleration for tasks which it could have been used for over the last 5 years, if people had seen enough value?
Some people seem focused on the issue of whether, for instance, vendor specific coding (e.g. CUDA plugins for Photoshop) might be faster than OpenCL implementations.
Does it matter which is faster, if one has vastly greater support? 20% vs 100%?
With IvyBridge, a new age of software development may finally be born, and Haswell and future Intel CPUs could help develop it even further. The technologies that NV has been building into its GPUs for years, and AMD has also started to focus on, could possibly finally become useful to the public at large!