Thursday, August 07, 2008
Microsoft Accelerator versus Apple OpenCL
"OpenCL is going to give each of us the power of a true supercomputer on our desktops."
OpenCL is a language for GPGPU based on C99, that Apple created with cooperation from others. C99 is a revision of the C standard (ISO/IEC 9899:1999 in 1999, hence the name) which has several amendments itself. There is another revision started 2007 - "C1x". Computing is kind of like this. OpenCL's purpose - to use the power of GPU for stuff beyond graphics, as we know.
OpenCL will no doubt show up at SIGGRAPH in LA 11-15th August
Patricia Harrell, director of stream computing at AMD, agrees: "I think OpenCL makes it possible to move GPU computing from being a speciality technology to something that any developer can count on being in a system that he is writing for."
Acceptance of a Khronos-backed GPU-computing standard is not universal. "The Khronos group is the everyone-apart-from-Microsoft group," says Andrew Richards, chief executive of Codeplay, a parallel-software specialist based in Edinburgh.
For the past few years, a team at Microsoft Research has been working on an experimental system called Accelerator. "There are several other initiatives to make it easier to program GPUs," says Satnam Singh, who is investigating acceleration technologies at Microsoft Research in Cambridge.
"I don't think Microsoft will be sitting and watching. I would never underestimate Microsoft's ability to come up with alternative positions."
OpenCL is a language for GPGPU based on C99, that Apple created with cooperation from others. C99 is a revision of the C standard (ISO/IEC 9899:1999 in 1999, hence the name) which has several amendments itself. There is another revision started 2007 - "C1x". Computing is kind of like this. OpenCL's purpose - to use the power of GPU for stuff beyond graphics, as we know.
OpenCL will no doubt show up at SIGGRAPH in LA 11-15th August
Patricia Harrell, director of stream computing at AMD, agrees: "I think OpenCL makes it possible to move GPU computing from being a speciality technology to something that any developer can count on being in a system that he is writing for."
Acceptance of a Khronos-backed GPU-computing standard is not universal. "The Khronos group is the everyone-apart-from-Microsoft group," says Andrew Richards, chief executive of Codeplay, a parallel-software specialist based in Edinburgh.
For the past few years, a team at Microsoft Research has been working on an experimental system called Accelerator. "There are several other initiatives to make it easier to program GPUs," says Satnam Singh, who is investigating acceleration technologies at Microsoft Research in Cambridge.
"I don't think Microsoft will be sitting and watching. I would never underestimate Microsoft's ability to come up with alternative positions."