Topic: Catalyst 12.4 + OpenCL 1.2 (Read 6494 times)

CPU usage is minimal with 4 GPUs using intensity 9, when intensity is raised to 10 however, CPU usage jumps to almost 100%. 5 GPUs doesn't cause a BSOD, but the screen cuts out without causing a reboot, the system just hangs... Hashrate is unchanged as well.. at least it didn't go down at all, if anything miners seem to be hashing like .5 MH/s faster (for 7970s mind you). Stability seems maybe a bit better as well, where my system almost certainly would have rebooted with 12.3, instead caused a driver crash, which reset my GPUs to stock clocks.

Please let us know how much CPU usage you have.

Catalyst 12.4 beta drivers available here:
http://www.tomshardware.com/forum/347950-33-catalyst-beta

OpenCL 1.2 beta:
http://blogs.amd.com/developer/2012/03/23/welcome-to-opencl%E2%84%A2-1-2-check-out-these-beta-drivers-that-contain-a-complete-opencl%E2%84%A2-1-2-solution/

From page:
Last December we released preview drivers that contained much of new the functionality defined in the OpenCL 1.2™ specification.  AMD is an ardent promoter of OpenCL™, and today, I am pleased to make available beta drivers that contain a complete  – beta level – implementation of the OpenCL™ 1.2 specification.  You can download the drivers from our APP SDK downloads page.

The OpenCL  1.2 beta includes  the following OpenCL™ 1.2 functionality:

    Host access flags for memory objects enable more efficient buffer handling and provide added protection. For example, a buffer that is created as “write only” cannot be read from the host.
    Pattern based GPU buffer and image initialization can help eliminate need for certain buffer/image transfers
    Memory objects migration supports transfer of buffers prior to need
    New generalized image creation API
    Enhanced image/buffer map operations
    OpenCL 1.2 CPU device partition including partition of a CPU after addition to a context
    Generalized 1D and 2D images, image arrays,  and image<-> buffer interop
    Libraries support including the separation of compile and link phases and the ability to compile with external symbols
    Kernel reflection, the ability to query a kernel’s arguments
    Support for printf as a built in function


Not a developer, but could this be the release we've been waiting for? Planning on trying it on my 7970 rig tonight, see if 5+ GPUs work again at least..