<html> <head> <base href="https://bugs.freedesktop.org/" /> </head> <body> <div> <a class="bz_bug_link bz_status_NEW " title="NEW - Mesa 10.6.3 linker is slow" href="https://bugs.freedesktop.org/show_bug.cgi?id=91857#c15">Comment # 15</a> on <a class="bz_bug_link bz_status_NEW " title="NEW - Mesa 10.6.3 linker is slow" href="https://bugs.freedesktop.org/show_bug.cgi?id=91857">bug 91857</a> from <a class="email" href="mailto:CliveMcCarthy@gmail.com" title="Clive McCarthy <CliveMcCarthy@gmail.com>"> Clive McCarthy</a> <pre>Clearly, whatever my *weird* fragment shader is doing to the linker, shouldn't impede the excellent optimizations that are being made to the actual shader executable. My own work-around on an i3 is to show a fixed image that allows a projector to be focused for the ~90 seconds while the shader links. Factors of two, in the linking of the shader, are not an issue. There is a two or three orders of magnitude class of problem somewhere. An i7 far outclasses the Atom processor but the Nvidia linker gets the job done in 120ms and the i7 30 to 40 seconds. Profiling the linker, while it's working on the shader, should show up where all the cycles are being consumed. There are an awful lot of i7 cycles in 30 seconds. I built Ian Romanick's standalone Mesa compiler and use it as an adjunct to my system just to make sure my shaders are compliant and portable. I'm tempted to try and build the Mesa linker too but I don't know that I have the skills to do the job.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are the QA Contact for the bug.</li> </ul> </body> </html>