Re: [patch, fortran-dev] Use fixed variable sizes for stride calculations

2012-04-13 Thread Thomas Koenig
Hi Janne, I thought I already approved this a few weeks ago sorry, I totally missed that. Comes from having a computer crash on you... Thanks for the review! Thomas

Re: [patch, fortran-dev] Use fixed variable sizes for stride calculations

2012-04-13 Thread Janne Blomqvist
On Wed, Apr 11, 2012 at 10:51, Thomas Koenig wrote: > Hi, > > >> this patch uses division by known sizes (which can usually be replaced >> by a simple shift because intrinsics have sizes of power of two) instead >> of division by the size extracted from the array descriptor itself. >> >> This shou

Re: [patch, fortran-dev] Use fixed variable sizes for stride calculations

2012-04-11 Thread Thomas Koenig
Hi, this patch uses division by known sizes (which can usually be replaced by a simple shift because intrinsics have sizes of power of two) instead of division by the size extracted from the array descriptor itself. This should save about 20 cycles for a single calculation. I'll go through the

Re: [patch, fortran-dev] Use fixed variable sizes for stride calculations

2012-03-28 Thread Janne Blomqvist
On Sat, Mar 24, 2012 at 16:13, Thomas Koenig wrote: > Hello world, > > this patch uses division by known sizes (which can usually be replaced > by a simple shift because intrinsics have sizes of power of two) instead > of division by the size extracted from the array descriptor itself. > > This sh

[patch, fortran-dev] Use fixed variable sizes for stride calculations

2012-03-24 Thread Thomas Koenig
Hello world, this patch uses division by known sizes (which can usually be replaced by a simple shift because intrinsics have sizes of power of two) instead of division by the size extracted from the array descriptor itself. This should save about 20 cycles for a single calculation. I'll go thr