-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 08/28/2011 06:51 PM, Bryan Cain wrote: > On 08/28/2011 07:38 PM, Eric Anholt wrote: >> On Sat, 27 Aug 2011 20:18:55 -0700, Kenneth Graunke <kenn...@whitecape.org> >> wrote: >>> From: Bryan Cain <bryanca...@gmail.com> >>> >>> Using multiply and reciprocal for integer division involves potentially >>> lossy floating point conversions. This is okay for older GPUs that >>> represent integers as floating point, but undesirable for GPUs with >>> native integer division instructions. >>> >>> TGSI, for example, has UDIV/IDIV instructions for integer division, >>> so it makes sense to handle this directly. Likewise for i965. >>> >>> Signed-off-by: Bryan Cain <bryanca...@gmail.com> >>> Signed-off-by: Kenneth Graunke <kenn...@whitecape.org> >> >>> --- >>> case ir_binop_div: >>> - if (lowering(DIV_TO_MUL_RCP)) >>> + if (lowering(INT_DIV_TO_MUL_RCP) && >>> ir->operands[1]->type->is_integer()) >>> + int_div_to_mul_rcp(ir); >>> + else if (lowering(DIV_TO_MUL_RCP)) >>> div_to_mul_rcp(ir); >>> break; >>> >> Sure looks odd to me for one of these to be checking the type and ther >> other not. > > It works, though. If it's not an integer type, it's going to be a float > type.
Right, but it could be an integer type and !lower(INT_DIV_TO_MUL_RCP). In which case if lower(DIV_TO_MUL_RCP), do_to_mul_rcp will get called on an integer type. Right? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/ iEYEARECAAYFAk5b0X4ACgkQX1gOwKyEAw+IEQCfV1mHz5ZrwLMvPpbAzAGdlorg zNcAn2iMVg8U7J5BVGiRugikqKYI5D7C =myKt -----END PGP SIGNATURE----- _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
