e5b0a782fadfadb353260bea7b98ed3e42bd7ab6 |
|
05-Oct-2016 |
Martin Storsjo <martin@martin.st> |
armv8: Use neg instead of sub #0, neg as replacement for rsb When the arm code has been ported to aarch64, occurrances of the rsb instruction have been replaced according to the following pattern: arm: rsb rB, rA, #X aarch64: sub xTMP, xA, #X neg xB, xTMP When the immediate constant #X is zero, we can just as well use a plain neg instruction without any extra subtraction. Change-Id: I637be20b469d8d4e7fac712e8039c7e3eedb8c54
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|
6958350a651e38f4f4062d9bbf17e01a584670b2 |
|
05-Oct-2016 |
Martin Storsjo <martin@martin.st> |
armv8: Remove unnecessary sign extensions After loading a single byte with ldrb and immediately moving it into a vector register with dup .8b, there's no point in sign extending it to 64 bit inbetween. Similarly in the second case, when adding a number of pixels, there's no need to sign extend between additions, just do the additions using the 32 bit register names. Change-Id: Ie37f47b93138643ffd05169a3792bc39967dbea3
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|
d91f49ad65795b8d3223f1aba481bf3931b291e6 |
|
05-Sep-2016 |
Martin Storsjo <martin@martin.st> |
armv8: Add missing sign extension for WORD32 parameters For WORD32 parameters, the upper half of the parameter registers is undefined (even though it is zero in most cases). When easily doable, use the 'w' register names for such parameters instead of doing the sign extension with a separate instruction. This fixes crashes in some configurations, when built with clang, which can pass nonzero bits in the high half of registers more aggressively than GCC. Also fix the parameter/register mappings; use wN register names where applicable, and fix incorrect parameter placement. (Some functions still were documented to have only 4 parameters in registers.) Change-Id: Icedf6503b064149d1a651ff1c6a76de5c01722ea
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|
db02f57bdcb3b101ca658e6c6dd1f5437539b06a |
|
09-Jun-2015 |
Martin Storsjo <martin@martin.st> |
armv8: Remove redundant NEON element size declarations When specifying one specific lane of the vector, the number of lanes don't need to be specified. The clang built-in assembler doesn't allow the redundant declarations, while binutils gas work fine with both forms. Change-Id: I86077ce0774d4594a1295b6860e4944df87dde2f
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|
cd9e51fc2b6091aa344363ee8a4f881e938ff80c |
|
18-Apr-2015 |
Lajos Molnar <lajos@google.com> |
fix file permissions Change-Id: Ia4f99d5b963acd8d8a1afc2fbdf06b122d898f63
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|
8d3d303c7942ced6a987a52db8977d768dc3605f |
|
13-Mar-2015 |
Hamsalekha S <hamsalekha.s@ittiam.com> |
Initial version Change-Id: I7efe9a589cd24edf86e8d086b40c27cbbf8b4017
/external/libavc/common/armv8/ih264_intra_pred_luma_16x16_av8.s
|