Make frame-wide filter-type decision fully RD-based.
Implements several heuristics to prune mode search
Merge "Refactor SSE2 8x8 functional units"
Merge "Fix to comp_inter_joint_search_thresh feature."
Fix to comp_inter_joint_search_thresh feature.
Added two new skip experiments.
Merge "Adding write_skip_coeff function."
Merge "Enable early termination in rd search"
Merge "Replacing 64 / MI_SIZE with MI_BLOCK_SIZE."
Adding write_skip_coeff function.
Merge "Inline a few intra predictors"
Enable early termination in rd search
Replacing 64 / MI_SIZE with MI_BLOCK_SIZE.
Merge "Adding write_selected_txfm_size function."
Refactor SSE2 8x8 functional units
Merge "Use pmovmskb to skip quantize loops over empty coefficients."
Merge "Remove unused function vp9_build_inter4x4_predictors_mbuv()."
Merge "Adjust Speed 0 settings."
Merge "Removing redundant struct from union b_mode_info."
Merge "Added a speed feature use_square_partition_only"
Removing redundant struct from union b_mode_info.
Adding write_selected_txfm_size function.
Added a speed feature use_square_partition_only
Use pmovmskb to skip quantize loops over empty coefficients.
Remove unused function vp9_build_inter4x4_predictors_mbuv().
Revert "New motion threshold factor - speed feature."
new unit test for cpu-speed
Speed feature to binary search dir intramodes
Merge "Clean-up in forward update to use mapping tables"
Tx size selection enhancements
Clean-up in forward update to use mapping tables
Merge "Removing unused implicit segmentation code."
Merge "Make get_coef_context() branchless."
Merge "Removing vp9_mbpitch.c, moving vp9_setup_block_dptrs to vp9_block.h."
Merge "Additional vp9_decodemv.c cleanup."
Removing unused implicit segmentation code.
Merge "Add speed feature to disable splitmv"
Add speed feature to disable splitmv
Calculate rd cost per transformed block
Merge "Update quantize SSSE3 SIMD to cover 32x32 transform case also."
fix the mismatch again in cpu_used 2
use partitioning from last frame
Removing vp9_mbpitch.c, moving vp9_setup_block_dptrs to vp9_block.h.
Make get_coef_context() branchless.
Additional vp9_decodemv.c cleanup.
Merge "Quantize (64-bit only, for now) SSSE3 SIMD."
Merge "Removing vp9_modecont.{h, c}."
Merge "Moving encoder subexp encoding functions to subexp.{h, c}."
Merge "Adding vp9_rb_read_signed_literal function."
Merge "Inlining decode_atom, decode_sb_intra, and decode_sb."
Merge "Cleanup inside vp9_decodemv.c."
Update quantize SSSE3 SIMD to cover 32x32 transform case also.
Quantize (64-bit only, for now) SSSE3 SIMD.
Removing vp9_modecont.{h, c}.
Merge "New motion threshold factor - speed feature."
New motion threshold factor - speed feature.
Adding vp9_rb_read_signed_literal function.
Merge "Enable SSE2 4x4 ADST/DCT transform"
SSE2 version of vp9_short_fdct32x32_rd.
Moving encoder subexp encoding functions to subexp.{h, c}.
Merge "fixed a bug where sse is not populated"
Merge "add Neon optimized add constant residual functions"
Merge "fix test compile error"
Merge "Inline vp9_get_coef_context() (and remove vp9_ prefix)."
Merge "Minor change to prevent one level of dereference in cost_coeffs()."
add Neon optimized add constant residual functions
Merge "Cosmetic reordering of FRAME_CONTEXT members."
Inlining decode_atom, decode_sb_intra, and decode_sb.
Enable SSE2 4x4 ADST/DCT transform
fixed a bug where sse is not populated
Merge "Fix switch statement in 8x8 transform"
Cosmetic reordering of FRAME_CONTEXT members.
Cleanup inside vp9_decodemv.c.
Merge "Removing CONFIG_DEBUG checks on assertions."
Fix switch statement in 8x8 transform
Merge "Some minor optimizations for cost_coeffs()."
Merge "Make coefficient skip condition an explicit RD choice."
Inline vp9_get_coef_context() (and remove vp9_ prefix).
Merge "Decoder's code cleanup."
Removing CONFIG_DEBUG checks on assertions.
Minor change to prevent one level of dereference in cost_coeffs().
Some minor optimizations for cost_coeffs().
Make coefficient skip condition an explicit RD choice.
Merge "Minor cleanups"
Merge "Optimize partition search order"
Optimize partition search order
Merge "Fix tile independence with both column tiling and static_thresh set."
Merge "variance_test: add missing ClearSystemState..."
Fix tile independence with both column tiling and static_thresh set.
Add Neon optimized loop filter functions.
Merge "General cleanup in segmentation-related code."
Merge "Moving subexp encoding functions in separate vp9_dsubexp.c file."
Inline quantize so idiv instruction gets removed from inner loop.
Merge "Auto adapt step size feature."