Vittorio Giovara authored 8 years ago and Anton Mitrofanov committed 7 years ago

Add 'i_bitdepth' to x264_param_t with the corresponding '--output-depth' CLI
option to set the bit depth at runtime.

Drop the 'x264_bit_depth' global variable. Rather than hardcoding it to an
incorrect value, it's preferable to induce a linking failure. If applications
relies on this symbol this will make it more obvious where the problem is.

Add Makefile rules that compiles modules with different bit depths. Assembly
on x86 is prefixed with the 'private_prefix' define, while all other archs
modify their function prefix internally.

Templatize the main C library, x86/x86_64 assembly, ARM assembly, AARCH64
assembly, PowerPC assembly, and MIPS assembly.

The depth and cache CLI filters heavily depend on bit depth size, so they
need to be duplicated for each value. This means having to rename these
filters, and adjust the callers to use the right version.

Unfortunately the threaded input CLI module inherits a common.h dependency
(input/frame -> common/threadpool -> common/frame -> common/common) which
is extremely complicated to address in a sensible way. Instead duplicate
the module and select the appropriate one at run time.

Each bitdepth needs different checkasm compilation rules, so split the main
checkasm target into two executables.

Vittorio Giovara authored 8 years ago and Anton Mitrofanov committed 7 years ago

qp is modified to require a valid value before use, while qp_max is set
to maximum allowable value (and clipped later on).

This is needed so that param functions do not depend on bit depth size.

Henrik Gramner authored 7 years ago and Anton Mitrofanov committed 7 years ago

Also drop the x264 prefix from all static cabac arrays.

Henrik Gramner authored 7 years ago and Anton Mitrofanov committed 7 years ago

Use dword instead of qword entries. Cuts the size of the tables in half
which allows each table fit inside a single cache line.

When PIC is disabled dwords are enough to store absolute addresses.

When PIC is enabled we can store dword offsets relative to the start of
the table and simply add the address of the table to the offset in order
to calculate the full address. This approach also have the advantage of
eliminating a whole bunch of run-time .data relocations.

Henrik Gramner authored 7 years ago and Anton Mitrofanov committed 7 years ago

On ELF platforms such symbols needs to be flagged as functions with the
correct visibility to please certain linkers in some scenarios.

Henrik Gramner authored 7 years ago and Anton Mitrofanov committed 7 years ago

The standard section for read-only data on Windows is .rdata. Nasm will
flag non-standard sections as executable by default which isn't ideal.

Henrik Gramner authored 7 years ago and Anton Mitrofanov committed 7 years ago

There are 32 pseudo-instructions for each floating-point comparison
instruction, but only 8 of them are actually valid in legacy-encoded mode.
The remaining 24 requires the use of VEX-encoded (v-prefixed) instructions
and can therefore be disregarded for this purpose.

Fix MSVS fprofiled build for win64

Some cpuflags would previously be displayed incorrectly when running older
operating systems without AVX support on modern CPU:s.

Uses gathers and scatters in combination with conflict detections to
vectorize the scalar part.

Also improve the checkasm test to try different mb_y values and check
for out-of-bounds writes.

James Darnley authored 7 years ago and Henrik Gramner committed 7 years ago

Upstreaming this from FFmpeg. Unused in x264.

Martin Storsjö authored 7 years ago and Anton Mitrofanov committed 7 years ago

The existing functions could easily be used by just calling them
twice - this would give the following cycle numbers from checkasm:

var2_8x8_c:      4110
var2_8x8_neon:   1505
var2_8x16_c:     8019
var2_8x16_neon:  2545

However, by merging both passes into the same function, we get the
following speedup:
var2_8x8_neon:   1205
var2_8x16_neon:  2327

Martin Storsjö authored 7 years ago and Anton Mitrofanov committed 7 years ago

The existing functions could easily be used by just calling them
twice - this would give the following cycle numbers from checkasm:

             Cortex A7     A8     A9   A53
var2_8x8_c:       7302   5342   5050  4400
var2_8x8_neon:    2645   1612   1932  1715
var2_8x16_c:     14300  10528  10020  8637
var2_8x16_neon:   5127   2695   3217  2651

However, by merging both passes into the same function, we get the
following speedup:
var2_8x8_neon:    2312   1190   1389  1300
var2_8x16_neon:   4862   2130   2293  2422

Henrik Gramner authored 8 years ago and Anton Mitrofanov committed 7 years ago

These levels were added in the 2016-10 revision of the H.264 specification and
improves support for content with high resolutions and/or high frame rates.

Level 6.2 supports 8K resolution at 120 fps.

Also shrink the x264_levels array by using smaller data types.

Henrik Gramner authored 8 years ago and Anton Mitrofanov committed 7 years ago

Makes it possible to use slicing with resolutions larger than 2^24 pixels.

Henrik Gramner authored 8 years ago and Anton Mitrofanov committed 7 years ago

Use a dynamic size depending on the MV range. Reduces memory consumption by
up to a few megabytes.

Drop a related old miscompilation check since it may otherwise cause an
out-of-bounds memory access.

Also remove an unused extern variable declaration.

Use the correct ctxIdxInc calculation for coded_block_flag.

Change V and H intra prediction in lossless (TransformBypassModeFlag == 1)
macroblocks to correctly adhere to the specification. Affects lossless
encoding with 8x8dct or mix of lossless with normal macroblocks.

8x8dct has already been disabled in lossless mode for some time due to
being out-of-spec but this will allow us to re-enable it again.

Could occur on the 1st pass in combination with --fake-interlaced and
some input heights due to allocating a too small buffer.

Functions that uses self-relative expressions in the form of [foo-$$]
appears to cause issues on 64-bit Mach-O systems when assembled with nasm.
Temporarily disable those functions on macho64 for the time being until
we've figured out the root cause.

Anton Mitrofanov authored 7 years ago and Henrik Gramner committed 7 years ago

Only check if option -Werror=unknown-warning-option is supported before adding it

Martin Storsjö authored 7 years ago and Henrik Gramner committed 7 years ago

Prior to this, this loop hasn't run at all. The condition has been
the same since it was introduced in 5b0cb86f.

This issue was pointed out by a clang warning.