1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-11-26 19:01:44 +02:00

build: Prefer NASM assembler over YASM

NASM is more actively maintained and permits generating dependency information
as a sideeffect of assembling, thus cutting build times in half.

(Cherry-picked from libav commit 57b753b445)

Signed-off-by: James Almer <jamrial@gmail.com>
This commit is contained in:
Diego Biurrun 2016-10-13 18:33:15 +00:00 committed by James Almer
parent 3cc73d3d6d
commit 4f9297ac3b
5 changed files with 15 additions and 13 deletions

View File

@ -6,7 +6,7 @@ os:
addons:
apt:
packages:
- yasm
- nasm
- diffutils
compiler:
- clang
@ -17,7 +17,7 @@ cache:
before_install:
- if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew update --all; fi
install:
- if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew install yasm; fi
- if [ "$TRAVIS_OS_NAME" == "osx" ]; then brew install nasm; fi
script:
- mkdir -p ffmpeg-samples
- ./configure --samples=ffmpeg-samples --cc=$CC

View File

@ -22,6 +22,8 @@ version <next>:
- headphone audio filter
- superequalizer audio filter
- roberts video filter
- The x86 assembler default switched from yasm to nasm, pass
--x86asmexe=yasm to configure to restore the old behavior.
version 3.3:
- CrystalHD decoder moved to new decode API

6
configure vendored
View File

@ -3258,7 +3258,7 @@ pkg_config_default=pkg-config
ranlib_default="ranlib"
strip_default="strip"
version_script='--version-script'
x86asmexe_default="yasm"
x86asmexe_default="nasm"
windres_default="windres"
nvcc_default="nvcc"
nvccflags_default="-gencode arch=compute_30,code=sm_30 -O2"
@ -5506,7 +5506,7 @@ EOF
}
if ! disabled_any asm mmx x86asm; then
for program in $x86asmexe yasm nasm; do
for program in $x86asmexe nasm yasm; do
probe_x86asm $program
test -n "$x86asm_type" && break
done
@ -5518,7 +5518,7 @@ EOF
esac
check_x86asm "movbe ecx, [5]" && enable x86asm ||
die "yasm/nasm not found or too old. Use --disable-x86asm for a crippled build."
die "nasm/yasm not found or too old. Use --disable-x86asm for a crippled build."
check_x86asm "vextracti128 xmm0, ymm0, 0" || disable avx2_external
check_x86asm "vpmacsdd xmm0, xmm1, xmm2, xmm3" || disable xop_external
check_x86asm "vfmaddps ymm0, ymm1, ymm2, ymm3" || disable fma4_external

View File

@ -161,8 +161,8 @@ do{
For x86, mark registers that are clobbered in your asm. This means both
general x86 registers (e.g. eax) as well as XMM registers. This last one is
particularly important on Win64, where xmm6-15 are callee-save, and not
restoring their contents leads to undefined results. In external asm (e.g.
yasm), you do this by using:
restoring their contents leads to undefined results. In external asm,
you do this by using:
cglobal function_name, num_args, num_regs, num_xmm_regs
In inline asm, you specify clobbered registers at the end of your asm:
__asm__(".." ::: "%eax").
@ -199,12 +199,12 @@ actual lines causing issues.
Inline asm vs. external asm
---------------------------
Both inline asm (__asm__("..") in a .c file, handled by a compiler such as gcc)
and external asm (.s or .asm files, handled by an assembler such as yasm/nasm)
and external asm (.s or .asm files, handled by an assembler such as nasm/yasm)
are accepted in FFmpeg. Which one to use differs per specific case.
- if your code is intended to be inlined in a C function, inline asm is always
better, because external asm cannot be inlined
- if your code calls external functions, yasm is always better
- if your code calls external functions, external asm is always better
- if your code takes huge and complex structs as function arguments (e.g.
MpegEncContext; note that this is not ideal and is discouraged if there
are alternatives), then inline asm is always better, because predicting

View File

@ -71,9 +71,9 @@ Mac OS X on PowerPC or ARM (iPhone) requires a preprocessor from
assembly functions. Put the Perl script somewhere
in your PATH, FFmpeg's configure will pick it up automatically.
Mac OS X on amd64 and x86 requires @command{yasm} to build most of the
Mac OS X on amd64 and x86 requires @command{nasm} to build most of the
optimized assembly functions. @uref{http://www.finkproject.org/, Fink},
@uref{http://www.gentoo.org/proj/en/gentoo-alt/prefix/bootstrap-macos.xml, Gentoo Prefix},
@uref{https://wiki.gentoo.org/wiki/Project:Prefix, Gentoo Prefix},
@uref{https://mxcl.github.com/homebrew/, Homebrew}
or @uref{http://www.macports.org, MacPorts} can easily provide it.
@ -141,7 +141,7 @@ them under @command{MinGW-w64 Win64 Shell} and @command{MinGW-w64 Win32 Shell}.
pacman -S make pkgconf diffutils
# mingw-w64 packages and toolchains
pacman -S mingw-w64-x86_64-yasm mingw-w64-x86_64-gcc mingw-w64-x86_64-SDL
pacman -S mingw-w64-x86_64-nasm mingw-w64-x86_64-gcc mingw-w64-x86_64-SDL2
@end example
To target 32 bits replace @code{x86_64} with @code{i686} in the command above.
@ -159,7 +159,7 @@ You will need the following prerequisites:
@item @uref{http://code.google.com/p/msinttypes/, msinttypes}
(if using MSVC 2012 or earlier)
@item @uref{http://msys2.github.io/, MSYS2}
@item @uref{http://yasm.tortall.net/, YASM}
@item @uref{http://www.nasm.us/, NASM}
(Also available via MSYS2's package manager.)
@end itemize