Instead of loading the registers one lane at a time, load full
registers and then transpose them.
This is faster, reducing the runtime for the function from about
506 cycles to 434 cycles (tested on a Cortex A8).
This also avoids an issue which seems like a cpu bug, present
on Sony Xperia T (cpu implementer 0x51 architecture 7 variant 0x1
part 0x04d). On such a device, it seemed like the "vswp q9, q10"
could start executing before the previous
vld4.u8 {d20[x],d21[x],d22[x],d23[x]}, [r3], r1
had finished and written back their result. Changing the
"vswp q9, q10" into "vswp q10, q9", or into separate
"vswp d18, d20; vswp d19, d21" (or the other way around) seemed to
avoid the issue. This happened occasionally (a couple times per
100000 invocations or so).
This fixes building with gnu binutils, which don't support this
nonstandard form of the instructions.
Once Apple's tools support the proper standard form of the
instructions, the code should be updated to use that everywhere
instead, and these macros should be removed.
This makes sure the windows version of these functions behave
more like the posix version. The posix *snprintf returns how
much would have been written if the buffer had been large
enough, which we don't know easily in the windows versions.
This basically means that we can assume that the return value is
>= 0 now, which can simplify the calling code.
Windows Phone lacks the old CreateThread/beginthreadex APIs for
creating threads. (Technically, the functions still do exist,
but they aren't officially supported and aren't visible in the
headers when targeting Windows Phone.)
Building code that uses the Windows Runtime language extensions
requires building with the -ZW option.
According to the Win64 ABI, these registers need to be preserved,
and compilers are allowed to rely on their content to stay
available - not only for float usage but for any usage, anywhere,
in the calling C++ code.
This adds a macro which pushes the clobbered registers onto the
stack if targeting win64 (and a matching one which restores them).
The parameter to the macro is the number of xmm registers used
(e.g. if using xmm0 - xmm7, the parameter is 8), or in other
words, the number of the highest xmm register used plus one.
This is similar to how the same issue is handled for the NEON
registers q4-q7 with the vpush instruction, except that they needed
to be preserved on all platforms, not only on one particular platform.
This allows removing the XMMREG_PROTECT_* hacks, which can
easily fail if the compiler chooses to use the callee saved
xmm registers in an unexpected spot.
This is what nasm ended up assembling movsx with 32 bit input to
anyway.
Keep using plain movsx for 16 bit input.
This fixes building with yasm in 64 bit mode.
Remap q5 to q8, q6 to q9, q7 to q10 and q8 to q11, and push
q4 to the stack.
This was missed previously since the codec unittest doesn't
test encoding with loop filter enabled yet.
According to the calling convention, the registers q4-q7 should be
preserved by functions. The caller (generated by the compiler) could
be using those registers anywhere for any intermediate data.
Functions that use more than 12 of the qX registers must push
the clobbered registers on the stack in order to be able to restore them
afterwards.
In functions that don't use all 16 registers, but clobber some of
the callee saved registers q4-q7, one or more of them are remapped
to reduce the number of registers that have to be saved/restored.
This incurs a very small (around 0.5%) slowdown in the decoder and
encoder.
According to the calling convention, the registers q4-q7 should be
preserved by functions. The caller (generated by the compiler) could
be using those registers anywhere for any intermediate data.
Functions that use 12 or less of the qX registers can avoid
violating the calling convention by simply using other registers instead
of the callee saved registers q4-q7.
This change only remaps the registers used within functions - therefore
this does not affect performance at all. E.g. in functions using
registers q0-q7, we now use q0-q3 and q8-q11 instead.
Now calling WelsThreadJoin is enough to finish and clean up
the thread on all platforms.
This unifies the thread cleanup code between windows and unix.
Now all of the threading code should use the exact same codepaths
between windows and unix.
Arm assembly has got two variants of the syntax, the old legacy
syntax, and the new modern UAL (unified assembly language) syntax.
Most arm assembly is the same in the both syntaxes, but some
uncommon cases change the order of suffixes - the "subscs"
instruction would be written "subcss" in the old syntax.
The apple tools default to UAL, while the GNU tools (e.g. in
android) require you to specify ".syntax unified" to enable the
new syntax. When enabling the new syntax with the GNU tools, some
cases of "sub r0, r1, lsl #1" needs to be written explicitly as
"sub r0, r0, r1, lsl #1", handled in the previous commit.
This allows using the same, modern syntax for things like subscs,
without needing to have two alternate forms of writing it.
On arm, the exact same detection is done in WelsCPUFeatureDetect,
but in the x86 version of that function we use x86 cpuid for getting
the core count, and this is not available on all processors. For the
case when cpuid can't tell the core count, use the NDK function as
higher level API.
The thread lib itself doesn't build properly on android yet, but will
do so soon.
This allows making the WelsMultipleEventsWaitSingleBlocking
function work properly in unix, without polling. If a master
event is provided, the function first waits for a signal on
that event - once such a signal is received, it is assumed that
one of the individual events in the list have been signalled as
well. Then the function can proceed to check each of the semaphores
in the list using sem_trywait to find the first one of them that
has been signalled. Assuming that the master event is signalled
in pair with the other events, one of the sem_trywait calls
should succeed.
The same master event is also used in
WelsMultipleEventsWaitAllBlocking, to keep the semaphore values
in sync across calls to the both functions.
All users of the function passed the value corresponding to
"infinite", and the (currently unused) unix implementation of it
only supported infinite wait as well.
