Error out early instead of producing bogus mismatch errors in case
of an incorrect cpu mask for example.

Changes stem from redesigning the reduction stage of the multisymbol
decode function.
* No longer use adapt4 for 5 possible symbol values
* Specialize reduction for 4/8/16 decode functions
* Modify control flow

+------------------------+--------------+--------------+---------------+
|                        |  Neoverse V1 |  Neoverse N1 |   Cortex A72  |
|                        | (Graviton 3) | (Graviton 2) |  (Graviton 1) |
+------------------------+-------+------+-------+------+-------+-------+
|                        |  Old  |  New |  Old  |  New |  Old  |  New  |
+------------------------+-------+------+-------+------+-------+-------+
| decode_bool_neon       |  13.0 | 12.9 |  14.9 | 14.0 |  39.3 |  29.0 |
+------------------------+-------+------+-------+------+-------+-------+
| decode_bool_adapt_neon |  15.4 | 15.6 |  17.5 | 16.8 |  41.6 |  33.5 |
+------------------------+-------+------+-------+------+-------+-------+
| decode_bool_equi_neon  |  11.3 | 12.0 |  14.0 | 12.2 |  35.0 |  26.3 |
+------------------------+-------+------+-------+------+-------+-------+
| decode_hi_tok_c        |  73.7 | 57.8 |  73.4 | 60.5 | 130.1 | 103.9 |
+------------------------+-------+------+-------+------+-------+-------+
| decode_hi_tok_neon     |  63.3 | 48.2 |  65.2 | 51.2 | 119.0 | 105.3 |
+------------------------+-------+------+-------+------+-------+-------+
| decode_symbol_\        |  28.6 | 22.5 |  28.4 | 23.5 |  67.8 |  55.1 |
| adapt4_neon            |       |      |       |      |       |       |
+------------------------+-------+------+-------+------+-------+-------+
| decode_symbol_\        |  29.5 | 26.6 |  29.0 | 28.8 |  76.6 |  74.0 |
| adapt8_neon            |       |      |       |      |       |       |
+------------------------+-------+------+-------+------+-------+-------+
| decode_symbol_\        |  31.6 | 31.2 |  33.3 | 33.0 |  77.5 |  68.1 |
| adapt16_neon           |       |      |       |      |       |       |
+------------------------+-------+------+-------+------+-------+-------+

Both POSIX and the C standard places several environmental limits on
setjmp() invocations, with essentially anything beyond comparing the
return value with a constant as a simple branch condition being UB.

We were previously performing a function call using the setjmp()
return value as an argument, which is technically not allowed
even though it happened to work correctly in practice.

Some systems may loosen those restrictions and allow for more
flexible usage, but we shouldn't be relying on that.

Will be used to gate code using vec_absd and other useful instructions.

Kyle Siefring authored 1 year ago and Henrik Gramner committed 1 year ago

Port improvements from the hi token functions to the rest of the symbol
adaption functions. These weren't originally ported since they didn't
work with arbitrary padding. In practice, zero padding is already used
and only the tests need to be updated.

Results - Neoverse N1

Old:
msac_decode_symbol_adapt4_c:         41.4 ( 1.00x)
msac_decode_symbol_adapt4_neon:      31.0 ( 1.34x)
msac_decode_symbol_adapt8_c:         54.5 ( 1.00x)
msac_decode_symbol_adapt8_neon:      32.2 ( 1.69x)
msac_decode_symbol_adapt16_c:        85.6 ( 1.00x)
msac_decode_symbol_adapt16_neon:     37.5 ( 2.28x)

New:
msac_decode_symbol_adapt4_c:         41.5 ( 1.00x)
msac_decode_symbol_adapt4_neon:      27.7 ( 1.50x)
msac_decode_symbol_adapt8_c:         55.7 ( 1.00x)
msac_decode_symbol_adapt8_neon:      30.1 ( 1.85x)
msac_decode_symbol_adapt16_c:        82.4 ( 1.00x)
msac_decode_symbol_adapt16_neon:     35.2 ( 2.34x)

It was originally disabled due to older meson versions mixing the output
of 'meson test -v' from different tests, which made the log difficult to
read. Newer versions however caches the output from each test as it runs
and prints it in one contiguous block, so that's no longer an issue.

On AArch64, the performance counter registers usually are
restricted and not accessible from user space.

On macOS, we currently use mach_absolute_time() as timer on
aarch64. This measures wallclock time but with a very coarse
resolution.

There is a private API, kperf, that one can use for getting
high precision timers though. Unfortunately, it requires running
the checkasm binary as root (e.g. with sudo).

Also, as it is a private, undocumented API, it can potentially
change at any time.

This is handled by adding a new meson build option, for switching
to this timer. If the timer source in checkasm could be changed
at runtime with an option, this wouldn't need to be a build time
option.

This allows getting benchmarks like this:

mc_8tap_regular_w16_hv_8bpc_c:              1522.1 ( 1.00x)
mc_8tap_regular_w16_hv_8bpc_neon:            331.8 ( 4.59x)

Instead of this:

mc_8tap_regular_w16_hv_8bpc_c:                 9.0 ( 1.00x)
mc_8tap...

Prints a list of cpuflags available for the current architecture.

Flags which are supported on the current system will be printed in
green, and flags which are unsupported in red with a ~ prefix.

Arpad Panyik authored 1 year ago and Martin Storsjö committed 1 year ago

Add run-time CPU feature detection for DotProd, i8mm, SVE and SVE2.
SVE and SVE2 are AArch64-only features.

The 6-tap sub-pel filter specialisation uses different code paths for
sharp (8-tap) and regular/smooth (6-tap) filtering kernels.

This patch enables benchmarking for the different code paths.

* Process the entire buffer to get better coverage of eob handling.

* Use a more reasonable buffer size.

* Ignore trailing dif bits to allow for more implementation flexibility.

Henrik Gramner authored 1 year ago and Henrik Gramner committed 1 year ago

Default to using the number of logical cores divided by thread count.

inv_txfm_add_4x4_identity_identity_0_8bpc_c:      534.6 ( 1.00x)
inv_txfm_add_4x4_identity_identity_0_8bpc_rvv:     72.2 ( 7.40x)
inv_txfm_add_4x4_identity_identity_1_8bpc_c:      534.7 ( 1.00x)
inv_txfm_add_4x4_identity_identity_1_8bpc_rvv:     72.3 ( 7.40x)

Only print the paths relative to the argon directory. This avoids
excessive terminal line wrapping due to long path names which
otherwise interferes with the '\r' usage for progress reporting.

Matthias Dressel authored 1 year ago and Ronald S. Bultje committed 1 year ago

Allows to explicitly enable/disable seek-stress tests.

jinbo authored 1 year ago and Hecai Yuan committed 1 year ago

Relative speedup over C code:

msac_decode_bool_c:                            0.5 ( 1.00x)
msac_decode_bool_lsx:                          0.5 ( 1.09x)
msac_decode_bool_adapt_c:                      0.7 ( 1.00x)
msac_decode_bool_adapt_lsx:                    0.6 ( 1.20x)
msac_decode_symbol_adapt4_c:                   1.3 ( 1.00x)
msac_decode_symbol_adapt4_lsx:                 1.0 ( 1.30x)
msac_decode_symbol_adapt8_c:                   2.1 ( 1.00x)
msac_decode_symbol_adapt8_lsx:                 1.0 ( 2.05x)
msac_decode_symbol_adapt16_c:                  3.7 ( 1.00x)
msac_decode_symbol_adapt16_lsx:                0.8 ( 4.77x)

Also prefer re-setting the signal handler upon intercept in combination
with SA_RESETHAND over re-raising exceptions with the SIG_DFL handler.

This was missed in 2ef970a8.

Also print this text for EXCEPTION_IN_PAGE_ERROR on Windows.

Henrik Gramner authored 1 year ago and Henrik Gramner committed 1 year ago

The max_width/max_height values can exceed 16-bit range.

longjmp on Windows uses SEH to unwind on ARM/ARM64 too, just like on
x86_64, thus use RtlCaptureContext/RtlRestoreContext instead of
setjmp/longjmp on those architectures as well.

Henrik Gramner authored 1 year ago and Henrik Gramner committed 1 year ago

Some functionality is only available on WINAPI_PARTITION_DESKTOP systems.

Henrik Gramner authored 1 year ago and Henrik Gramner committed 1 year ago

This allows for the use of standard VT100 escape codes for text coloring,
which simplifies things by eliminating a bunch of Windows-specific code.

This is only supported since Windows 10. Things will still run on
older systems, just without colored text output.

Integrates --bench-c into --bench to simplify benchmarks.

Henrik Gramner authored 1 year ago and Henrik Gramner committed 1 year ago

Pack two indices into each byte instead of storing them separately.

Reduces memory usage by up to 16 kB per sb128 in streams that uses
screen content tools when frame-threading is enabled, at the cost
of some additional computational overhead for packing/unpacking.