flac dev - May 2015 - A patch for precompute_partition_info_sums_()

Here I attach a preliminary version of a patch for
precompute_partition_info_sums_()
function that should accelerate encoding of 24-bit input data.

1) SSE2, SSSE3 and AVX2 versions of this function should be updated, too

2) the patch also changes
     if(FLAC__bitmath_ilog2(default_partition_samples) + bps +
FLAC__MAX_EXTRA_RESIDUAL_BPS < 32)
into (in effect)
     if(FLAC__bitmath_ilog2(default_partition_samples) + bps +
FLAC__MAX_EXTRA_RESIDUAL_BPS <= 32)

So, should the value of FLAC__MAX_EXTRA_RESIDUAL_BPS be increased (to 5? or to
6?)

3) any comments?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: precomp_partitions.patch
Type: application/octet-stream
Size: 3069 bytes
Desc: not available
Url :
http://lists.xiph.org/pipermail/flac-dev/attachments/20150523/a55b2278/attachment.obj

lvqcl wrote:
> 3) any comments?
Well, it compiles and passes the test suite. However, I am concerned
that that the test suite may not actually fully cover this function
(maybe its time to figure out gcov again). I know we had a problem 
in this bit of code before.

Is it possible to write a unit test for this functon?

Erik
-- 
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/

Erik de Castro Lopo wrote:
> Is it possible to write a unit test for this functon?
I'm not a programmer, so I don't know how to make a proper unit test
here.
And I don't know what to test. Because...
> I know we had a problem in this bit of code before.
...the problem was not in the code itself, but in "an assumption that
the average residual magnitude will not be more than "bps" bits".