decompress_generic: optimize match copy
authorDave Watson <davejwatson@fb.com>
Thu, 24 Jan 2019 22:17:24 +0000 (14:17 -0800)
committerDave Watson <davejwatson@fb.com>
Fri, 8 Feb 2019 21:57:23 +0000 (13:57 -0800)
commit5dfa7d422ba6c184a7c7694f56bcd36e38e5ed1a
treee4f93476a19502e37d5f32e07cd5f812869715a6
parent28356e02ad6f6dac529302cedf707712c5b628fe
decompress_generic: optimize match copy

Add an LZ4_wildCopy16, that will wildcopy, potentially smashing up
to 16 bytes, and use it for match copy.  On x64, this avoids many
blocked loads due to store forwarding, similar to issue #411.
lib/lz4.c