2024-01-03 11:52:02 +01:00
|
|
|
-------------------------------------------------------------------
|
|
|
|
|
Wed Jan 3 01:32:19 UTC 2024 - Stefan Brüns <stefan.bruens@rwth-aachen.de>
|
|
|
|
|
|
|
|
|
|
- Update to 1.6.0:
|
|
|
|
|
* Changes
|
|
|
|
|
+ Replaced the rANS codec SIMD gathers with simulated gathers
|
|
|
|
|
via scalar memory fetches. This helps AMD Zen4, but
|
|
|
|
|
importantly it also fixes a disastrous performance regression
|
|
|
|
|
caused by Intel's DownFall microcode fix.
|
|
|
|
|
+ There is an impact on pre-DownFall speeds, but we should
|
|
|
|
|
focus on patched CPUs as a priority.
|
|
|
|
|
+ A small speed up to the rans_F_to_s3 function used by
|
|
|
|
|
order-0 rans decode.
|
|
|
|
|
+ Small speed up to SIMD rans32x16 order-1 encoder by reducing
|
|
|
|
|
cache misses. Also sped up the rans4x8 order-1 encoder,
|
|
|
|
|
particularly on AMD Zen4.
|
|
|
|
|
+ Now supports building with "zig cc"
|
|
|
|
|
* Bug fixes
|
|
|
|
|
+ Improve robustness of name tokeniser when given non 7-bit
|
|
|
|
|
ASCII and on machines where "char" defaults to unsigned.
|
|
|
|
|
+ Also fixed a 1 byte buffer read-overrun in name tokeniser.
|
|
|
|
|
+ Fix name tokeniser encoder failure with some duplicated
|
|
|
|
|
streams.
|
|
|
|
|
+ Fixed rans_set_cpu to work multiple times, as well as
|
|
|
|
|
reinstating the ability to change decode and encode side
|
|
|
|
|
independently (accidentally lost in commit 958032c). No
|
|
|
|
|
effect on usage, but it improves the test coverage.
|
|
|
|
|
+ Added a round-trip fuzz tester to test the ability to
|
|
|
|
|
encode. The old fuzz testing was decode streams only.
|
|
|
|
|
+ Fixed bounds checking in rans_uncompress_O0_32x16_avx2,
|
|
|
|
|
fixing buffer read overruns.
|
|
|
|
|
+ Removed undefined behaviour in transpose_and_copy(),
|
|
|
|
|
fixing zig cc builds.
|
|
|
|
|
|
2023-05-12 11:50:10 +02:00
|
|
|
-------------------------------------------------------------------
|
|
|
|
|
Thu May 4 14:47:16 UTC 2023 - Andrea Manzini <andrea.manzini@suse.com>
|
|
|
|
|
|
|
|
|
|
- Update to 1.5.0:
|
|
|
|
|
* Significant speed ups to the fqzcomp codec via code restructuring
|
|
|
|
|
and use of memory prefetch instructions. Encode is 30-40% faster
|
|
|
|
|
and decode 5-8% faster.
|
|
|
|
|
* Remove unused ax_with_libdeflate.m4 file from build system
|
|
|
|
|
|
|
|
|
|
- removed patch fix_ix86_build.patch already merged in upstream
|
|
|
|
|
|
|
|
|
|
- Update to 1.4.0:
|
|
|
|
|
* This is almost entirely minor bug fixing with a few small updates.
|
|
|
|
|
* Optimise compression / speed of the name tokeniser
|
|
|
|
|
* Improvements for Intel -m32 builds, including better AVX2 validation
|
|
|
|
|
* Detect Neon capability at runtime via operating system APIs.
|
|
|
|
|
* Update hts_pack to operate in line with CRAMcodecs spec, where the number of symbols > 16.
|
|
|
|
|
* Fixed too-stringent buffer overflow checking in O1 rans decoder.
|
|
|
|
|
|
|
|
|
|
|
2022-09-11 08:19:50 +02:00
|
|
|
-------------------------------------------------------------------
|
|
|
|
|
Thu Sep 8 21:52:47 UTC 2022 - Stefan Brüns <stefan.bruens@rwth-aachen.de>
|
|
|
|
|
|
|
|
|
|
- Update to 1.3.0:
|
|
|
|
|
* The primary change in this release is a new SIMD enabled
|
|
|
|
|
rANS codec.
|
|
|
|
|
There is a 32-way unrolled rANS implementation. This is accessed
|
|
|
|
|
using the existing rans 4x16 API with the RANS_ORDER_X32 bit set.
|
|
|
|
|
* Improved memory allocation via a new htscodecs_tls_alloc function.
|
|
|
|
|
* Some external functions have been renamed, with the old ones
|
|
|
|
|
still existing in a deprecated fashion.
|
|
|
|
|
* Improved test framework with an "entropy" tool that iterates
|
|
|
|
|
over all entropy encoders.
|
|
|
|
|
* Reworked fuzzing infrastructure.
|
|
|
|
|
* Small speed improvements to various rANS encoders and decoders.
|
|
|
|
|
* Substantial memory reduction to the name tokeniser (tok3).
|
|
|
|
|
* Fixed undefined behaviour in our use of _builtin_clz().
|
|
|
|
|
* Fixed a few redundant #includes.
|
|
|
|
|
* Work around strict aliasing bugs, uncovered with gcc -O2.
|
|
|
|
|
* Fixed an issue with encoding data blocks close to 2GB in size.
|
|
|
|
|
* Fix encode error with large blocks using RANS_ORDER_STRIPE.
|
|
|
|
|
- Add fix_ix86_build.patch
|
|
|
|
|
|
2022-05-02 11:30:00 +02:00
|
|
|
-------------------------------------------------------------------
|
|
|
|
|
Wed Apr 20 19:45:48 UTC 2022 - Ferdinand Thiessen <rpm@fthiessen.de>
|
|
|
|
|
|
|
|
|
|
- Created initial package for libhts
|
|
|
|
|
|
