GNU C Library 2.26 Released With Per-Thread Cache For Greater Performance
Today marks the release of GNU C Library version 2.26 and it's a big release with one feature we have been looking forward to.
The big feature we've been looking forward to in glibc 2.26 is its malloc per-thread cache now being merged. This per-thread cache requires no locks and has been shown to boost performance for code paths with frequent allocating and freeing of small amounts of memory. This feature is enabled by default in glibc 2.26 but can be disabled at build-time.
There are also other glibc 2.26 features including Unicode 10.0 support, glibc tunables enabled by default, wrappers for new Linux system calls, 128-bit floating point support in the math library, the DNS stub resolver is in better shape, various bug fixes, and a hand full of CVE security fixes.
More details on the GNU C Library 2.26 release can be found via this morning's info-gnu announcement.
The big feature we've been looking forward to in glibc 2.26 is its malloc per-thread cache now being merged. This per-thread cache requires no locks and has been shown to boost performance for code paths with frequent allocating and freeing of small amounts of memory. This feature is enabled by default in glibc 2.26 but can be disabled at build-time.
There are also other glibc 2.26 features including Unicode 10.0 support, glibc tunables enabled by default, wrappers for new Linux system calls, 128-bit floating point support in the math library, the DNS stub resolver is in better shape, various bug fixes, and a hand full of CVE security fixes.
More details on the GNU C Library 2.26 release can be found via this morning's info-gnu announcement.
8 Comments