File likwid.changes of Package likwid
-------------------------------------------------------------------
Mon Dec 4 13:31:09 UTC 2023 - Christian Goll <cgoll@suse.com>
- enable likwid-setFrequencies, likwid-setFreq and likwid-accessD
for non x86_64 and disabled i568 builds
-------------------------------------------------------------------
Fri Nov 17 00:30:24 UTC 2023 - Christian Goll <cgoll@suse.com>
- Updated to 5.3.3 release. As this is a major step, just the
highlights are mentioned:
* Measure multiple event sets in a round*robin fashion (no multiplexing!)
* Event options to filter the counter increments
* Whole LIKWID functionality is exposed as API for C/C++ and Lua
* New functions in the Marker API to switch event sets and get intermediate
results
* Topology code relies on hwloc. CPUID is still included but only as fallback
* Most LIKWID applications are written in Lua (only exception likwid*bench)
* Monitoring daemon likwid*agent with multiple output backends
* Basic support for Linux perf_event interface instead of native access.
Currently only core*local counters working, Uncore is experimental
* Support to build against a existing Lua installation (5.1 * 5.3 tested)
* Support for CPU frequency manipulation, Lua interface updated
* Access module checks for LLNL's msr_safe kernel module
* Support for counter registers that are only available when
HyperThreading is off
* Fix for non*HyperThreading counters (PMC4*7) on Intel Broadwell
* Socket measurements can be used for all cores on the socket in
metric formulas.
* Support for Nvidia GPU monitoring (with NvMarkerAPI)
* New clock frequency backend (with less overhead)
* Generation of benchmarks for likwid*bench on*the*fly from ptt files
* Switch back to C*based metric calculator (less overhead)
* Interface function to performance groups, create your own.
* Integration of GOTCHA for hooking into client application at runtime
* Thread*local initialization of streams for likwid*bench
* Enhanced support for SLURM with likwid*mpirun
* New MPI and Hybrid pinning features for likwid*mpirun
* Interface to enable the membind kernel memory policy
* JSON output filter file (use *o output.json)
- removed following patches:
* likwid-3.1.3-loadData-for-i386.patch
* likwid-32bit.patch
* likwid-fix-i586.patch
* likwid-libdir.patch
* likwid-optflags.patch
* likwid-perlapp.patch
-------------------------------------------------------------------
Fri Feb 6 15:14:17 UTC 2015 - mardnh@gmx.de
- updated source url
- updated patch likwid-libdir.patch
- updated patch likwid-optflags.patch
- added likwid-perlapp.patch for setting the correct prefix
- fixed license
- update to version 3.1.3:
+ Release 3.1.3
- Full Uncore Support for Nehalem EX and Westmere EX
- Atom Silvermont (Avoton + BayTrail) support
- Read Marker API results and derived metrics in instrumented application
(Patch by Julian Kunkel)
- More low-level benchmarks for likwid-bench
- Kernel module for enabling the RDPMC instruction
- Use RDPMC for fixed and general purpose core-local counters if possible
- New undocumented but working events
- Support for all RAPL domains in likwid-powermeter
- Better PCI device lookup
- Lots of bug fixes
+ Release 3.1.2
- New application likwid-setFrequencies and corresponding daemon likwid-setFreq
- likwid-powermeter can measure multiple sockets at once
- likwid-memsweeper also cleans last level cache from dirty cachelines
- Sanitized groups for Intel IvyBridge and SandyBridge
- Automatic lookup of Uncore PCI devices with exclusion of non-existent devices
- Reduced connect time to likwid-accessD
- Multiple workgroups for likwid-bench are evaluated correctly
- MSR device file handling on Intel Xeon Phi improved
- Improved commandline argument handling for all LIKWID tools
- man pages sanitized and a new one for likwid-bench
- Moved setuid command to sbin directory
- Lots of bugfixes
+ Release 3.1
- Full support for Intel Xeon Phi
- Support for Intel IvyBridge-EP (partial Uncore support)
- Support for Intel Haswell
- Support for AMD Kabini architecture
- Improved Intel Atom support
- Adding expression based thread group syntax for all LIKWID tools
- Add scatter option to thread groups
- Add possibility to dump topology info to cfg file. This enables simpler usage
of LIKWID in a restricted cpuset
- Using the Marker API a variable number of threads may enter a region now
- Simple lock file based control to synchronize Monitoring and Profiling in a
cluster environment
- Using custom event sets on Intel Processors fixed counters are automatically
added and connected derived metrics runtime, clock and CPI are output
- Oversubscription of processors with round robin placement is now default
- Add event for TM/TM2 temperature sensors on Intel processors
- Improved RDTSC timing measurement
- Review of performance groups
- Performance improvements in Marker API
- Lots of bugfixes and improvements
+ Release 3.0
- New application likwid-memsweeper to cleanup ccNUMA memory domains.
This functionality is also integrated in likwid-pin.
- Support for Intel SandyBridge Uncore (partial)
- Support for Intel IvyBridge (only core part)
- Initial support for Intel Xeon Phi (KNC)
- Better support for AMD Interlagos
- The OpenMP type is now detected automatically while pinning
- Lots of bugfixes and improvements
- Marker API works now for threaded code and accessDaemon
- Uncore support for timeline mode
- Convenient macro wrapper for Marker API
- Data volume as new metric in all memory/cache groups
- Updated Wiki documentation
-------------------------------------------------------------------
Mon Aug 20 23:47:47 CET 2012 - pascal.bleser@opensuse.org
- initial version (2.3.0)
