Version 0.1.0 (2008-10-15):
	* Initial public release

Version 0.1.1 (2009-07-06):
	* Added intel_fn11 fields to cpu_raw_data_t to handle
	  new processor topology enumeration required on Core i7
	* Support for Intel Nehalem architecture CPUs (Core i7, Xeon i7)

Version 0.1.2 (2009-09-26):
	* Added support for MSR reading through self-extracting kernel driver
	  on Win32.

Version 0.1.3 (2010-04-20):
	* Added support for greater more accurate CPU clock measurements with
	  cpu_clock_by_ic()

Version 0.2.0 (2011-10-11):
	* Support for AMD Bulldozer CPUs, 128-bit SSE unit size checking.
	  A backwards-incompatible change, since the sizeof cpu_id_t is now
	  different.

Version 0.2.1 (2012-05-26):
	* Support for Ivy Bridge, and detecting the presence of the RdRand
	  instruction.

Version 0.2.2 (2015-11-04):
	* Support for newer processors up to Haswell and Vishera
	* Fix clock detection in cpu_clock_by_ic() for Bulldozer
	* Support for detection of AVX/AVX2/BMI1/BMI2
	* More entries supported in cpu_msrinfo()
	* Rename of some CPU codenames, made more consistent
	* *BSD and Solaris support (unofficial)

Version 0.3.0 (2016-07-09):
	* A backwards-incompatible change, since the sizeof cpu_raw_data_t and
	  cpu_id_t are now different.
	* Support for processors up to Skylake.
	* Fix clock detection in cpu_clock_by_ic() for Skylake.
	* Support up to 8 subleaf entries for CPUID leaf 04 and detection
	  of L4 cache.
	* MSR functions supported on FreeBSD.
	* INFO_VOLTAGE request supported by cpu_msrinfo().

Version 0.4.0 (2016-09-30):
	* A backwards-incompatible change, since the sizeof cpu_raw_data_t and
	  cpu_id_t are now different.
	* Better detection of AMD clock multiplier with msrinfo.
	* Support for Intel SGX detection

Version 0.4.1 (2019-02-05):
	* Better support for Skylake Core i5 (#76)
	* Misdiagnosis microarchitecture for i3-3220T (#81)
	* Ability to dump MSR values to a file (PR #82)
	* AMD Ryzen support (#86)
	* Support for Coffee and Kaby Lake (#104)
	* Support for Raven Ridge and Threadripper (#106)
	* Support for Pinnacle Ridge (#111)
	* Fix P-III Celeron misdetection
	* Support for Skylake-X (#116)
	* Support for Zen+ Threadripper

Version 0.5.0 (2020-05-23):
	* A backwards-incompatible change, since the sizeof cpu_raw_data_t and
	  cpu_id_t are now different.
	* Support for Hygon Dhyana (#123)
	* Support for Zen 2 (#125, #137)
	* Stubs for AArch64 (#127)
	* Support for Intel AVX512VNNI detection (#129)
	* Typo fixes (#130)
	* Support for get_total_cpus on Haiku (#131)
	* Fix build on PPC (#132)
	* Support for CMake build system (#138) and vcpkg packaging (#140)
	* Support for Cannon Lake
	* Support for Coffee Lake-U and Coffee Lake Refresh
	* Support for Comet Lake
	* Support for Ice Lake
	* Support for Renoir
	* Fix ThreadSanitizer report (#142)
	* Better support of L1 intruction cache (--l1i-assoc, --l1i-cacheline)
	* Support for Ivy Bridge-E (Xeon)
	* Support up to 4 subleaf entries for CPUID leaf 0x8000001D and
	  fix detection of L3 cache on Zen 2
	* Support for RDSEED/ADX/SHA_NI detection (AMD)
	* Support for ABM/AVX512VBMI/AVX512VBMI2 detection (Intel)