v0.8.2

StochasticRounding v0.8.2

Diff since v0.8.1

Merged pull requests:

• BFloat16s v0.5 compat (#76) (@milankl)

v0.8.1

StochasticRounding v0.8.1

Diff since v0.8.0

Merged pull requests:

• round :Down removed (#75) (@milankl)

v0.8.0

StochasticRounding v0.8.0

Diff since v0.7.0

Merged pull requests:

• Complete restructure (#73) (@milankl)

Closed issues:

• Float16_chance_roundup direct from a Float64? (#63)
• Restructure (#70)
• trunc for stochastic rounding (#72)

v0.7.0

StochasticRounding v0.7.0

Diff since v0.6.3

Closed issues:

• Using SR.jl types in conjunction with DE.jl (#57)
• Float64sr via Double64? (#59)
• GenericFFT.jl compatibility (#64)

Merged pull requests:

• add Float32sr methods for DE.jl (#58) (@tomkimpson)
• Float64sr via Double64 (#61) (@milankl)
• Added missing function definitions resolves #64 (#65) (@avleenk2312)
• BFloat16s = "0.1,0.2,0.3,0.4" (#69) (@milankl)
• rand, randn, zeros, ones constructors (#71) (@milankl)

v0.6.3

StochasticRounding v0.6.3

Diff since v0.6.0

Closed issues:

• register v0.6 (#45)
• [BUG] Invalid stochastic rounding with infinite (#48)
• Algorithm shortcut (#49)
• Complex numbers (#60)
• Float16sr subnormals also rounds somewhere else then up/down (#62)

Merged pull requests:

• github action ci (#47) (@milankl)
• Shortcuts (#50) (@milankl)
• convert tabs to spaces (#52) (@milankl)
• Indent (#53) (@milankl)
• bump to v0.6.2 (#54) (@milankl)
• new tagbot yml (#55) (@milankl)
• Add support for converting between Stochastic FP numbers, and convenience methods for array conversions (#56) (@jprhyne)
• Float16 stochastic round of subnormals (#67) (@milankl)
• bump to v0.6.3 (#68) (@milankl)

v0.6.0

StochasticRounding v0.6.0

Diff since v0.5.0

Closed issues:

• register v0.5 (#40)
• odd-float32 x can be round to prevfloat(x) (#43)

Merged pull requests:

• Fastfloatfaster (#41) (@milankl)
• Fastfloatfaster (#42) (@milankl)
• Proper rounding for all and for v0.6 (#44) (@milankl)

v0.5.0

StochasticRounding v0.5.0

Diff since v0.4.1

Merged pull requests:

• fastfloat16sr first implementation (#36) (@milankl)
• Fastfloat16 (#37) (@milankl)
• FastFloat16sr with && return 0 (#38) (@milankl)
• no unnecessary type casting (#39) (@milankl)

v0.4.1

StochasticRounding v0.4.1

Diff since v0.4.0

Merged pull requests:

• Removed conversion to BFloat16sr (#35) (@milankl)

v0.4.0

StochasticRounding v0.4.0

Diff since v0.3.2

Merged pull requests:

• Remove _basetable (#29) (@milankl)
• Mk/max (#30) (@milankl)
• Stochastic round to NaN? (#31) (@milankl)
• bump to v0.4 (#33) (@milankl)

v0.3.2

StochasticRounding v0.3.2

Diff since v0.3.1

Closed issues:

• register v0.3.1 (#22)
• register v0.3.2 (#26)

Merged pull requests:

• Mk/subnormals (#23) (@milankl)
• Stochastic rounding for subformals of Float16/BFloat16 (#24) (@milankl)
• bump version to v0.3.2 (#25) (@milankl)
• BFloat16 compat limited (#27) (@milankl)