v0.12.22

LoopVectorization v0.12.22

Diff since v0.12.21

Merged pull requests:

• Rename YisyAIFramework.jl to DianoiaML.jl (#267) (@SkyWorld117)

v0.12.21

LoopVectorization v0.12.21

Diff since v0.12.20

v0.12.20

LoopVectorization v0.12.20

Diff since v0.12.19

Closed issues:

• at-avx: Wrong result when using indices and views (#263)
• Not correctly handling a sincos kernel (#265)

Merged pull requests:

• Optimize solve_unroll_lagrange (#264) (@YingboMa)

v0.12.19

LoopVectorization v0.12.19

Diff since v0.12.18

Closed issues:

• Don't handle constant indices by taking views (#148)
• @avx occasional slowdown compared to @inbounds @simd (#214)

Merged pull requests:

• Pass arguments to avx! in registers. (#261) (@chriselrod)

v0.12.18

LoopVectorization v0.12.18

Diff since v0.12.17

Closed issues:

• @avxt triggers DivideError and SegFault (#255)
• SegFault triggered by unknown reason (#260)

v0.12.17

LoopVectorization v0.12.17

Diff since v0.12.16

v0.12.16

LoopVectorization v0.12.16

Diff since v0.12.15

Closed issues:

• Wrong result with mix of multiplication and addition within avx accumulation (#259)

v0.12.15

LoopVectorization v0.12.15

Diff since v0.12.14

v0.12.14

LoopVectorization v0.12.14

Diff since v0.12.13

Closed issues:

• BoundsError with v0.12.13 (#256)
• MethodError with v0.12.13 (#257)
• Precompilation fails on 32bit systems (#258)

v0.12.13

LoopVectorization v0.12.13

Diff since v0.12.12

Closed issues:

• Windows failure on 1.6 (#227)
• Segfault with image filter example (#235)
• A bug? (#237)
• segfault with static sizes and related issues (#238)
• avx gives incorrect results comparing to the code without it (#239)
• Scope problem with log10 (#249)
• Hangs with div in indices? (#250)
• Bug in scalar reduction? (#251)
• Failure with reduction to scalar over three arrays with triple-repeated index (#252)
• Float32 constructor with Int64 argument? (#253)

Merged pull requests:

• Update matrix_multiplication.md (#236) (@el-oso)
• add another assumption of at-avx to the README.md (#240) (@ranocha)
• add Trixi to users of LoopVectorization (#247) (@ranocha)
• update docstrings to not point to Threads.at-threads anymore (#248) (@ranocha)
• Indexing changes and more precompiles (#254) (@chriselrod)