6877 Commits (15716c1d38136a9a4e382a536e174f956dac39cb)

Author SHA1 Message Date
James Jackson-South 15716c1d38 Fix Color conversion for differing types 4 years ago
Dmitry Pentin 7d524bda2b Fixed compilation error 4 years ago
Dmitry Pentin e9a25c74e0 Added xml docs to the SpectralConverter 4 years ago
Dmitry Pentin 13602169c0 Moved cancellation token to a more appropriate place 4 years ago
Dmitry Pentin 717b166e41 Removed unnecessary small allocation from color post processing, removed test benchmark call from sandbox project 4 years ago
Dmitry Pentin eb120ba1a5 gfoidl const Size struct jit inlining in color converters 4 years ago
James Jackson-South cc28bc00df Use Rgb24 for Async 4 years ago
Brian Popow 07cef0c5d1 Add SSE2 version of TransformOne 4 years ago
Brian Popow 23ee80be31 Add SSE2 version of TransformTwo 4 years ago
Brian Popow 8e515c1423 Rename ITransform to ITransformTwo to better reflect that actually two transforms take place 4 years ago
Brian Popow b52abeffc0 Remove not needed allocation 4 years ago
Brian Popow 383f455d3a Add AVX2 version of Vp8_Sse4X4 4 years ago
James Jackson-South 81433c2f52 Remove more scalar bounds checks 4 years ago
Justin Hopper c0ee67b5b2 Added missing CancellationToken parameters to Image 4 years ago
Brian Popow 554c0d7b52 Add AVX2 version of QuantizeBlock 4 years ago
Brian Popow 83da0e0694 Reverse array access order to avoid bounds checks 4 years ago
Dmitry Pentin e8b59c6bb4 Fixed access modifier & docs 4 years ago
Dmitry Pentin 97a0dccfc9 Removed redundant if check (min call) 4 years ago
Dmitry Pentin 0d3e7fff53 Removed redundant if check, removed used code 4 years ago
Dmitry Pentin 503b379e89 Removed JpegBlockPostProcessor abstraction layer 4 years ago
Dmitry Pentin 140b07854e Removed redundant code from Block8x8F and Block8x8 4 years ago
Brian Popow 0215e99696 Avoid pinning, avoid using LoadScalarVector128 4 years ago
Brian Popow cb08407728 Use nint in for loop 4 years ago
Brian Popow 81070c4e61 Add missing #pragma warning restore SA1503 4 years ago
Brian Popow 798e9c3ad6 Add SSE2 version of FTransform2 4 years ago
Brian Popow 38fd3a8458 Avoid bounds checks in IsFlat 4 years ago
Brian Popow 4bb56eea71 Define mask and shuffle vectors as static readonly 4 years ago
Dmitry Pentin 9c3d348b4d Separated Block8x8 into scalar and intrinsic code bases 4 years ago
Dmitry Pentin 525cfc89ae Removed redundant/unused code from Block8x8 4 years ago
Dmitry Pentin 60fd3b0659 Removed GenericBlock8x8<T> 4 years ago
Brian Popow 98f5a42816 Add SSE2 version off FTransform 4 years ago
Dmitry Pentin 7a2c22a966 Removed obsolete/redundant/unused code from Block8x8F 4 years ago
Brian Popow 7775c34304 Group loading y, u, v together 4 years ago
Brian Popow 2d60b73b14 Rename size to count, add DebugGuard 4 years ago
Brian Popow c174ab42be Remove SSE2 version of VectorMismatch: Profiling does not show any speedup 4 years ago
Brian Popow 22537b226b Revert "Allocate clean buffers": the tmp buffers does not need to be clean, they will be overwritten anyway 4 years ago
Brian Popow cded607d5c Allocate clean buffers 4 years ago
Brian Popow 2ca81aec3c Fill buffers with default values only in Debug mode 4 years ago
Brian Popow 6293f72c80 Use ref parameters in ConvertYuv444ToBgrSse41 4 years ago
Brian Popow 65870b96f4 Avoid branching inside loop 4 years ago
Brian Popow cc5f7af71c Better version of LoadHigh 4 years ago
Brian Popow 5403fbd8b2 Add better version of ReduceSum for Vector 256 4 years ago
Brian Popow 110ff3d9e8
Avoid using Span<int> tmp 4 years ago
Dmitry Pentin 586df2e7b1 Fixed crash bug 4 years ago
Brian Popow 491b742ae4 Add SSE2 version of VectorMismatch 4 years ago
Dmitry Pentin 28147059b4 Removed bound checks from DCT adjustment methods 4 years ago
Brian Popow 6393484e42 Remove duplicate FTransform method 4 years ago
Brian Popow a45f49517b Avoid bounds checks in VectorMismatch 4 years ago
Brian Popow 8192e116f2 Add AVX2 version of AddVector 4 years ago
Brian Popow 14508498d9 Seal classes and avoid calling SuppressFinalize 4 years ago