Making it easier to work with shaders https://git.yummers.dev/slang.git/atom?h=master 2025-01-22T17:40:15+00:00 2025-01-22T17:40:15+00:00 Yong He yonghe@outlook.com 2025-01-22T17:40:15+00:00 urn:sha1:8000e0ede34e920cc7f37d69a335d74a472eff42 * Cache and reuse glsl module. * Fix. * Implement record/replay for the new api. * Fix record replay. * Fix test. 2024-10-29T06:49:26+00:00 Ellie Hermaszewska ellieh@nvidia.com 2024-10-29T06:49:26+00:00 urn:sha1:f65d756bff8d4c5cbc15bd0322a2ae8e6b896a21 * format * Minor test fixes * enable checking cpp format in ci 2024-08-27T16:53:16+00:00 kaizhangNV 149626564+kaizhangNV@users.noreply.github.com 2024-08-27T16:53:16+00:00 urn:sha1:59c23b96b3649e77f5428c32e47b4401a802c604 * Migrate cpu-hello-world to new slang API Migrate cpu-hello-world to new slang API, and also convert this example as one of the unit test. * Add 'shader-object' to slang-unit-test * Convert ray-tracing example into unit-test Convert ray-tracing example into unit-test * Fix some replay bugs: - Wrong decode type in 'getEntryPointHostCallable'. - Mistakes in computing the output buffer size. - Wrong decode type in array size in specialize() call. - When capture entrypoint, we should increase the reference count for the allocated entrypoint recorder object, because that is allocated by record layer, it should be owned by the layer, user should not be able to free it. - Improve json consumer on the prelude text. * Test verify change: In our test, we add a "callIdx" string at beginning of the hash-code string, as there could be more than one modules in the example, so they could call 'getEntryPointHash' multiple times, in order for the test can identify them, add "callIdx: <number>" as the key word. 2024-07-23T15:45:26+00:00 kaizhangNV 149626564+kaizhangNV@users.noreply.github.com 2024-07-23T15:45:26+00:00 urn:sha1:986256ffb92ab7c8fc7cf9f2c424919a439a824f * Add decoder * Add a replay executable to consume the decoded content Add file-processor.cpp/h where we implement the logic to process the captured file block by block. Each block is: function header + parameter buffer + function tailer + function output[optional]. After reading one block, the block of data is sent to decoder module to dispatch the corresponding API. Add slang-decoder.cpp/h where we implement the logic to dispatch the slang API according to the input block data. - Rename api_callId.h to capture-format.h - Renmae capture_utility.cpp to capture-utility.cpp - Renmae capture_utility.h to capture-utility.h - Change the #include file name accordingly. * Reorganize source files structure Move all the capture logic code into `capture` directory. - the capture code will be build with slang dll. Move all the replay logic code into `relay` directoy. - the replay code is not part of slang dll, it will be built as a stand alone binary and link against slang dll. Change the #include file names accordingly. Add tools/slang-replay/main.cpp for the slang-replay stand alone binary place holder. Will implement it later. Update premake5.lua accordingly. * Update cmake files Update cmake files to change the build process for capture and relay system. - capture component should be build with slang dll, so we should not include replay component. - replay component should be a separate executable tool, which should not include capture component. - In order to easy use our current cmake infrastructure, move the shared files to a `util` folder - change the header include path * Redesgin the interfaces of consumers Fix some issues in capture Finish implementing all slang-decoder functions * Fix the AppleClang build issue * Address few comments - Fix the weird indent issues. - Correct the function name for CreateGlobalSession() - Rename file-processor to captureFile-processor to be more specific. - Use Slang::List instead of std::vector * record/replay: name refactor change Refactor the naming. Change the name "encoder/capture" to "record".