diff --git a/CMakePresets.json b/CMakePresets.json index 748637f7b38f..beb38ec85f56 100644 --- a/CMakePresets.json +++ b/CMakePresets.json @@ -13,10 +13,7 @@ "generator": "Ninja", "binaryDir": "${sourceDir}/build", "cacheVariables": { - "CMAKE_BUILD_TYPE": "Debug", - "CMAKE_CXX_FLAGS_RELEASE": "-O2", - "CMAKE_CXX_FLAGS_RELWITHDEBINFO": "-O2 -g", - "CMAKE_CXX_FLAGS_MINSIZEREL": "-Os" + "CMAKE_BUILD_TYPE": "Debug" } }, { @@ -41,13 +38,11 @@ }, { "name": "msvc-release", + "inherits": "msvc-debug", "displayName": "MSVC (Release)", - "description": "Debug build using Ninja generator and MSVC with vcpkg dependencies.", - "generator": "Ninja", - "binaryDir": "${sourceDir}/build", + "description": "Release build using Ninja generator and MSVC with vcpkg dependencies.", "cacheVariables": { - "CMAKE_BUILD_TYPE": "Release", - "CMAKE_TOOLCHAIN_FILE": "$env{VCPKG_ROOT}/scripts/buildsystems/vcpkg.cmake" + "CMAKE_BUILD_TYPE": "Release" } }, { diff --git a/README_cmake.md b/README_cmake.md index 362d4594b577..a59ab16f68b5 100644 --- a/README_cmake.md +++ b/README_cmake.md @@ -377,7 +377,6 @@ targets (like tests and documentation) are built. | Option | Default | Description | | ---------------------- | -------------------- | ---------------------------------------------------------------------------------------- | | `WITH_TESTS` | `ON` | Enable building unit and integration tests | -| `WITH_APPS` | `ON` | Enable testing sample applications (run `ctest -L apps` to actually build and test them) | | `WITH_PYTHON_BINDINGS` | `ON` if Python found | Enable building Python 3.x bindings | | `WITH_DOCS` | `OFF` | Enable building the documentation via Doxygen | | `WITH_UTILS` | `ON` | Enable building various utilities including the trace visualizer | diff --git a/src/CodeGen_PyTorch.cpp b/src/CodeGen_PyTorch.cpp index b52dc3ab0572..b17cc80e7084 100644 --- a/src/CodeGen_PyTorch.cpp +++ b/src/CodeGen_PyTorch.cpp @@ -172,7 +172,7 @@ void CodeGen_PyTorch::compile(const LoweredFunc &f, bool is_cuda) { << c_print_name(buffer_args[i].name) << "_buffer.host_dirty()," << "\"device not synchronized for buffer " << c_print_name(buffer_args[i].name) - << ", make sure all update stages are excplicitly computed on GPU." + << ", make sure all update stages are explicitly computed on GPU." << "\");\n"; stream << get_indent(); stream @@ -386,7 +386,7 @@ inline int test1_th_(at::Tensor &_buf, float _alpha, int32_t _beta) { AT_ASSERTM(err == 0, "Halide call failed"); // Make sure data is on device - AT_ASSERTM(!_buf_buffer.host_dirty(),"device not synchronized for buffer _buf, make sure all update stages are excplicitly computed on GPU."); + AT_ASSERTM(!_buf_buffer.host_dirty(),"device not synchronized for buffer _buf, make sure all update stages are explicitly computed on GPU."); _buf_buffer.device_detach_native(); return 0; diff --git a/src/EliminateBoolVectors.h b/src/EliminateBoolVectors.h index bdb3882909ed..77a7331ccff5 100644 --- a/src/EliminateBoolVectors.h +++ b/src/EliminateBoolVectors.h @@ -16,7 +16,7 @@ namespace Internal { * u16x8, u16x8), where the first argument is a vector of integers representing * a mask. This pass converts vectors of bools to vectors of integers to meet * this requirement. This is done by injecting intrinsics to convert bools to - * architecture-specific masks, and using a select_mask instrinsic instead of a + * architecture-specific masks, and using a select_mask intrinsic instead of a * Select node. This also converts any intrinsics that operate on vectorized * conditions to a *_mask equivalent (if_then_else, require). Because the masks * are architecture specific, they may not be stored or loaded. On Stores, the diff --git a/src/Schedule.h b/src/Schedule.h index 0532b6ce1038..f0c5d0797be2 100644 --- a/src/Schedule.h +++ b/src/Schedule.h @@ -575,7 +575,7 @@ class FuncSchedule { // @} /** You may explicitly specify an estimate of some of the function - * dimensions. See \ref Func::estimate */ + * dimensions. See \ref Func::set_estimate */ // @{ const std::vector &estimates() const; std::vector &estimates(); diff --git a/src/autoschedulers/adams2019/autotune_loop.sh b/src/autoschedulers/adams2019/autotune_loop.sh index 7f99692402fc..3de57185aeca 100755 --- a/src/autoschedulers/adams2019/autotune_loop.sh +++ b/src/autoschedulers/adams2019/autotune_loop.sh @@ -1,3 +1,5 @@ +#!/bin/bash + # Build the generator to autotune. This script will be autotuning the # autoscheduler's cost model training pipeline, which is large enough # to be interesting. diff --git a/src/autoschedulers/mullapudi2016/AutoSchedule.cpp b/src/autoschedulers/mullapudi2016/AutoSchedule.cpp index 0ea9bacb8a72..d8227393e286 100644 --- a/src/autoschedulers/mullapudi2016/AutoSchedule.cpp +++ b/src/autoschedulers/mullapudi2016/AutoSchedule.cpp @@ -1957,7 +1957,7 @@ Partitioner::GroupAnalysis Partitioner::analyze_group(const Group &g, bool show_ float load_slope = arch_params.balance / arch_params.last_level_cache_size; for (const auto &f_load : group_load_costs) { internal_assert(g.inlined.find(f_load.first) == g.inlined.end()) - << "Intermediates of inlined pure fuction \"" << f_load.first + << "Intermediates of inlined pure function \"" << f_load.first << "\" should not have been in the group_load_costs\n"; const auto &alloc_reg = get_element(alloc_regions, f_load.first); diff --git a/src/runtime/HalideBuffer.h b/src/runtime/HalideBuffer.h index 6f8265478ab1..6aa89846c080 100644 --- a/src/runtime/HalideBuffer.h +++ b/src/runtime/HalideBuffer.h @@ -836,7 +836,7 @@ class Buffer { // @{ // The overload with one argument is 'explicit', so that - // (say) int is not implicitly convertable to Buffer + // (say) int is not implicitly convertible to Buffer explicit Buffer(int first) { static_assert(!T_is_void, "To construct an Buffer, pass a halide_type_t as the first argument to the constructor"); diff --git a/src/runtime/HalideRuntime.h b/src/runtime/HalideRuntime.h index b457e02d67ff..355469ee5acb 100644 --- a/src/runtime/HalideRuntime.h +++ b/src/runtime/HalideRuntime.h @@ -187,7 +187,7 @@ typedef bool (*halide_semaphore_try_acquire_t)(struct halide_semaphore_t *, int) /** A task representing a serial for loop evaluated over some range. * Note that task_parent is a pass through argument that should be - * passed to any dependent taks that are invokved using halide_do_parallel_tasks + * passed to any dependent taks that are invoked using halide_do_parallel_tasks * underneath this call. */ typedef int (*halide_loop_task_t)(void *user_context, int min, int extent, uint8_t *closure, void *task_parent); @@ -929,7 +929,7 @@ extern void halide_memoization_cache_evict(void *user_context, uint64_t eviction * the case where halide_memoization_cache_lookup is handling multiple * buffers. (This corresponds to memoizing a Tuple in Halide.) Note * that the host pointer must be sufficient to get to all information - * the relase operation needs. The default Halide cache impleemntation + * the release operation needs. The default Halide cache impleemntation * accomplishes this by storing extra data before the start of the user * modifiable host storage. * diff --git a/src/runtime/HalideRuntimeD3D12Compute.h b/src/runtime/HalideRuntimeD3D12Compute.h index 5814f85a8de2..96634330be15 100644 --- a/src/runtime/HalideRuntimeD3D12Compute.h +++ b/src/runtime/HalideRuntimeD3D12Compute.h @@ -78,7 +78,7 @@ struct halide_d3d12compute_command_queue; extern int halide_d3d12compute_acquire_context(void *user_context, struct halide_d3d12compute_device **device_ret, struct halide_d3d12compute_command_queue **queue_ret, bool create); -/** This call balances each successfull halide_d3d12compute_acquire_context call. +/** This call balances each successful halide_d3d12compute_acquire_context call. * If halide_d3d12compute_acquire_context is replaced, this routine must be replaced * as well. */ diff --git a/src/runtime/HalideRuntimeHexagonDma.h b/src/runtime/HalideRuntimeHexagonDma.h index ccabac597628..42b1ea35dc31 100644 --- a/src/runtime/HalideRuntimeHexagonDma.h +++ b/src/runtime/HalideRuntimeHexagonDma.h @@ -50,7 +50,7 @@ typedef enum { extern const struct halide_device_interface_t *halide_hexagon_dma_device_interface(); /** This API is used to set up the DMA device interface to be used for DMA transfer. This also internally - * creates the DMA device handle and populates all the Buffer releated parameters (width, height, stride) + * creates the DMA device handle and populates all the Buffer related parameters (width, height, stride) * to be used for DMA configuration. */ extern int halide_hexagon_dma_device_wrap_native(void *user_context, struct halide_buffer_t *buf, @@ -90,7 +90,7 @@ extern int halide_hexagon_dma_prepare_for_copy_to_device(void *user_context, str extern int halide_hexagon_dma_unprepare(void *user_context, struct halide_buffer_t *buf); /** This API is used to setup the hexagon Operation modes. We will setup the necessary Operating frequency - * based on the power mode choosen. Check the structure halide_hexagon_power_mode_t defined in Halide HalideRuntimeHexagonHost.h + * based on the power mode chosen. Check the structure halide_hexagon_power_mode_t defined in Halide HalideRuntimeHexagonHost.h * for the supported power modes. */ extern int halide_hexagon_dma_power_mode_voting(void *user_context, halide_hexagon_power_mode_t cornercase); diff --git a/src/runtime/HalideRuntimeHexagonHost.h b/src/runtime/HalideRuntimeHexagonHost.h index c5756504ac4b..ff0092ffd111 100644 --- a/src/runtime/HalideRuntimeHexagonHost.h +++ b/src/runtime/HalideRuntimeHexagonHost.h @@ -80,7 +80,7 @@ typedef enum halide_hexagon_power_mode_t { } halide_hexagon_power_mode_t; /** More detailed power settings to control Hexagon. - * @param set_mips - Set to TRUE to requst MIPS + * @param set_mips - Set to TRUE to request MIPS * @param mipsPerThread - mips requested per thread, to establish a minimal clock frequency per HW thread * @param mipsTotal - Total mips requested, to establish total number of MIPS required across all HW threads * @param set_bus_bw - Set to TRUE to request bus_bw diff --git a/src/runtime/HalideRuntimeMetal.h b/src/runtime/HalideRuntimeMetal.h index 802d6659317f..fef18c895101 100644 --- a/src/runtime/HalideRuntimeMetal.h +++ b/src/runtime/HalideRuntimeMetal.h @@ -87,7 +87,7 @@ struct halide_metal_command_queue; extern int halide_metal_acquire_context(void *user_context, struct halide_metal_device **device_ret, struct halide_metal_command_queue **queue_ret, bool create); -/** This call balances each successfull halide_metal_acquire_context call. +/** This call balances each successful halide_metal_acquire_context call. * If halide_metal_acquire_context is replaced, this routine must be replaced * as well. */