| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Implement 6D parallelization with 1D and no tiling | Marat Dukhan | 2020-12-05 | 1 | -0/+271 |
| | | |||||
| * | Use cpuinfo_get_current_uarch_index_with_default for parallelization with uarch | Marat Dukhan | 2020-05-26 | 1 | -8/+8 |
| | | | | | index | ||||
| * | 3D/4D/5D parallelization functions with 1D or no tiling | Marat Dukhan | 2020-05-26 | 1 | -0/+667 |
| | | |||||
| * | Fast path using atomic decrement instead of atomic compare-and-swap | Marat Dukhan | 2020-05-02 | 1 | -37/+153 |
| | | | | | 50% higher throughput on x86 (disabled on other platforms) | ||||
| * | Replace atomic fetch_sub with decrement_fetch primitive | Marat Dukhan | 2020-04-10 | 1 | -13/+13 |
| | | | | | Decrement-fetch is a closer match to the primitive used in implementation | ||||
| * | Rewrite work spreading between threads | Marat Dukhan | 2020-04-10 | 1 | -27/+27 |
| | | | | | | | | - Avoid word x word -> doubleword multiplication - Avoid doubleword / word -> word division - Replace remaining division with multiplication via FXdiv - Improve portability through removal of platform-dependent multiply_divide function | ||||
| * | Windows implementation using Events | Marat Dukhan | 2020-04-07 | 1 | -3/+3 |
| | | |||||
| * | Optimized pthreadpool_parallelize_* functions | Marat Dukhan | 2020-04-05 | 1 | -397/+717 |
| | | | | | | Eliminate function call and division per each processed item in the multi-threaded case | ||||
| * | Refactor pthreadpool implementation | Marat Dukhan | 2020-04-01 | 1 | -0/+1010 |
| Split implementation into two types of components: - Components dependent on threading API - Portable components | |||||
 |
index : platform_external_pthreadpool | |
| Unnamed repository; edit this file 'description' to name the repository. |
| aboutsummaryrefslogtreecommitdiffstats |