Repository navigation

Implements `top_k` in dpctl.tensor - #1921

Merged

ndgrigorian merged 30 commits into

masterfrom

feature/topk

Jan 5, 2025

Merged

Implements `top_k` in dpctl.tensor#1921
ndgrigorian merged 30 commits into
masterfrom
feature/topk

Commits on Jan 5, 2025
- Implements `top_k`
  ndgrigorian
  committed
  0b353bf
- Add `top_k` to doc sources
  ndgrigorian
  committed
  f809573
- Use `std::size_t` for `k` instead of `py::ssize_t`
  ndgrigorian
  committed
  b59e59d
- Add implementation of `top_k` using radix sort
  ndgrigorian
  committed
  b4d7ba4
- Reduce code duplication by using std::optional in py_topk directly
  ndgrigorian
  committed
  c37f1de
- Clean up top_k docstring
  ndgrigorian
  committed
  1f0e9fd
- Move top_k into _sorting.py
  ndgrigorian
  committed
  fa74e5e
- top_k raises when k > 1 and input is a scalar
  ndgrigorian
  committed
  cd83881
- Fix bug in top_k partial merge sort implementation
  ndgrigorian
  committed
  70c2ef0
- Add test file for top_k functionality
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  65f14be
- Add sort_utils.hpp with iota_impl
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  275827a
- Simplify write-out kernels in topk implementation (avoid recomputing gid)
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  cd1243f
- Add explicit read into registers
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  823c201
- Use unique_ptr as temporary owner of USM allocation
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  3a4d8ab
- Factor out write-out kernel into separate detail function
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  b15e302
- Typo: smart_malloc_jost -> smart_malloc_host
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  4ef615e
- Reimplemented map_back_impl to process few elements per work-item
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  2608aea
- Replace use of std::log2(size_t_value)
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  15da303
- Remove dead code disable via preprocessor conditional
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  fb9bc43
- Add information displayed on failure, renamed variables
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  6cf36b1
- Add comment to explain ceil_log2 algo
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  4ec252e
- Add static assers to async_smart_free
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  1436522
- Skip top_k test known to fail for some CPU architectures
  ndgrigorian
  committed
  f0e3245
- Fixed blunder in work-item id to data_id computation
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  8f4dcdd
- Replace map_back_impl in sort_utils
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  b6806ba
- Use get_global_linear_id instead of get_global_id and rely on implicit conversion
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  b47d9f3
- Counters in one-workgroup kernel to use uint16_t from uint32_t
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  0493485
- Apply work-around for failing tests with CPU device and short sub-groups
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  5125e11
- Remove skipping of tests for i1/i2 dtypes since work-around
  
  oleksandr-pavlyk
  authored and
  ndgrigorian
  committed
  505b64c
- Add entry to changelog for top_k
  ndgrigorian
  committed
  8c6abf5

Commits

Commits on Jan 5, 2025