onnxruntime/include/onnxruntime/core
Maximilian Müller 2eeafc37bc
Enable global TRT timing cache (#17865)
I am adding a new `trt_timing_cache_path` option. Internally it is
handled as `global_cache_path_` and will be set via a fall through
approach:
1. no path provided => workdir
2. `trt_engine_cache_path` provided but no `trt_timing_cache_path` =>
`trt_engine_cache_path`
3. `trt_timing_cache_path` provided => `trt_timing_cache_path` (if not
provided `trt_engine_cache_path` will still be workdir)

### Motivation and Context

A TRT timing cache can be reused across multiple models as it only holds
kernel timings and it is common that network "patterns" are reused. This
can accelerate build times a lot.

---------

Co-authored-by: Carson M <carson@pyke.io>
2023-10-27 09:23:19 -07:00
..
common remove gsl to make status.h independent from gsl (#17402) 2023-09-13 21:47:43 -07:00
eager Run clang-format in CI (#15524) 2023-04-18 09:26:58 -07:00
framework implement isinf20 and isnan20 (#17874) 2023-10-24 10:58:54 -07:00
graph Functions Ahead Of Time inlininng (#17764) 2023-10-23 17:42:20 -07:00
optimizer fix compilation error in no absl build (#15769) 2023-05-02 08:20:49 -07:00
platform Re-implement stacktrace (#17173) 2023-08-16 16:07:49 -07:00
providers Enable global TRT timing cache (#17865) 2023-10-27 09:23:19 -07:00
session Functions Ahead Of Time inlininng (#17764) 2023-10-23 17:42:20 -07:00