1cc1dc7a3Sopenharmony_ci# Profiling astcenc 2cc1dc7a3Sopenharmony_ci 3cc1dc7a3Sopenharmony_ciThis page contains some developer notes on profiling `astcenc` using command 4cc1dc7a3Sopenharmony_ciline tools on Linux hosts. 5cc1dc7a3Sopenharmony_ci 6cc1dc7a3Sopenharmony_ci## Building for profiling 7cc1dc7a3Sopenharmony_ci 8cc1dc7a3Sopenharmony_ciIt is recommended to profile release builds, but you will need debug symbols. 9cc1dc7a3Sopenharmony_ciIt is also recommended to disable link-time optimization to get call stacks 10cc1dc7a3Sopenharmony_cithat vaguely resemble the source code, although beware because this means that 11cc1dc7a3Sopenharmony_ciyou are not quite profiling the reality of full release builds. 12cc1dc7a3Sopenharmony_ci 13cc1dc7a3Sopenharmony_ciBoth of these can be achieved using the following CMake build type: 14cc1dc7a3Sopenharmony_ci 15cc1dc7a3Sopenharmony_ci```shell 16cc1dc7a3Sopenharmony_ci -DCMAKE_BUILD_TYPE=RelWithDebInfo 17cc1dc7a3Sopenharmony_ci``` 18cc1dc7a3Sopenharmony_ci 19cc1dc7a3Sopenharmony_ci## Running Callgrind tools 20cc1dc7a3Sopenharmony_ci 21cc1dc7a3Sopenharmony_ciWe provide a helper script that wraps Callgrind for hotspot profiling, although 22cc1dc7a3Sopenharmony_cibeware that it only currently supports profiling LDR input images and the 23cc1dc7a3Sopenharmony_cisingle compression mode. 24cc1dc7a3Sopenharmony_ci 25cc1dc7a3Sopenharmony_ciThis script requires the following tools on your `PATH`: 26cc1dc7a3Sopenharmony_ci 27cc1dc7a3Sopenharmony_ci * valgrind 28cc1dc7a3Sopenharmony_ci * gprof2dot 29cc1dc7a3Sopenharmony_ci * dot 30cc1dc7a3Sopenharmony_ci 31cc1dc7a3Sopenharmony_ciRun the helper script from the root of the repository using, e.g.: 32cc1dc7a3Sopenharmony_ci 33cc1dc7a3Sopenharmony_ci```shell 34cc1dc7a3Sopenharmony_cipython3 ./Test/astc_profile_valgrind.py <image.png> --test-quality fastest 35cc1dc7a3Sopenharmony_ci``` 36cc1dc7a3Sopenharmony_ci 37cc1dc7a3Sopenharmony_ciThe output will be two files: 38cc1dc7a3Sopenharmony_ci 39cc1dc7a3Sopenharmony_ci- perf_<quality>.png: an annotated call graph. 40cc1dc7a3Sopenharmony_ci- perf_<quality>.txt: the top N functions table. 41cc1dc7a3Sopenharmony_ci 42cc1dc7a3Sopenharmony_ci### Viewing disassembly 43cc1dc7a3Sopenharmony_ci 44cc1dc7a3Sopenharmony_ciStandard syntax x86-64 disassembly can be generated using: 45cc1dc7a3Sopenharmony_ci 46cc1dc7a3Sopenharmony_ci```shell 47cc1dc7a3Sopenharmony_ciobjdump -C -M intel --no-show-raw -d -S <binary> > dis.txt 48cc1dc7a3Sopenharmony_ci``` 49cc1dc7a3Sopenharmony_ci 50cc1dc7a3Sopenharmony_ci- - - 51cc1dc7a3Sopenharmony_ci 52cc1dc7a3Sopenharmony_ci_Copyright © 2020-2022, Arm Limited and contributors. All rights reserved._ 53
