Controlling Output File Names
Intel® SDE includes many analysis tools. These tools normally dump their analysis results into output file. Intel® SDE provides a central mechanism for controlling output files. Additionally Intel® SDE and Intel® Pin might emit log files for indicating error conditions.
By default the output files are created at the current working directory. Intel® SDE allows the user to specify a directory for all output files, use this option with the -odir knob. Alternatively, Intel® SDE allow the user to modify the working directory which ends with the same result. Use this option with the knobs -work-dir.
Multi-process applications create child processes and by default Intel® SDE inject itself to these child process. Preventing the multiple process from writing to the same output file is critical to ensure that the output from each process is intact. Intel® SDE provides an option to attach the process ID for every output file and avoid these conflicts. Additionally, on Linux Intel® SDE preserves the process ID hierarchy. Intel® SDE default behavior is keep the output file of the main process as defined by the tool (or the user), and then attach the process ID (and on Linux the parent process ID) to the output files of the child processes.
Here is the list of options available for controlling log files and output file names.
- -child_pid
Add process id to child process output file name [default 1]
- -file_suffix
Common suffix to add to output files (before .txt) [default ‘’]
- -i
Add process id to all output file name [default 0]
- -odir
Add directory to output file name [default ‘’]
- -pid_hierarchy
Add parent process id with pid to output file name (Linux only) [default 1]
- -work_dir
SDE working directory for files [default ‘’]
- -logfile
The log file path and file name [default pintool.log]
- -pin_log_file
Specify a pin log file [default pin-log.txt]
- -unique_logfile
The pin/sde log file names will contain the pid [default 0]