ProcessSpeed

Basic Usage

processspeed.sh <file>

Where <file> is a text file containing output from the Linux time command in the format produced by time -p or similar timing utilities.

Parameters

This tool takes only one positional argument and has no configurable parameters:

Input

<file>: Input file containing timing data with lines starting with "real", "user", and "sys" followed by tab-separated time values in minutes:seconds format (e.g., "real\t1m23.456s")

Input Format

The input file should contain timing output from the Linux time command with lines in this format:

real	1m23.456s
user	0m45.123s
sys	0m12.789s

Each line consists of:

Label: "real", "user", or "sys"
Tab separator
Time value: In format XmY.ZZZs where X is minutes, Y.ZZZ is seconds

Output Format

The tool outputs a header line followed by timing data converted to decimal seconds:

#real	user	sys
83.456	45.123	12.789

All time values are converted to decimal seconds for easy parsing and analysis.

Examples

Basic Usage

# Run a command with timing
time -p my_program > timing_output.txt

# Process the timing results
processspeed.sh timing_output.txt

This will parse the timing output and display the results in decimal seconds.

Batch Processing Multiple Timing Files

# Process multiple timing files
for file in timing_*.txt; do
    echo "Processing $file"
    processspeed.sh "$file"
done

Process multiple timing result files to summarize performance data across different runs.

Creating CSV Output

# Create CSV-formatted output for spreadsheet analysis
echo "filename,real,user,sys" > timing_summary.csv
for file in timing_*.txt; do
    result=$(processspeed.sh "$file" | tail -1)
    echo "$file,$result" >> timing_summary.csv
done

Generate a CSV file with timing summaries from multiple runs for further analysis.

Algorithm Details

ProcessSpeed2 implements a straightforward time parsing algorithm using standard Java I/O classes:

File Processing Implementation

TextFile-based Reading: Uses TextFile.nextLine() method for sequential file processing with automatic resource management
String.startsWith() Detection: Identifies timing lines using exact prefix matching ("real\t", "user\t", "sys\t")
String.split() Extraction: Isolates time values using tab-delimited splitting with split("\t")[1] array indexing
toSeconds() Method: Custom conversion method implementing string manipulation and arithmetic conversion

Time Conversion Implementation

The toSeconds() method implements this specific algorithm:

String.replaceAll("s", ""): Removes trailing 's' character using regular expression replacement
String.split("m"): Separates on 'm' delimiter, creating split[0]=minutes, split[1]=seconds
Double.parseDouble(): Converts string components to double-precision floating point numbers
Arithmetic Formula: Calculates 60*minutes + seconds using standard floating-point arithmetic
Tools.format("%.3f"): Uses BBTools Tools class for 3-decimal formatting with printf-style format specifiers

Output Stream Management

The tool coordinates output through System.out using specific formatting:

Header Line: System.println("#real\tuser\tsys") writes tab-delimited column headers
Synchronized Printing: System.print() calls ensure real/user values followed by System.print() with newline for sys
Tools.format() Precision: Maintains consistent "%.3f\t" and "%.3f\n" format strings for decimal alignment

Use Cases

This utility is particularly useful for:

Performance Analysis: Converting timing data to machine-readable decimal format for statistical analysis
Benchmarking: Processing timing results from multiple test runs
Automation: Integrating timing data collection into automated testing pipelines
Comparison Studies: Standardizing timing output format for comparison across different tools or parameters

Memory Requirements

ProcessSpeed2 is designed for minimal memory usage:

Default Memory: 120MB (-Xmx120m) is allocated, which is sufficient for processing timing files of any reasonable size
Streaming Processing: Reads and processes the input file line-by-line, maintaining constant memory usage regardless of input size
No Data Accumulation: Does not store processed data in memory, immediately outputting converted values

Error Handling

The tool expects properly formatted input and will encounter issues with:

Malformed Time Strings: Input not matching the expected XmY.ZZZs format
Missing Components: Lines lacking proper tab separation or missing time components
Invalid Numbers: Non-numeric values in the minutes or seconds fields

For best results, ensure input files contain only properly formatted timing output from the Linux time command.

Related Tools

ProcessSpeed2 is part of the BBTools ecosystem for performance analysis:

BBTools Performance Utilities: Many BBTools scripts generate timing information that can be processed with this tool
Benchmarking Scripts: Can be integrated with shell scripts that benchmark BBTools performance
Statistical Analysis: Output can be fed into statistical analysis tools for performance characterization

Support

For questions and support:

Email: bbushnell@lbl.gov
Documentation: bbmap.org