Modifying a computer program configuration based on variable-bin histograms

A method of modifying a configuration of a computer program, comprising:

• receiving at a second computer, a plurality of encoded sample values from a sample collector on a first computer, the encoded sample values generated from the operation of the computer program;
• for each of the encoded sample values, determining an associated histogram structure based on a seed value, wherein the seed value comprises a random seed used to set a first value for a pseudorandom sequence and further specifies how the pseudorandom sequence is generated;
• selecting an analysis number within the range of values in the determined associated histogram structures of the received encoded sample values;
• determining a total for each selected analysis number, based on the selected analysis number, a bin analysis range, the received bin values and the determined histogram structure for each received bin value;
• generating a plurality of totals and associated analysis numbers by repeating the selecting and determining a number of times;
• generating a result based on the plurality of generated totals and associated analysis numbers;
• generating a configuration setting of a computer program configured to operate on the first computer based on the result;
• and
• sending the configuration setting to the first computer.

A system for modifying a configuration of a computer program, comprising:

• an encoded data receiver configured to receive a plurality of encoded data values from a sample collector on a first computer, the encoded values generated based on the operation of the computer program;
• a bin generation value selector, configured to select a bin generation value for each of the encoded data values based on a seed value, wherein the seed value comprises a random seed used to set a first value for a pseudorandom sequence and further specifies how the pseudorandom sequence is generated;
• a histogram structure reconstructor implemented on a processor that configured, when executed, reconstructs, for each of the received encoded data values, an associated histogram structure based on the bin generation value associated with each encoded data value;
• a bin data analyzer configured to generate a result based on the received encoded data values and the reconstructed histogram structure associated with each encoded data value;
• and
• a configuration generator configured to generate a configuration for a computer program setting based on the result and send the configuration setting to the first computer.

A computer-readable memory having computer-executable instructions stored thereon that, when executed by a computing device, cause the computing device to perform a method of modifying a configuration of a computer program, comprising:

• monitoring an operation of the computer program, the operation being performed on a first computer;
• collecting a plurality of data values based on the monitoring, the data values having a given data type, wherein the data type comprises the numeric range of possible values of the data values;
• generating a histogram structure for each of selected ones of the plurality of data values plotted on the histogram structure, the histogram structure being based on the data type and a seed value, wherein the seed value comprises a random seed used to set a first value for a pseudorandom sequence and further specifies how the pseudorandom sequence is generated;
• generating an encoded sample value for each of the selected ones of the plurality of data values indicating the generated histogram structure associated with each selected data value;
• sending the encoded sample values to a second computer;
• receiving a configuration setting from the second computer based on the encoded sample values and the seed value;
• and
• modifying the configuration of the computer program by applying the configuration setting to the computer program.

A method of sending data from a first computer to a second computer, comprising:

• collecting a plurality of data values on the first computer, the data values having a given data type, wherein the data type comprises the numeric range of possible values of the data values;
• generating a histogram structure for each of the plurality of data values based on the data type and a seed value, wherein the seed value comprises a random seed used to set a first value for a pseudorandom sequence and further specifies how the pseudorandom sequence is generated;
• generating an encoded sample value for each of the plurality of data values, based on each data value and the generated histogram structure associated with each data value;
• and
• sending the encoded sample values and the seed value to the second computer, wherein at the second computer, based on the encoded sample values and the seed value, a cumulative distribution of the collected plurality of data values is generated.