This is works in the perf.data importer and should work (though I haven't tested it yet) with samply record on Linux.
It creates a synthetic thread per CPU, and also a combined thread (the main thread of the synthetic "CPU" process) which contains all samples from all threads. Example: https://share.firefox.dev/3W8qt2f
This is works in the perf.data importer and should work (though I haven't tested it yet) with
samply recordon Linux.It creates a synthetic thread per CPU, and also a combined thread (the main thread of the synthetic "CPU" process) which contains all samples from all threads. Example: https://share.firefox.dev/3W8qt2f