1 i synthesize instructions events 2 b synthesize branches events (branch misses for Arm SPE) 3 c synthesize branches events (calls only) 4 r synthesize branches events (returns only) 5 x synthesize transactions events 6 w synthesize ptwrite events 7 p synthesize power events (incl. PSB events for Intel PT) 8 o synthesize other events recorded due to the use 9 of aux-output (refer to perf record) 10 I synthesize interrupt or similar (asynchronous) events 11 (e.g. Intel PT Event Trace) 12 e synthesize error events 13 d create a debug log 14 f synthesize first level cache events 15 m synthesize last level cache events 16 M synthesize memory events 17 t synthesize TLB events 18 a synthesize remote access events 19 g synthesize a call chain (use with i or x) 20 G synthesize a call chain on existing event records 21 l synthesize last branch entries (use with i or x) 22 L synthesize last branch entries on existing event records 23 s skip initial number of events 24 q quicker (less detailed) decoding 25 A approximate IPC 26 Z prefer to ignore timestamps (so-called "timeless" decoding) 27 28 The default is all events i.e. the same as --itrace=ibxwpe, 29 except for perf script where it is --itrace=ce 30 31 In addition, the period (default 100000, except for perf script where it is 1) 32 for instructions events can be specified in units of: 33 34 i instructions 35 t ticks 36 ms milliseconds 37 us microseconds 38 ns nanoseconds (default) 39 40 Also the call chain size (default 16, max. 1024) for instructions or 41 transactions events can be specified. 42 43 Also the number of last branch entries (default 64, max. 1024) for 44 instructions or transactions events can be specified. 45 46 Similar to options g and l, size may also be specified for options G and L. 47 On x86, note that G and L work poorly when data has been recorded with 48 large PEBS. Refer linkperf:perf-intel-pt[1] man page for details. 49 50 It is also possible to skip events generated (instructions, branches, transactions, 51 ptwrite, power) at the beginning. This is useful to ignore initialization code. 52 53 --itrace=i0nss1000000 54 55 skips the first million instructions. 56 57 The 'e' option may be followed by flags which affect what errors will or 58 will not be reported. Each flag must be preceded by either '+' or '-'. 59 The flags are: 60 o overflow 61 l trace data lost 62 63 If supported, the 'd' option may be followed by flags which affect what 64 debug messages will or will not be logged. Each flag must be preceded 65 by either '+' or '-'. The flags are: 66 a all perf events 67 o output to stdout 68 69 If supported, the 'q' option may be repeated to increase the effect. 70