1 Turbolizer 2 ========== 3 4 Turbolizer is a HTML-based tool that visualizes optimized code along the various 5 phases of Turbofan's optimization pipeline, allowing easy navigation between 6 source code, Turbofan IR graphs, scheduled IR nodes and generated assembly code. 7 8 Turbolizer consumes .json files that are generated per-function by d8 by passing 9 the '--trace-turbo' command-line flag. 10 11 Host the turbolizer locally by starting a web server that serves the contents of 12 the turbolizer directory, e.g.: 13 14 cd src/tools/turbolizer 15 python -m SimpleHTTPServer 8000 16 17 Optionally, profiling data generated by the perf tools in linux can be merged 18 with the .json files using the turbolizer-perf.py file included. The following 19 command is an example of using the perf script: 20 21 perf script -i perf.data.jitted -s turbolizer-perf.py turbo-main.json 22 23 The output of the above command is a json object that can be piped to a file 24 which, when uploaded to turbolizer, will display the event counts from perf next 25 to each instruction in the disassembly. Further detail can be found in the 26 bottom of this document under "Using Perf with Turbo." 27 28 Using the python interface in perf script requires python-dev to be installed 29 and perf be recompiled with python support enabled. Once recompiled, the 30 variable PERF_EXEC_PATH must be set to the location of the recompiled perf 31 binaries. 32 33 Graph visualization and manipulation based on Mike Bostock's sample code for an 34 interactive tool for creating directed graphs. Original source is at 35 https://github.com/metacademy/directed-graph-creator and released under the 36 MIT/X license. 37 38 Icons derived from the "White Olive Collection" created by Breezi released under 39 the Creative Commons BY license. 40 41 Using Perf with Turbo 42 --------------------- 43 44 In order to generate perf data that matches exactly with the turbofan trace, you 45 must use either a debug build of v8 or a release build with the flag 46 'disassembler=on'. This flag ensures that the '--trace-turbo' will output the 47 necessary disassembly for linking with the perf profile. 48 49 The basic example of generating the required data is as follows: 50 51 perf record -k mono /path/to/d8 --turbo --trace-turbo --perf-prof main.js 52 perf inject -j -i perf.data -o perf.data.jitted 53 perf script -i perf.data.jitted -s turbolizer-perf.py turbo-main.json 54 55 These commands combined will run and profile d8, merge the output into a single 56 'perf.data.jitted' file, then take the event data from that and link them to the 57 disassembly in the 'turbo-main.json'. Note that, as above, the output of the 58 script command must be piped to a file for uploading to turbolizer. 59 60 There are many options that can be added to the first command, for example '-e' 61 can be used to specify the counting of specific events (default: cycles), as 62 well as '--cpu' to specify which CPU to sample.