1## Inferno 2 3 4 5### Description 6 7Inferno is a flamegraph generator for native (C/C++) Android apps. It was 8originally written to profile and improve surfaceflinger performance 9(Android compositor) but it can be used for any native Android application 10. You can see a sample report generated with Inferno 11[here](./report.html). Report are self-contained in HTML so they can be 12exchanged easily. 13 14Notice there is no concept of time in a flame graph since all callstack are 15merged together. As a result, the width of a flamegraph represents 100% of 16the number of samples and the height is related to the number of functions on 17the stack when sampling occurred. 18 19 20 21 22In the flamegraph featured above you can see the main thread of SurfaceFlinger. 23It is immediatly apparent that most of the CPU time is spent processing messages 24`android::SurfaceFlinger::onMessageReceived`. The most expensive task is to ask 25 the screen to be refreshed as `android::DisplayDevice::prepare` shows in orange 26. This graphic division helps to see what part of the program is costly and 27where a developer's effort to improve performances should go. 28 29### Example of bottleneck 30 31A flamegraph give you instant vision on the CPU cycles cost centers but 32it can also be used to find specific offenders. To find them, look for 33plateaus. It is easier to see an example: 34 35 36 37In the previous flamegraph, two 38plateaus (due to `android::BufferQueueCore::validateConsistencyLocked`) 39are immediately apparent. 40 41### How it works 42Inferno relies on simpleperf to record the callstack of a native application 43thousands of times per second. Simpleperf takes care of unwinding the stack 44either using frame pointer (recommended) or dwarf. At the end of the recording 45`simpleperf` also symbolize all IPs automatically. The record are aggregated and 46dumps dumped to a file `perf.data`. This file is pulled from the Android device 47and processed on the host by Inferno. The callstacks are merged together to 48visualize in which part of an app the CPU cycles are spent. 49 50### How to use it 51 52Open a terminal and from `simpleperf/scripts` directory type: 53``` 54./inferno.sh (on Linux/Mac) 55inferno.bat (on Windows) 56``` 57 58Inferno will collect data, process them and automatically open your web browser 59to display the HTML report. 60 61### Parameters 62 63You can select how long to sample for, the color of the node and many other 64things. Use `-h` to get a list of all supported parameters. 65 66``` 67./inferno.sh -h 68``` 69 70### Troubleshooting 71 72#### Messy flame graph 73A healthy flame graph features a single call site at its base (see [here](./report.html)). 74If you don't see a unique call site like `_start` or `_start_thread` at the base 75from which all flames originate, something went wrong. : Stack unwinding may 76fail to reach the root callsite. These incomplete 77callstack are impossible to merge properly. By default Inferno asks 78 `simpleperf` to unwind the stack via the kernel and frame pointers. Try to 79 perform unwinding with dwarf `-du`, you can further tune this setting. 80 81 82#### No flames 83If you see no flames at all or a mess of 1 level flame without a common base, 84this may be because you compiled without frame pointers. Make sure there is no 85` -fomit-frame-pointer` in your build config. Alternatively, ask simpleperf to 86collect data with dward unwinding `-du`. 87 88 89 90#### High percentage of lost samples 91 92If simpleperf reports a lot of lost sample it is probably because you are 93unwinding with `dwarf`. Dwarf unwinding involves copying the stack before it is 94processed. Try to use frame pointer unwinding which can be done by the kernel 95and it much faster. 96 97The cost of frame pointer is negligible on arm64 parameter but considerable 98 on arm 32-bit arch (due to register pressure). Use a 64-bit build for better 99 profiling. 100 101#### run-as: package not debuggable 102If you cannot run as root, make sure the app is debuggable otherwise simpleperf 103will not be able to profile it. 104
