Misc. Build Options

• -llvm-features=<features> Control LLVM target features used when generating machine code. These are architecture-specific capabilities, such as instruction set extensions, not Go language features or build tags.

  Pass a comma-separated list without spaces. Use +name to enable a feature and -name to disable it. For example, when building for an x86-64 CPU that supports AVX2:

  GOOS=linux GOARCH=amd64 tinygo build -llvm-features=+avx2 -o app .
  

  TinyGo appends these settings to the selected target’s features list. LLVM also derives defaults from the selected CPU. Normally the target defaults are sufficient; changing them can produce code that will not run on the intended hardware or WebAssembly engine.

  Available feature names depend on the target architecture and LLVM version. To list them, install LLVM and use its llvm-as and llc tools:

  llvm-as < /dev/null | llc -march=arm -mattr=help
  

  Replace arm with the LLVM architecture, such as aarch64, x86, riscv32, or wasm32. The output lists CPUs and features with brief descriptions. Use the LLVM version reported by tinygo version to get a matching list. These tools are not required for normal TinyGo builds and may need to be installed separately.

  See LLVM’s llc documentation for more about target features.

• -no-debug Disable outputting debug symbols. This can be useful for WebAssembly, as there is no debugger for .wasm files yet and .wasm files are generally served directly. Avoiding debug symbols can have a big impact on generated binary size, reducing them by more than half. This is not necessary on microcontrollers because debugging symbols are not flashed to the microcontroller. Additionally, you will need it when you use tinygo gdb. In general, it is recommended to include debug symbols unless you have a good reason not to. Note: while there is some support for debug symbols, only line numbers have been implemented so far. That means single-stepping and stacktraces work just fine, but no variables can be inspected.

• -size Print size (none, short, or full) of the output (linked) binary. Note that the calculated size includes RAM reserved for the stack.

  • none (default) Print nothing.

  • short Print size statistics, roughly like what the size binutils program would print but with useful flash and RAM columns:

    code    data     bss |   flash     ram
    5780     144    2132 |    5924    2276
    

  • full Try to determine per package how much space is used. Note that these calculations are merely guesses and can sometimes be way off due to various reasons like inlining:

    code  rodata    data     bss |   flash     ram | package
    876        0       4       0 |     880       4 | (bootstrap)
    38         0       0       0 |      38       0 | device/arm
    0          0       0      66 |       0      66 | machine
    2994     440     124       0 |    3558     124 | main
    948      127       4       1 |    1079       5 | runtime
    4856     567     132      67 |    5555     199 | (sum)
    5780       -     144    2132 |    5924    2276 | (all)
    

• -print-allocs=<regex> Print heap allocations in the program and why they happen. You can use -print-allocs=. if you want to see all heap allocations in the program. You can also match the function name, for example you can use -print-allocs=strconv to print heap allocations that contain the strconv substring (which includes all functions from the strconv standard library package).