LLVM-REMARK-SIZE-DIFF(1) LLVM LLVM-REMARK-SIZE-DIFF(1)

llvm-remark-size-diff - diff size remarks

llvm-remark-size-diff [options] file_a file_b --parser parser

llvm-remark-size-diff diffs size remarks in two remark files: file_a and file_b.

llvm-remark-size-diff can be used to gain insight into which functions were impacted the most by code generation changes.

In most common use-cases file_a and file_b will be remarks output by compiling a fixed source with differing compilers or differing optimization settings.

llvm-remark-size-diff handles both YAML and bitstream remarks.

Select the type of input remark parser. Required. * yaml: The tool will parse YAML remarks. * bitstream: The tool will parse bitstream remarks.
Output style. * human: Human-readable textual report. Default option. * json: JSON report.
Pretty-print JSON output. Optional.

If output is not set to JSON, this does nothing.

Output file for the report. Outputs to stdout by default.

The human-readable format for llvm-remark-size-diff is composed of two sections:

  • Per-function changes.
  • A high-level summary of all changes.

Suppose you are comparing two remark files OLD and NEW.

For each function with a changed instruction count in OLD and NEW, llvm-remark-size-diff will emit a line like below:

(++|--|==) (>|<) function_name, N instrs, M stack B

A breakdown of the format is below:

(++|--|==)
Which of OLD and NEW the function_name is present in.
  • ++: Only in NEW. ("Added")
  • --: Only in OLD. ("Removed")
  • ==: In both.
(>|<)
Denotes if function_name has more instructions or fewer instructions in the second file.
  • >: More instructions in second file than first file.
  • <: Fewer instructions in second file than in first file.
The name of the changed function.
Second file instruction count - first file instruction count.
Second file stack byte count - first file stack byte count.

llvm-remark-size-diff will output a high-level summary after printing all changed functions.

instruction count: N (inst_pct_change%)
stack byte usage: M (sb_pct_change%)
Sum of all instruction count changes between the second and first file.
Percent increase or decrease in instruction count between the second and first file.
Sum of all stack byte count changes between the second and first file.
Percent increase or decrease in stack byte usage between the second and first file.

Suppose we are comparing two files, OLD and NEW.

llvm-remark-size-diff will output JSON as follows.

"Files": [
  "A": "path/to/OLD",
  "B": "path/to/NEW"
]
"InBoth": [
  ...
],
"OnlyInA": [
  ...
],
"OnlyInB": [
  ...
]
Original paths to remark files.
  • A: Path to the first file.
  • B: Path to the second file.
Functions present in both files.
Functions only present in the first file.
Functions only present in the second file.

The InBoth, OnlyInA, and OnlyInB sections contain size information for each function in the input remark files.

{
  "FunctionName" : "function_name"
  "InstCount": [
      INST_COUNT_A,
      INST_COUNT_B
    ],
  "StackSize": [
      STACK_BYTES_A,
      STACK_BYTES_B
    ],
}
Name of the function.
Instruction counts for the function.
  • INST_COUNT_A: Instruction count in OLD.
  • INST_COUNT_B: Instruction count in NEW.
Stack byte counts for the function.
  • STACK_BYTES_A: Stack bytes in OLD.
  • STACK_BYTES_B: Stack bytes in NEW.

Function JSON does not contain the diffs. Tools consuming JSON output from llvm-remark-size-diff are responsible for computing the diffs separately.

To compute the diffs:

  • Instruction count diff: INST_COUNT_B - INST_COUNT_A
  • Stack byte count diff: STACK_BYTES_B - STACK_BYTES_A

llvm-remark-size-diff returns 0 on success, and a non-zero value otherwise.

Maintained by the LLVM Team (https://llvm.org/).

2003-2024, LLVM Project

2024-05-21 17