[ty] Add environment variable to dump Salsa memory usage stats (#18928)

## Summary

Setting `TY_MEMORY_REPORT=full` will generate and print a memory usage
report to the CLI after a `ty check` run:

```
=======SALSA STRUCTS=======
`Definition`                                       metadata=7.24MB   fields=17.38MB  count=181062
`Expression`                                       metadata=4.45MB   fields=5.94MB   count=92804
`member_lookup_with_policy_::interned_arguments`   metadata=1.97MB   fields=2.25MB   count=35176
...
=======SALSA QUERIES=======
`File -> ty_python_semantic::semantic_index::SemanticIndex`
    metadata=11.46MB  fields=88.86MB  count=1638
`Definition -> ty_python_semantic::types::infer::TypeInference`
    metadata=24.52MB  fields=86.68MB  count=146018
`File -> ruff_db::parsed::ParsedModule`
    metadata=0.12MB   fields=69.06MB  count=1642
...
=======SALSA SUMMARY=======
TOTAL MEMORY USAGE: 577.61MB
    struct metadata = 29.00MB
    struct fields = 35.68MB
    memo metadata = 103.87MB
    memo fields = 409.06MB
```

Eventually, we should integrate these numbers into CI in some form. The
one limitation currently is that heap allocations in salsa structs (e.g.
interned values) are not tracked, but memoized values should have full
coverage. We may also want a peak memory usage counter (that accounts
for non-salsa memory), but that is relatively simple to profile manually
(e.g. `time -v ty check`) and would require a compile-time option to
avoid runtime overhead.
This commit is contained in:
Ibraheem Ahmed
2025-06-26 17:27:51 -04:00
committed by GitHub
parent a1579d82d0
commit 6f7b1c9bb3
79 changed files with 905 additions and 207 deletions

View File

@@ -30,7 +30,7 @@ pub fn resolve_module(db: &dyn Db, module_name: &ModuleName) -> Option<Module> {
///
/// This query should not be called directly. Instead, use [`resolve_module`]. It only exists
/// because Salsa requires the module name to be an ingredient.
#[salsa::tracked]
#[salsa::tracked(heap_size=get_size2::GetSize::get_heap_size)]
pub(crate) fn resolve_module_query<'db>(
db: &'db dyn Db,
module_name: ModuleNameIngredient<'db>,
@@ -95,7 +95,7 @@ impl std::fmt::Display for SystemOrVendoredPathRef<'_> {
/// Resolves the module for the file with the given id.
///
/// Returns `None` if the file is not a module locatable via any of the known search paths.
#[salsa::tracked]
#[salsa::tracked(heap_size=get_size2::GetSize::get_heap_size)]
pub(crate) fn file_to_module(db: &dyn Db, file: File) -> Option<Module> {
let _span = tracing::trace_span!("file_to_module", ?file).entered();
@@ -297,7 +297,7 @@ impl SearchPaths {
/// The editable-install search paths for the first `site-packages` directory
/// should come between the two `site-packages` directories when it comes to
/// module-resolution priority.
#[salsa::tracked(returns(deref))]
#[salsa::tracked(returns(deref), heap_size=get_size2::GetSize::get_heap_size)]
pub(crate) fn dynamic_resolution_paths(db: &dyn Db) -> Vec<SearchPath> {
tracing::debug!("Resolving dynamic module resolution paths");