zhangyiss/spack
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Optimize reuse from buildcaches (#30806)

Browse Source 
* database: don't sort on return from query_local

* ASP-based solver: don't build the hash-lookup dictionary twice

Building this dictionary twice and traversing all the specs
might be time-consuming for large buildcaches.
This commit is contained in:
Massimiliano Culpo 2022-08-02 20:43:14 +02:00 committed by GitHub
parent aeac72e1e3
commit 6e0b7959a4
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 13 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
lib/spack/spack/database.py
View File

@ -1525,9 +1525,14 @@ def _query(
_query.__doc__ += _query_docstring
def query_local(self, *args, **kwargs):
"""Query only the local Spack database."""
"""Query only the local Spack database.
This function doesn't guarantee any sorting of the returned
data for performance reason, since comparing specs for __lt__
may be an expensive operation.
"""
with self.read_transaction():
return sorted(self._query(*args, **kwargs))
return self._query(*args, **kwargs)
if query_local.__doc__ is None:
query_local.__doc__ = ""

15
lib/spack/spack/solver/asp.py
View File

@ -670,7 +670,7 @@ def stringify(x):
if result.satisfiable:
# build spec from the best model
builder = SpecBuilder(specs, reuse=reuse)
builder = SpecBuilder(specs, hash_lookup=setup.reusable_and_possible)
min_cost, best_model = min(models)
tuples = [(sym.name, [stringify(a) for a in sym.arguments]) for sym in best_model]
answers = builder.build_specs(tuples)
@ -718,6 +718,7 @@ def __init__(self, tests=False):
# hashes we've already added facts for
self.seen_hashes = set()
self.reusable_and_possible = {}
# id for dummy variables
self._condition_id_counter = itertools.count()
@ -1730,6 +1731,7 @@ def _facts_from_concrete_spec(self, spec, possible):
# be dependencies (don't tell it about the others)
h = spec.dag_hash()
if spec.name in possible and h not in self.seen_hashes:
self.reusable_and_possible[h] = spec
try:
# Only consider installed packages for repo we know
spack.repo.path.get(spec)
@ -1887,7 +1889,7 @@ class SpecBuilder(object):
#: Attributes that don't need actions
ignored_attributes = ["opt_criterion"]
def __init__(self, specs, reuse=None):
def __init__(self, specs, hash_lookup=None):
self._specs = {}
self._result = None
self._command_line_specs = specs
@ -1896,11 +1898,7 @@ def __init__(self, specs, reuse=None):
# Pass in as arguments reusable specs and plug them in
# from this dictionary during reconstruction
self._hash_lookup = {}
if reuse is not None:
for spec in reuse:
for node in spec.traverse():
self._hash_lookup.setdefault(node.dag_hash(), node)
self._hash_lookup = hash_lookup or {}
def hash(self, pkg, h):
if pkg not in self._specs:
@ -2205,8 +2203,7 @@ def _reusable_specs(self):
# Specs from buildcaches
try:
index = spack.binary_distribution.update_cache_and_get_specs()
reusable_specs.extend([s for s in index if not s.satisfies("dev_path=*")])
reusable_specs.extend(index)
except (spack.binary_distribution.FetchCacheError, IndexError):
# this is raised when no mirrors had indices.
# TODO: update mirror configuration so it can indicate that the