From 597013caa5e22c7134b6ca6e398659ba76047b15 Mon Sep 17 00:00:00 2001 From: Erik Johnston Date: Wed, 1 Jun 2016 18:01:22 +0100 Subject: [PATCH 1/3] Make cachedList go a bit faster --- synapse/metrics/metric.py | 22 +++++++++------- synapse/util/caches/descriptors.py | 42 +++++++++++++++++++++++------- 2 files changed, 45 insertions(+), 19 deletions(-) diff --git a/synapse/metrics/metric.py b/synapse/metrics/metric.py index 368fc2498..6f82b360b 100644 --- a/synapse/metrics/metric.py +++ b/synapse/metrics/metric.py @@ -15,6 +15,7 @@ from itertools import chain +from collections import Counter # TODO(paul): I can't believe Python doesn't have one of these @@ -55,30 +56,29 @@ class CounterMetric(BaseMetric): """The simplest kind of metric; one that stores a monotonically-increasing integer that counts events.""" + __slots__ = ("counts") + def __init__(self, *args, **kwargs): super(CounterMetric, self).__init__(*args, **kwargs) - self.counts = {} + self.counts = Counter() # Scalar metrics are never empty if self.is_scalar(): self.counts[()] = 0 def inc_by(self, incr, *values): - if len(values) != self.dimension(): - raise ValueError( - "Expected as many values to inc() as labels (%d)" % (self.dimension()) - ) + # if len(values) != self.dimension(): + # raise ValueError( + # "Expected as many values to inc() as labels (%d)" % (self.dimension()) + # ) # TODO: should assert that the tag values are all strings - if values not in self.counts: - self.counts[values] = incr - else: - self.counts[values] += incr + self.counts[values] += incr def inc(self, *values): - self.inc_by(1, *values) + self.counts[values] += 1 def render_item(self, k): return ["%s%s %d" % (self.name, self._render_key(k), self.counts[k])] @@ -132,6 +132,8 @@ class CacheMetric(object): This metric generates standard metric name pairs, so that monitoring rules can easily be applied to measure hit ratio.""" + __slots__ = ("name", "hits", "total", "size") + def __init__(self, name, size_callback, labels=[]): self.name = name diff --git a/synapse/util/caches/descriptors.py b/synapse/util/caches/descriptors.py index 758f5982b..4bbb16ed3 100644 --- a/synapse/util/caches/descriptors.py +++ b/synapse/util/caches/descriptors.py @@ -32,6 +32,7 @@ import os import functools import inspect import threading +import itertools logger = logging.getLogger(__name__) @@ -43,6 +44,14 @@ CACHE_SIZE_FACTOR = float(os.environ.get("SYNAPSE_CACHE_FACTOR", 0.1)) class Cache(object): + __slots__ = ( + "cache", + "max_entries", + "name", + "keylen", + "sequence", + "thread", + ) def __init__(self, name, max_entries=1000, keylen=1, lru=True, tree=False): if lru: @@ -293,16 +302,21 @@ class CacheListDescriptor(object): # cached is a dict arg -> deferred, where deferred results in a # 2-tuple (`arg`, `result`) - cached = {} + results = {} + cached_defers = {} missing = [] for arg in list_args: key = list(keyargs) key[self.list_pos] = arg try: - res = cache.get(tuple(key)).observe() - res.addCallback(lambda r, arg: (arg, r), arg) - cached[arg] = res + res = cache.get(tuple(key)) + if not res.called: + res = res.observe() + res.addCallback(lambda r, arg: (arg, r), arg) + cached_defers[arg] = res + else: + results[arg] = res.result except KeyError: missing.append(arg) @@ -340,12 +354,22 @@ class CacheListDescriptor(object): res = observer.observe() res.addCallback(lambda r, arg: (arg, r), arg) - cached[arg] = res + cached_defers[arg] = res - return preserve_context_over_deferred(defer.gatherResults( - cached.values(), - consumeErrors=True, - ).addErrback(unwrapFirstError).addCallback(lambda res: dict(res))) + if cached_defers: + return preserve_context_over_deferred(defer.gatherResults( + cached_defers.values(), + consumeErrors=True, + ).addCallback( + lambda res: { + k: v + for k, v in itertools.chain(results.items(), res) + } + )).addErrback( + unwrapFirstError + ) + else: + return results obj.__dict__[self.orig.__name__] = wrapped From e043ede4a2f18a47b67bf19368600183554824f7 Mon Sep 17 00:00:00 2001 From: Erik Johnston Date: Thu, 2 Jun 2016 11:52:32 +0100 Subject: [PATCH 2/3] Small optimisation to CacheListDescriptor --- synapse/metrics/metric.py | 22 ++++++++++------------ synapse/util/async.py | 9 +++++++++ synapse/util/caches/descriptors.py | 4 ++-- 3 files changed, 21 insertions(+), 14 deletions(-) diff --git a/synapse/metrics/metric.py b/synapse/metrics/metric.py index 6f82b360b..368fc2498 100644 --- a/synapse/metrics/metric.py +++ b/synapse/metrics/metric.py @@ -15,7 +15,6 @@ from itertools import chain -from collections import Counter # TODO(paul): I can't believe Python doesn't have one of these @@ -56,29 +55,30 @@ class CounterMetric(BaseMetric): """The simplest kind of metric; one that stores a monotonically-increasing integer that counts events.""" - __slots__ = ("counts") - def __init__(self, *args, **kwargs): super(CounterMetric, self).__init__(*args, **kwargs) - self.counts = Counter() + self.counts = {} # Scalar metrics are never empty if self.is_scalar(): self.counts[()] = 0 def inc_by(self, incr, *values): - # if len(values) != self.dimension(): - # raise ValueError( - # "Expected as many values to inc() as labels (%d)" % (self.dimension()) - # ) + if len(values) != self.dimension(): + raise ValueError( + "Expected as many values to inc() as labels (%d)" % (self.dimension()) + ) # TODO: should assert that the tag values are all strings - self.counts[values] += incr + if values not in self.counts: + self.counts[values] = incr + else: + self.counts[values] += incr def inc(self, *values): - self.counts[values] += 1 + self.inc_by(1, *values) def render_item(self, k): return ["%s%s %d" % (self.name, self._render_key(k), self.counts[k])] @@ -132,8 +132,6 @@ class CacheMetric(object): This metric generates standard metric name pairs, so that monitoring rules can easily be applied to measure hit ratio.""" - __slots__ = ("name", "hits", "total", "size") - def __init__(self, name, size_callback, labels=[]): self.name = name diff --git a/synapse/util/async.py b/synapse/util/async.py index 0d6f48e2d..40be7fe7e 100644 --- a/synapse/util/async.py +++ b/synapse/util/async.py @@ -102,6 +102,15 @@ class ObservableDeferred(object): def observers(self): return self._observers + def has_called(self): + return self._result is not None + + def has_succeeded(self): + return self._result is not None and self._result[0] is True + + def get_result(self): + return self._result[1] + def __getattr__(self, name): return getattr(self._deferred, name) diff --git a/synapse/util/caches/descriptors.py b/synapse/util/caches/descriptors.py index 4bbb16ed3..5be409727 100644 --- a/synapse/util/caches/descriptors.py +++ b/synapse/util/caches/descriptors.py @@ -311,12 +311,12 @@ class CacheListDescriptor(object): try: res = cache.get(tuple(key)) - if not res.called: + if not res.has_succeeded(): res = res.observe() res.addCallback(lambda r, arg: (arg, r), arg) cached_defers[arg] = res else: - results[arg] = res.result + results[arg] = res.get_result() except KeyError: missing.append(arg) From 58a224a6515dceacebc729f1e6fbb87a22f3a35a Mon Sep 17 00:00:00 2001 From: Erik Johnston Date: Fri, 3 Jun 2016 11:47:07 +0100 Subject: [PATCH 3/3] Pull out update_results_dict --- synapse/util/caches/descriptors.py | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/synapse/util/caches/descriptors.py b/synapse/util/caches/descriptors.py index 5be409727..799fd2a9c 100644 --- a/synapse/util/caches/descriptors.py +++ b/synapse/util/caches/descriptors.py @@ -32,7 +32,7 @@ import os import functools import inspect import threading -import itertools + logger = logging.getLogger(__name__) @@ -357,17 +357,16 @@ class CacheListDescriptor(object): cached_defers[arg] = res if cached_defers: + def update_results_dict(res): + results.update(res) + return results + return preserve_context_over_deferred(defer.gatherResults( cached_defers.values(), consumeErrors=True, - ).addCallback( - lambda res: { - k: v - for k, v in itertools.chain(results.items(), res) - } - )).addErrback( + ).addCallback(update_results_dict).addErrback( unwrapFirstError - ) + )) else: return results