Github user aarondav commented on a diff in the pull request:
https://github.com/apache/spark/pull/1707#discussion_r15726557
--- Diff:
core/src/main/scala/org/apache/spark/shuffle/ShuffleMemoryManager.scala ---
@@ -0,0 +1,114 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.shuffle
+
+import org.apache.spark.{SparkException, SparkConf}
+import scala.collection.mutable
+
+/**
+ * Allocates a pool of memory to task threads for use in shuffle
operations. Each disk-spilling
+ * collection (ExternalAppendOnlyMap or ExternalSorter) used by these
tasks can acquire memory
+ * from this pool and release it as it spills data out. When a task ends,
all its memory will be
+ * released by the Executor.
+ *
+ * This class tries to ensure that each thread gets a reasonable share of
memory, instead of some
+ * thread ramping up to a large amount first and then causing others to
spill to disk repeatedly.
+ * If there are N threads, it ensures that each thread can acquire at
least 1 / 2N of the memory
+ * before it has to spill, and at most 1 / N. Because N varies
dynamically, we keep track of the
+ * set of active threads and redo the calculations of 1 / 2N and 1 / N in
waiting threads whenever
+ * this set changes. This is all done by synchronizing access on "this" to
mutate state and using
+ * wait() and notifyAll() to signal changes.
+ */
+private[spark] class ShuffleMemoryManager(maxMemory: Long) {
+ private val threadMemory = new mutable.HashMap[Long, Long]() //
threadId -> memory bytes
+
+ def this(conf: SparkConf) = this(ShuffleMemoryManager.getMaxMemory(conf))
+
+ /**
+ * Try to acquire numBytes memory for the current thread, or return
false if the pool cannot
+ * allocate this much memory to it. This call may block until there is
enough free memory in
+ * some situations, to make sure each thread has a chance to ramp up to
a reasonable share of
+ * the available memory before being forced to spill.
+ */
+ def tryToAcquire(numBytes: Long): Boolean = synchronized {
+ val threadId = Thread.currentThread().getId
+
+ // Add this thread to the threadMemory map just so we can keep an
accurate count of the number
+ // of active threads, to let other threads ramp down their memory in
calls to tryToAcquire
+ threadMemory.getOrElseUpdate(threadId, 0L)
+
+ // Keep looping until we're either sure that we don't want to grant
this request (because this
+ // thread would have more than 1 / numActiveThreads of the memory) or
we have enough free
+ // memory to give it (we always let each thread get at least 1 / (2 *
numActiveThreads)).
+ while (true) {
+ val numActiveThreads = threadMemory.keys.size
+ val curMem = threadMemory(threadId)
+ if (curMem + numBytes > maxMemory / numActiveThreads) {
+ // We'd get more than 1 / numActiveThreads of the total memory;
don't allow that
+ return false
+ }
+ val bytesFree = maxMemory - threadMemory.values.sum
+ if (bytesFree >= numBytes) {
+ // Grant the request
+ threadMemory(threadId) = curMem + numBytes
+ // Notify other waiting threads because the # active of threads
may have increased, so
+ // they may cancel their current waits
+ notifyAll()
+ return true
+ } else if (curMem + numBytes <= maxMemory / (2 * numActiveThreads)) {
+ // This thread has so little memory that we want it to block and
acquire a bigger
+ // amount instead of cancelling the request. Wait on "this" for a
thread to call notify.
+ // Before doing the wait, however, also notify other current
waiters in case our thread
+ // becoming active just pushed them over the limit to give up
their own waits.
+ notifyAll()
+ wait()
+ } else {
+ // Thread would have between 1 / (2 * numActiveThreads) and 1 /
numActiveThreads memory
+ return false
+ }
+ }
+ false // Never reached
+ }
+
+ /** Release numBytes bytes for the current thread. */
+ def release(numBytes: Long): Unit = synchronized {
+ val threadId = Thread.currentThread().getId
+ val curMem = threadMemory.getOrElse(threadId, 0L)
+ if (curMem < numBytes) {
+ throw new SparkException(
+ s"Internal error: release called on ${numBytes} bytes but thread
only has ${curMem}")
+ }
+ threadMemory(threadId) -= numBytes
+ notifyAll() // Notify waiters who locked "this" in tryToAcquire that
memory has freed
+ }
+
+ /** Release all memory for the current thread and mark it as inactive
(e.g. when a task ends). */
+ def releaseMemoryForThisThread(): Unit = synchronized {
+ val threadId = Thread.currentThread().getId
+ threadMemory.remove(threadId)
+ notifyAll() // Notify waiters who locked "this" in tryToAcquire that
memory has freed
+ }
+}
+
+private object ShuffleMemoryManager {
+ def getMaxMemory(conf: SparkConf): Long = {
--- End diff --
Maybe a comment here on what this is doing (and why it uses two fractions,
possibly).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
