zhztheplayer commented on code in PR #5439: URL: https://github.com/apache/incubator-gluten/pull/5439#discussion_r1571725004
########## gluten-core/src/main/java/org/apache/gluten/memory/memtarget/spark/DynamicOffHeapSizingPolicyChecker.java: ########## @@ -0,0 +1,71 @@ +/* + * 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.gluten.memory.memtarget.spark; + +import org.apache.gluten.GlutenConfig; + +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.concurrent.atomic.AtomicLong; + +public final class DynamicOffHeapSizingPolicyChecker { + private static final Logger LOG = + LoggerFactory.getLogger(DynamicOffHeapSizingPolicyChecker.class); + private final long maxOnHeapMemoryInBytes = GlutenConfig.getConf().onHeapMemorySize(); + private final AtomicLong usedOffHeapBytes = new AtomicLong(); + + DynamicOffHeapSizingPolicyChecker() {} + + public boolean canBorrow(long size) { + if (size == 0) { + return true; + } + + long totalMemory = Runtime.getRuntime().totalMemory(); + long freeMemory = Runtime.getRuntime().freeMemory(); + long usedOnHeapBytes = (totalMemory - freeMemory); Review Comment: Thank for the explanation. > In this calculation if the on-heap was set to 10GB and 5GB is unused by any Java objects, usedOnHeapBytes would give us 5GB. The JVM memory footprint in this case is 10GB. With the calculations above, we would allow up to 5GB of memory to be used for off-heap. So in this case, the important one is that whether the total RSS of the JVM is calculated by which one of the following: 1. 10G total-heap + 5G off-heap = 15G 2. 5G used-heap + 5G off-heap = 10G The code seems to map to the second calculation. But I am concerned that probably the first one is the real situation. If user configured Gluten/Spark with a 10G memory limit but the actual RSS goes 15G, that is unexpected. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
