[
https://issues.apache.org/jira/browse/HIVE-23880?focusedWorklogId=470123&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-470123
]
ASF GitHub Bot logged work on HIVE-23880:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 13/Aug/20 07:50
Start Date: 13/Aug/20 07:50
Worklog Time Spent: 10m
Work Description: abstractdog commented on a change in pull request #1280:
URL: https://github.com/apache/hive/pull/1280#discussion_r469762209
##########
File path:
ql/src/java/org/apache/hadoop/hive/ql/exec/vector/expressions/aggregates/VectorUDAFBloomFilterMerge.java
##########
@@ -77,6 +75,211 @@ public void reset() {
// Do not change the initial bytes which contain
NumHashFunctions/NumBits!
Arrays.fill(bfBytes, BloomKFilter.START_OF_SERIALIZED_LONGS,
bfBytes.length, (byte) 0);
}
+
+ public boolean mergeBloomFilterBytesFromInputColumn(BytesColumnVector
inputColumn,
+ int batchSize, boolean selectedInUse, int[] selected, Configuration
conf) {
+ // already set in previous iterations, no need to call initExecutor again
+ if (numThreads == 0) {
+ return false;
+ }
+ if (executor == null) {
+ initExecutor(conf, batchSize);
+ if (!isParallel) {
+ return false;
+ }
+ }
+
+ // split every bloom filter (represented by a part of a byte[]) across
workers
+ for (int j = 0; j < batchSize; j++) {
+ if (!selectedInUse && inputColumn.noNulls) {
+ splitVectorAcrossWorkers(workers, inputColumn.vector[j],
inputColumn.start[j],
+ inputColumn.length[j]);
+ } else if (!selectedInUse) {
+ if (!inputColumn.isNull[j]) {
+ splitVectorAcrossWorkers(workers, inputColumn.vector[j],
inputColumn.start[j],
+ inputColumn.length[j]);
+ }
+ } else if (inputColumn.noNulls) {
+ int i = selected[j];
+ splitVectorAcrossWorkers(workers, inputColumn.vector[i],
inputColumn.start[i],
+ inputColumn.length[i]);
+ } else {
+ int i = selected[j];
+ if (!inputColumn.isNull[i]) {
+ splitVectorAcrossWorkers(workers, inputColumn.vector[i],
inputColumn.start[i],
+ inputColumn.length[i]);
+ }
+ }
+ }
+
+ return true;
+ }
+
+ private void initExecutor(Configuration conf, int batchSize) {
+ numThreads =
conf.getInt(HiveConf.ConfVars.TEZ_BLOOM_FILTER_MERGE_THREADS.varname,
+ HiveConf.ConfVars.TEZ_BLOOM_FILTER_MERGE_THREADS.defaultIntVal);
+ LOG.info("Number of threads used for bloom filter merge: {}",
numThreads);
+
+ if (numThreads < 0) {
+ throw new RuntimeException(
+ "invalid number of threads for bloom filter merge: " + numThreads);
+ }
+ if (numThreads == 0) { // disable parallel feature
+ return; // this will leave isParallel=false
+ }
+ isParallel = true;
+ executor = Executors.newFixedThreadPool(numThreads);
+
+ workers = new BloomFilterMergeWorker[numThreads];
+ for (int f = 0; f < numThreads; f++) {
+ workers[f] = new BloomFilterMergeWorker(bfBytes, 0, bfBytes.length);
+ }
+
+ for (int f = 0; f < numThreads; f++) {
+ executor.submit(workers[f]);
+ }
+ }
+
+ public int getNumberOfWaitingMergeTasks(){
+ int size = 0;
+ for (BloomFilterMergeWorker w : workers){
+ size += w.queue.size();
+ }
+ return size;
+ }
+
+ public int getNumberOfMergingWorkers() {
+ int working = 0;
+ for (BloomFilterMergeWorker w : workers) {
+ if (w.isMerging.get()) {
+ working += 1;
+ }
+ }
+ return working;
+ }
+
+ private static void splitVectorAcrossWorkers(BloomFilterMergeWorker[]
workers, byte[] bytes,
+ int start, int length) {
+ if (bytes == null || length == 0) {
+ return;
+ }
+ /*
+ * This will split a byte[] across workers as below:
+ * let's say there are 10 workers for 7813 bytes, in this case
+ * length: 7813, elementPerBatch: 781
+ * bytes assigned to workers: inclusive lower bound, exclusive upper
bound
+ * 1. worker: 5 -> 786
+ * 2. worker: 786 -> 1567
+ * 3. worker: 1567 -> 2348
+ * 4. worker: 2348 -> 3129
+ * 5. worker: 3129 -> 3910
+ * 6. worker: 3910 -> 4691
+ * 7. worker: 4691 -> 5472
+ * 8. worker: 5472 -> 6253
+ * 9. worker: 6253 -> 7034
+ * 10. worker: 7034 -> 7813 (last element per batch is: 779)
+ *
+ * This way, a particular worker will be given with the same part
+ * of all bloom filters along with the shared base bloom filter,
+ * so the bitwise OR function will not be a subject of threading/sync
issues.
+ */
+ int elementPerBatch =
+ (int) Math.ceil((double) (length - START_OF_SERIALIZED_LONGS) /
workers.length);
+
+ for (int w = 0; w < workers.length; w++) {
+ int modifiedStart = START_OF_SERIALIZED_LONGS + w * elementPerBatch;
+ int modifiedLength = (w == workers.length - 1)
+ ? length - (START_OF_SERIALIZED_LONGS + w * elementPerBatch) :
elementPerBatch;
+
+ ElementWrapper wrapper =
+ new ElementWrapper(bytes, start, length, modifiedStart,
modifiedLength);
+ workers[w].add(wrapper);
+ }
+ }
+
+ public void shutdownAndWaitForMergeTasks() {
+ /**
+ * Executor.shutdownNow() is supposed to send Thread.interrupt to worker
threads, and they are
+ * supposed to finish their work.
+ */
+ executor.shutdownNow();
+ try {
+ executor.awaitTermination(180, TimeUnit.SECONDS);
+ } catch (InterruptedException e) {
+ LOG.warn("Bloom filter merge is interrupted while waiting to finish,
this is unexpected",
+ e);
+ }
+ }
+ }
+
+ private static class BloomFilterMergeWorker implements Runnable {
+ private BlockingQueue<ElementWrapper> queue;
+ private byte[] bfAggregation;
+ private int bfAggregationStart;
+ private int bfAggregationLength;
+ AtomicBoolean isMerging = new AtomicBoolean(false);
+
+ public BloomFilterMergeWorker(byte[] bfAggregation, int bfAggregationStart,
+ int bfAggregationLength) {
+ this.bfAggregation = bfAggregation;
+ this.bfAggregationStart = bfAggregationStart;
+ this.bfAggregationLength = bfAggregationLength;
+ this.queue = new ArrayBlockingQueue<>(VectorizedRowBatch.DEFAULT_SIZE *
2);
+ }
+
+ public void add(ElementWrapper wrapper) {
+ queue.add(wrapper);
+ }
+
+ @Override
+ public void run() {
+ while (true) {
+ ElementWrapper currentBf = null;
+ try {
+ currentBf = queue.take();
+ // at this point we have a currentBf wrapper which contains the
whole byte[] of the
+ // serialized bloomfilter, but we only want to merge a modified
"start -> start+length"
+ // part of it, which is pointed by modifiedStart/modifiedLength
fields by ElementWrapper
+ merge(currentBf);
+ } catch (InterruptedException e) {// Executor.shutdownNow() is called
+ if (!queue.isEmpty()){
+ LOG.debug(
+ "bloom filter merge was interrupted while processing and queue
is still not empty"
+ + ", this is fine in case of shutdownNow");
+ }
+ while (!queue.isEmpty()) { // time to finish work if any
Review comment:
I'm implementing abort in the next patch, I agree it's worth taking care
of this
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 470123)
Time Spent: 7h (was: 6h 50m)
> Bloom filters can be merged in a parallel way in VectorUDAFBloomFilterMerge
> ---------------------------------------------------------------------------
>
> Key: HIVE-23880
> URL: https://issues.apache.org/jira/browse/HIVE-23880
> Project: Hive
> Issue Type: Improvement
> Reporter: László Bodor
> Assignee: László Bodor
> Priority: Major
> Labels: pull-request-available
> Attachments: lipwig-output3605036885489193068.svg
>
> Time Spent: 7h
> Remaining Estimate: 0h
>
> Merging bloom filters in semijoin reduction can become the main bottleneck in
> case of large number of source mapper tasks (~1000, Map 1 in below example)
> and a large amount of expected entries (50M) in bloom filters.
> For example in TPCDS Q93:
> {code}
> select /*+ semi(store_returns, sr_item_sk, store_sales, 70000000)*/
> ss_customer_sk
> ,sum(act_sales) sumsales
> from (select ss_item_sk
> ,ss_ticket_number
> ,ss_customer_sk
> ,case when sr_return_quantity is not null then
> (ss_quantity-sr_return_quantity)*ss_sales_price
> else
> (ss_quantity*ss_sales_price) end act_sales
> from store_sales left outer join store_returns on (sr_item_sk =
> ss_item_sk
> and
> sr_ticket_number = ss_ticket_number)
> ,reason
> where sr_reason_sk = r_reason_sk
> and r_reason_desc = 'reason 66') t
> group by ss_customer_sk
> order by sumsales, ss_customer_sk
> limit 100;
> {code}
> On 10TB-30TB scale there is a chance that from 3-4 mins of query runtime 1-2
> mins are spent with merging bloom filters (Reducer 2), as in:
> [^lipwig-output3605036885489193068.svg]
> {code}
> ----------------------------------------------------------------------------------------------
> VERTICES MODE STATUS TOTAL COMPLETED RUNNING PENDING
> FAILED KILLED
> ----------------------------------------------------------------------------------------------
> Map 3 .......... llap SUCCEEDED 1 1 0 0
> 0 0
> Map 1 .......... llap SUCCEEDED 1263 1263 0 0
> 0 0
> Reducer 2 llap RUNNING 1 0 1 0
> 0 0
> Map 4 llap RUNNING 6154 0 207 5947
> 0 0
> Reducer 5 llap INITED 43 0 0 43
> 0 0
> Reducer 6 llap INITED 1 0 0 1
> 0 0
> ----------------------------------------------------------------------------------------------
> VERTICES: 02/06 [====>>----------------------] 16% ELAPSED TIME: 149.98 s
> ----------------------------------------------------------------------------------------------
> {code}
> For example, 70M entries in bloom filter leads to a 436 465 696 bits, so
> merging 1263 bloom filters means running ~ 1263 * 436 465 696 bitwise OR
> operation, which is very hot codepath, but can be parallelized.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
