Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[VL][1.2] Port #6573 #7025 #7132 #7973

Merged
merged 3 commits into from
Nov 20, 2024
Merged

[VL][1.2] Port #6573 #7025 #7132 #7973

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
aa52e7c
[GLUTEN-7024][VL] Skip call collectMetrics when the task does not cal…
kecookier Aug 29, 2024
6771482
[GLUTEN-7130][CORE] Skip command execution when collect qe fallback s…
wForget Sep 5, 2024
fb8531f
[VL] Add config for show velox task metrics when finished (#6573)
Yohahaha Jul 25, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 10 additions & 3 deletions cpp/velox/compute/WholeStageResultIterator.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -311,15 +311,22 @@ void WholeStageResultIterator::collectMetrics() {
return;
}

if (veloxCfg_->get<bool>(kDebugModeEnabled, false)) {
auto planWithStats = velox::exec::printPlanWithStats(*veloxPlan_.get(), task_->taskStats(), true);
const auto& taskStats = task_->taskStats();
if (taskStats.executionStartTimeMs == 0) {
LOG(INFO) << "Skip collect task metrics since task did not call next().";
return;
}

if (veloxCfg_->get<bool>(kDebugModeEnabled, false) ||
veloxCfg_->get<bool>(kShowTaskMetricsWhenFinished, kShowTaskMetricsWhenFinishedDefault)) {
auto planWithStats = velox::exec::printPlanWithStats(*veloxPlan_.get(), taskStats, true);
std::ostringstream oss;
oss << "Native Plan with stats for: " << taskInfo_;
oss << "\n" << planWithStats << std::endl;
LOG(INFO) << oss.str();
}

auto planStats = velox::exec::toPlanStats(task_->taskStats());
auto planStats = velox::exec::toPlanStats(taskStats);
// Calculate the total number of metrics.
int statsNum = 0;
for (int idx = 0; idx < orderedNodeIds_.size(); idx++) {
Expand Down
4 changes: 3 additions & 1 deletion cpp/velox/compute/WholeStageResultIterator.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,9 @@ class WholeStageResultIterator : public ColumnarBatchIterator {

Metrics* getMetrics(int64_t exportNanos) {
collectMetrics();
metrics_->veloxToArrow = exportNanos;
if (metrics_) {
metrics_->veloxToArrow = exportNanos;
}
return metrics_.get();
}

Expand Down
3 changes: 3 additions & 0 deletions cpp/velox/config/VeloxConfig.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,6 +59,9 @@ const std::string kBloomFilterNumBits = "spark.gluten.sql.columnar.backend.velox
const std::string kBloomFilterMaxNumBits = "spark.gluten.sql.columnar.backend.velox.bloomFilter.maxNumBits";
const std::string kVeloxSplitPreloadPerDriver = "spark.gluten.sql.columnar.backend.velox.SplitPreloadPerDriver";

const std::string kShowTaskMetricsWhenFinished = "spark.gluten.sql.columnar.backend.velox.showTaskMetricsWhenFinished";
const bool kShowTaskMetricsWhenFinishedDefault = false;

const std::string kEnableUserExceptionStacktrace =
"spark.gluten.sql.columnar.backend.velox.enableUserExceptionStacktrace";
const bool kEnableUserExceptionStacktraceDefault = true;
Expand Down
2 changes: 1 addition & 1 deletion gluten-core/src/main/scala/org/apache/spark/sql/execution/GlutenImplicits.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -176,7 +176,7 @@ object GlutenImplicits {
withSQLConf(SQLConf.ADAPTIVE_EXECUTION_ENABLED.key -> "false") {
// AQE is not materialized, so the columnar rules are not applied.
// For this case, We apply columnar rules manually with disable AQE.
val qe = spark.sessionState.executePlan(logicalPlan)
val qe = spark.sessionState.executePlan(logicalPlan, CommandExecutionMode.SKIP)
processPlan(qe.executedPlan, concat.append, collectFallbackFunc)
}
} else {
Expand Down
7 changes: 7 additions & 0 deletions shims/common/src/main/scala/org/apache/gluten/GlutenConfig.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1685,6 +1685,13 @@ object GlutenConfig {
.booleanConf
.createWithDefault(true)

val COLUMNAR_VELOX_SHOW_TASK_METRICS_WHEN_FINISHED =
buildConf("spark.gluten.sql.columnar.backend.velox.showTaskMetricsWhenFinished")
.internal()
.doc("Show velox full task metrics when finished.")
.booleanConf
.createWithDefault(false)

val COLUMNAR_VELOX_MEMORY_USE_HUGE_PAGES =
buildConf("spark.gluten.sql.columnar.backend.velox.memoryUseHugePages")
.internal()
Expand Down
Loading