This is an automated email from the ASF dual-hosted git repository.
zhouyuan pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/gluten.git
The following commit(s) were added to refs/heads/main by this push:
new cd22fb88f5 [GLUTEN-6887][VL] Daily Update Velox Version (2026_08_12)
(#12755)
cd22fb88f5 is described below
commit cd22fb88f521133570f3b20853ce77566788b171
Author: Gluten Performance Bot
<[email protected]>
AuthorDate: Thu Aug 13 20:59:00 2026 +0100
[GLUTEN-6887][VL] Daily Update Velox Version (2026_08_12) (#12755)
* [GLUTEN-6887][VL] Daily Update Velox Version (dft-2026_08_12)
Upstream Velox's New Commits:
55acc87ce by Ismaël Mejía, perf(parquet): Dictionary passthrough and
selective flattening in Parquet writer
8f2e1ae75 by PRASHANT GOLASH, feat(nimble): Add reader-owned lazy I/O load
hook (#17999)
844c77cc9 by oerling, feat(torchwave): Multikernel parity with cg (#18390)
f0d0557ed by Pedro Pedreira, fix(io_uring): Handle ENOMEM from the
availability probe
b3701be01 by wankun, fix: Prevent cancelled direct coalesced loads from
exposing partial data
1d8d1e291 by LingBin, refactor: Use CHECK_AND_INC_MEM_OP_STATS for external
alloc/free stats
d3d59c862 by carlnayak, feat: Add xxhash128 Presto scalar function (#18436)
fc99399aa by Guilherme Kunigami, feat(remote): Only send selected rows to
the remote server (#18379)
91b49e730 by Krishnakanth, feat: Support kAnti and kLeftSemiFilter in
NestedLoopJoin (#18161)
5017bfaf1 by Pedro Pedreira, fix(vector): Support OPAQUE vectors in
saveVector/restoreVector
0814ab6c8 by dependabot[bot], build(deps): Bump fast-uri from 3.1.0 to
3.1.5 in /website
1d6881b4e by RindsSchei225e, refactor(writer): Rename VeloxWriter to Writer
(#18431)
b1261a0e9 by Shrinidhi Joshi, feat(exchange): Add
TransportKind::kMaterialized and TransportOptions to PartitionedOutput (#18425)
36a0651a7 by Krishna Pai, fix(ci): Bump Ubuntu tzdata pin to the currently
published build (fixes #18440)
deca9543f by dependabot[bot], build(deps): Bump websocket-driver from 0.7.4
to 0.7.5 in /website
a0bf9327b by Andrii Rosa, fix: Always initialize shell for uv tool
351b0a72f by Sreeni Viswanadha, perf: Skip string conversion for
single-element array_join (#18361)
71cd8338c by Matt Gara, fix(cudf): Synchronize MarkDistinct state across
streams
b603b5b26 by Muhammad Haseeb, feat(cudf): Transform and push down filters
in cuDF Iceberg
27196fead by Matt Gara, fix(cudf): Preserve zero-column nested loop join
row counts
d4b5285d5 by Xiaoxuan, feat: Add shared dictionary builder and encoding
(#18418)
fd6fbc5e9 by Pedro Pedreira, fix(fuzzer): Cap KHyperLogLog generated sketch
size
a6b2a8f29 by Bikramjeet Singh Vig, fix(dwrf): Apply key filters to
preserved flat maps
92422486e by Minni Mittal, feat(spark): Add randn scalar function
77558daec by Raymond Lin, feat(hive): Add read bytes and format to scan
batch event (#18407)
6d0667cb7 by dependabot[bot], build(deps): Bump postcss from 8.5.12 to
8.5.25 in /website
568652c5f by Kunal Pathak, perf(vector): Accumulate consecutive copy ranges
in locals in copyRangesImpl
5cbf734d1 by LingBin, refactor(memory): Remove dangling comment in
SharedArbitrator.h
9f4cf9938 by Hongze Zhang, perf: Optimize probing misses using bloom filter
for kLeft / kLeftSemiProject / kAnti hash join
d3c2c8cd6 by LingBin, docs(memory): Fix and clarify ArbitrationOperation
timestamp comments
a34b683f4 by Natasha Sehgal, feat(aggregates): VARBINARY compare in N-arg
min_by/max_by
1354a214b by Natasha Sehgal, test(fuzzer): Fuzz qdigest scalars over full
value range (#18412)
ef92cf1ba by zhichenxu-meta, fix(rpc): Park instead of busy-spinning under
mid-stream BATCH back-pressure (#18300)
2615dde9d by JiaKe, refactor(hashjoin): Refactor hashJoinTableType and
isLeftNullAwareJoinWithFilter API in HashJoinBridge
c1652ed62 by oerling, feat(torchwave): Fused in-place slice/select/scatter
ops (#18389)
908bf838e by abhinavmuk04, feat: Add Layer 2 AIMD-under-load test for RPC
framework (#18311)
Signed-off-by: glutenperfbot <[email protected]>
* exclude UT for now
Signed-off-by: Yuan <[email protected]>
* fix
Signed-off-by: Yuan <[email protected]>
---------
Signed-off-by: glutenperfbot <[email protected]>
Signed-off-by: Yuan <[email protected]>
Co-authored-by: glutenperfbot <[email protected]>
Co-authored-by: Yuan <[email protected]>
---
ep/build-velox/src/get-velox.sh | 4 ++--
.../scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala | 4 ++++
.../scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala | 8 ++++++++
.../scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala | 8 ++++++++
4 files changed, 22 insertions(+), 2 deletions(-)
diff --git a/ep/build-velox/src/get-velox.sh b/ep/build-velox/src/get-velox.sh
index 5df0b0c28a..477b75cf30 100755
--- a/ep/build-velox/src/get-velox.sh
+++ b/ep/build-velox/src/get-velox.sh
@@ -18,8 +18,8 @@ set -exu
CURRENT_DIR=$(cd "$(dirname "$BASH_SOURCE")"; pwd)
VELOX_REPO=https://github.com/IBM/velox.git
-VELOX_BRANCH=dft-2026_08_05
-VELOX_ENHANCED_BRANCH=ibm-2026_08_05
+VELOX_BRANCH=dft-2026_08_12
+VELOX_ENHANCED_BRANCH=ibm-2026_08_12
VELOX_HOME=""
RUN_SETUP_SCRIPT=ON
ENABLE_ENHANCED_FEATURES=OFF
diff --git
a/gluten-ut/spark35/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
b/gluten-ut/spark35/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
index e7a8c93ba8..22c72ba392 100644
---
a/gluten-ut/spark35/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
+++
b/gluten-ut/spark35/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
@@ -628,6 +628,8 @@ class VeloxTestSettings extends BackendTestSettings {
.exclude("length check for input string values: nested in map value")
.exclude("length check for input string values: nested in both map key and
value")
.exclude("length check for input string values: nested in array of struct")
+ .excludeGlutenTest("length check for input string values: nested in array
of struct")
+ .exclude("char type values should be padded: nested in array of struct")
.exclude("length check for input string values: nested in array of array")
// Following tests are excluded as these are overridden in Gluten test
suite..
// The overridden tests assert against Velox-specific error messages for
char/varchar
@@ -645,6 +647,8 @@ class VeloxTestSettings extends BackendTestSettings {
.exclude("length check for input string values: nested in array")
.exclude("length check for input string values: nested in struct of array")
.exclude("length check for input string values: nested in array of struct")
+ .excludeGlutenTest("length check for input string values: nested in array
of struct")
+ .exclude("char type values should be padded: nested in array of struct")
.exclude("length check for input string values: nested in array of array")
.exclude("length check for input string values: with implicit cast")
.exclude("char/varchar type values length check: partitioned columns of
other types")
diff --git
a/gluten-ut/spark40/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
b/gluten-ut/spark40/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
index f12483655b..2e068639f9 100644
---
a/gluten-ut/spark40/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
+++
b/gluten-ut/spark40/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
@@ -949,7 +949,15 @@ class VeloxTestSettings extends BackendTestSettings {
// Rewrite for different cache size.
.exclude("SPARK-36120: Support cache/uncache table with TimestampNTZ type")
enableSuite[GlutenFileSourceCharVarcharTestSuite]
+ // Velox Parquet writer (post dft-2026_08_12) introduced selective
per-column flattening
+ // that misses CONSTANT-encoded descendants inside nested complex types
(ARRAY<STRUCT>).
+ // Writes to ARRAY<STRUCT<char/varchar>> crash with VeloxRuntimeError
INVALID_STATE in the
+ // Arrow bridge instead of producing the expected error or result.
+ .exclude("length check for input string values: nested in array of struct")
+ .exclude("char type values should be padded: nested in array of struct")
enableSuite[GlutenDSV2CharVarcharTestSuite]
+ .exclude("length check for input string values: nested in array of struct")
+ .exclude("char type values should be padded: nested in array of struct")
enableSuite[GlutenColumnExpressionSuite]
// Velox raise_error('errMsg') throws a velox_user_error exception with
the message 'errMsg'.
// The final caught Spark exception's getCause().getMessage() contains
'errMsg' but does not
diff --git
a/gluten-ut/spark41/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
b/gluten-ut/spark41/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
index b196f7351f..e027321f18 100644
---
a/gluten-ut/spark41/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
+++
b/gluten-ut/spark41/src/test/scala/org/apache/gluten/utils/velox/VeloxTestSettings.scala
@@ -933,7 +933,15 @@ class VeloxTestSettings extends BackendTestSettings {
.exclude("SPARK-36120: Support cache/uncache table with TimestampNTZ type")
enableSuite[GlutenCacheTableInKryoSuite]
enableSuite[GlutenFileSourceCharVarcharTestSuite]
+ // Velox Parquet writer (post dft-2026_08_12) introduced selective
per-column flattening
+ // that misses CONSTANT-encoded descendants inside nested complex types
(ARRAY<STRUCT>).
+ // Writes to ARRAY<STRUCT<char/varchar>> crash with VeloxRuntimeError
INVALID_STATE in the
+ // Arrow bridge instead of producing the expected error or result.
+ .exclude("length check for input string values: nested in array of struct")
+ .exclude("char type values should be padded: nested in array of struct")
enableSuite[GlutenDSV2CharVarcharTestSuite]
+ .exclude("length check for input string values: nested in array of struct")
+ .exclude("char type values should be padded: nested in array of struct")
enableSuite[GlutenColumnExpressionSuite]
// Velox raise_error('errMsg') throws a velox_user_error exception with
the message 'errMsg'.
// The final caught Spark exception's getCause().getMessage() contains
'errMsg' but does not
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
