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 e733332184 [GLUTEN-6887][VL] Daily Update Velox Version (2026_05_18)
(#12102)
e733332184 is described below
commit e733332184fa2e0fe85f29c5384b8d5210ca972e
Author: Gluten Performance Bot
<[email protected]>
AuthorDate: Fri May 22 07:56:12 2026 +0100
[GLUTEN-6887][VL] Daily Update Velox Version (2026_05_18) (#12102)
* [GLUTEN-6887][VL] Daily Update Velox Version (dft-2026_05_18)
Upstream Velox's New Commits:
6af674c38 by Claude, feat(fuzzer): Add WindowNode-based alternate plan to
TopNRowNumberFuzzer (#17536)
bec8806bf by 1fanwang, fix(parquet): Include file column name in
schema-mismatch error (#16591) (#17353)
103b8c889 by Shaojie Li, fix(cudf): Show stats for adapter operators not in
plan tree (#17541)
81dff12ad by Suryadev Sahadevan Rajesh, refactor(encoding): Move encoding
selection files to `selection/` subfolder (#17540)
472d3196b by beliefer, fix: Pass request type to
SelectiveDecimalColumnReader (#17463)
fb0826ace by Masha Basmanova, docs: Add PR review scripts and style guide
(#17524)
d41a22889 by Xiao Du, refactor: Enforce callers to provide metadataIoStats
to TabletReader (#17534)
951677f88 by Shruti Shivakumar, refactor(cudf): Alphabetize objects in
Velox-cuDF CMake files (#17517)
1718279d7 by Krishna Pai, build(docker): Pin tzdata across all velox-dev
images (#17535)
2a5e3f905 by Shrinidhi Joshi, fix: Drop HashTable cache entry on builder
failure (#17527)
982637f85 by Masha Basmanova, feat: Support dialect-specific type coercion
(#17519)
ea683f05a by Chengcheng Jin, fix(cudf): Fix CudfSplitReader ReaderOptions
init (#17532)
8801a43ff by Scott Lee, Add reallocateBytes to MemoryAllocator to avoid
unnecessary memcpy (#17525)
a983d9b07 by Madhurendra Purbay, perf(simd): Add arch-aware boolean mask
helpers (#17257)
2cea959b7 by rexan, feat: Add support for TZDIR environment variable
(#15871)
02f22e8ac by Raymond Lin, feat: Add partition key propagation to
ScanBatchEvent callback (#17513)
53b6a8b03 by Huameng (Michael) Jiang, fix(tracer): Fix TableWrite trace
replayer to register connector and preserve serde parameters (#17526)
5680a3226 by Xiaoxuan Meng, Refactor ReaderOptions IoStatistics from raw
pointers to shared_ptr (#17518)
fd130f44a by mohsaka, feat: Add iceberg data file statistics (#17388)
805db6b72 by Pramod Satya, fix: Validate reduce_agg initial state (#17398)
6af81f0e7 by Shaojie Li, perf(hashtable): Add adaptive prefetch to hashRows
normalizedKey path (#17495)
78ebafc59 by Rui Mo, fix(spark): Remove TIMESTAMP_NTZ type (#17512)
6800d5ba1 by mohsaka, feat: Add FileMetadata return to Writer::close() and
introduce WriterConfig constants (#17509)
888bab856 by Orri Erling, feat(wave): Add durable kernel caching
infrastructure to wave/common (#17474)
d0a53253c by Facebook GitHub Bot, Re-sync with internal repository
Signed-off-by: glutenperfbot <[email protected]>
* Trigger CI
* Trigger PR CI
Signed-off-by: Reema Alzaid <[email protected]>
* Fix Velox API compatibility
* Trigger PR CI
Signed-off-by: Reema Alzaid <[email protected]>
* Trigger CI for IBM Velox 2026-05-18 fix
* Trigger CI for IBM Velox 2026-05-18 fix
* [VL][Iceberg] Propagate VELOX_ENABLE_PARQUET to Gluten backend
---------
Signed-off-by: glutenperfbot <[email protected]>
Signed-off-by: Reema Alzaid <[email protected]>
Co-authored-by: glutenperfbot <[email protected]>
Co-authored-by: Reema Alzaid <[email protected]>
---
cpp/velox/CMakeLists.txt | 12 ++++++++++++
cpp/velox/substrait/SubstraitToVeloxPlanValidator.cc | 3 ++-
cpp/velox/utils/ConfigExtractor.cc | 6 +++---
ep/build-velox/src/get-velox.sh | 4 ++--
4 files changed, 19 insertions(+), 6 deletions(-)
diff --git a/cpp/velox/CMakeLists.txt b/cpp/velox/CMakeLists.txt
index 532dd6e6a7..a8b0d668ca 100644
--- a/cpp/velox/CMakeLists.txt
+++ b/cpp/velox/CMakeLists.txt
@@ -239,6 +239,18 @@ endif()
add_library(velox SHARED ${VELOX_SRCS})
+# Keep ABI sensitive Velox feature macros in sync
+set(GLUTEN_VELOX_CMAKE_CACHE "${VELOX_BUILD_PATH}/CMakeCache.txt")
+set(GLUTEN_VELOX_ENABLE_PARQUET ${VELOX_ENABLE_PARQUET})
+if(NOT GLUTEN_VELOX_ENABLE_PARQUET AND EXISTS "${GLUTEN_VELOX_CMAKE_CACHE}")
+ file(STRINGS "${GLUTEN_VELOX_CMAKE_CACHE}" GLUTEN_VELOX_ENABLE_PARQUET
+ REGEX "^VELOX_ENABLE_PARQUET:BOOL=(ON|TRUE|1)$")
+endif()
+if(GLUTEN_VELOX_ENABLE_PARQUET)
+ target_compile_definitions(velox PRIVATE VELOX_ENABLE_PARQUET)
+ message(STATUS "Propagating VELOX_ENABLE_PARQUET to Gluten Velox backend")
+endif()
+
if(ENABLE_GLUTEN_VCPKG AND NOT CMAKE_SYSTEM_NAME MATCHES "Darwin")
# Hide some symbols to avoid conflict.
target_link_options(
diff --git a/cpp/velox/substrait/SubstraitToVeloxPlanValidator.cc
b/cpp/velox/substrait/SubstraitToVeloxPlanValidator.cc
index 6b10b805ed..3ea800b60c 100644
--- a/cpp/velox/substrait/SubstraitToVeloxPlanValidator.cc
+++ b/cpp/velox/substrait/SubstraitToVeloxPlanValidator.cc
@@ -25,6 +25,7 @@
#include "velox/exec/Aggregate.h"
#include "velox/expression/Expr.h"
#include "velox/expression/SignatureBinder.h"
+#include "velox/type/TypeCoercer.h"
namespace gluten {
namespace {
@@ -1159,7 +1160,7 @@ bool
SubstraitToVeloxPlanValidator::validateAggRelFunctionType(const ::substrait
bool resolved = false;
for (const auto& signature : signaturesOpt.value()) {
- exec::SignatureBinder binder(*signature, types);
+ exec::SignatureBinder binder(*signature, types,
facebook::velox::TypeCoercer::defaults());
if (binder.tryBind()) {
TypePtr resolveType = nullptr;
try {
diff --git a/cpp/velox/utils/ConfigExtractor.cc
b/cpp/velox/utils/ConfigExtractor.cc
index 6006d990db..b0b6c1a99a 100644
--- a/cpp/velox/utils/ConfigExtractor.cc
+++ b/cpp/velox/utils/ConfigExtractor.cc
@@ -25,7 +25,7 @@
#include "utils/Macros.h"
#include "velox/connectors/hive/HiveConfig.h"
#include "velox/connectors/hive/storage_adapters/s3fs/S3Config.h"
-#include "velox/dwio/parquet/writer/Writer.h"
+#include "velox/dwio/parquet/writer/WriterConfig.h"
namespace gluten {
@@ -229,7 +229,7 @@ std::shared_ptr<facebook::velox::config::ConfigBase>
createHiveConnectorSessionC
configs[facebook::velox::connector::hive::HiveConfig::kFileColumnNamesReadAsLowerCaseSession]
=
!conf->get<bool>(kCaseSensitive, false) ? "true" : "false";
configs[facebook::velox::connector::hive::HiveConfig::kPartitionPathAsLowerCaseSession]
= "false";
- configs[facebook::velox::parquet::WriterOptions::kParquetWriteTimestampUnit]
= std::string("6");
+
configs[facebook::velox::parquet::WriterConfig::kParquetSessionWriteTimestampUnit]
= std::string("6");
configs[facebook::velox::connector::hive::HiveConfig::kReadTimestampUnitSession]
= std::string("6");
configs[facebook::velox::connector::hive::HiveConfig::kMaxPartitionsPerWritersSession]
=
conf->get<std::string>(kMaxPartitions, "10000");
@@ -243,7 +243,7 @@ std::shared_ptr<facebook::velox::config::ConfigBase>
createHiveConnectorSessionC
conf->get<bool>(kAllowInt32Narrowing, true) ? "true" : "false";
configs[facebook::velox::connector::hive::HiveConfig::kOrcUseColumnNamesSession]
=
conf->get<bool>(kOrcUseColumnNames, true) ? "true" : "false";
- configs[facebook::velox::parquet::WriterOptions::kParquetWritePageSize] =
+
configs[facebook::velox::parquet::WriterConfig::kParquetSessionWritePageSize] =
conf->get<std::string>(kWriteParquetPageSizeBytes, "1MB");
overwriteVeloxConf(conf.get(), configs, kDynamicBackendConfPrefix);
diff --git a/ep/build-velox/src/get-velox.sh b/ep/build-velox/src/get-velox.sh
index a6631e1e76..51034c77f7 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_05_15
-VELOX_ENHANCED_BRANCH=ibm-2026_05_15
+VELOX_BRANCH=dft-2026_05_18
+VELOX_ENHANCED_BRANCH=ibm-2026_05_18
VELOX_HOME=""
RUN_SETUP_SCRIPT=ON
ENABLE_ENHANCED_FEATURES=OFF
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
