GlutenPerfBot opened a new pull request, #12560:
URL: https://github.com/apache/gluten/pull/12560

   Upstream Velox's New Commits:
   
   ```
   b7cded725 by Shruti Shivakumar, fix(cudf): Return NULL instead of NaN for 
avg of empty groups
   0b1741281 by Krishna Pai, fix(ci): Use Velox-owned fuzzer images
   f4042228a by Karthikeyan, fix(dwio): Fix dictionary filter cache-mask 
extraction on ARM NEON
   43a1fb41b by zhichenxu-meta, feat(rpc): Add kInvalidRequest error kind for 
non-retryable failures (18095)
   94d87c285 by Bradley Dice, build: Upgrade bundled gflags to 2.3.0
   77d1607a0 by Krishnakanth, fix(core): Preserve markerName in 
UnnestNode::Builder
   adfc53cae by Krishnakanth, fix(core): Validate UnnestNode unnestNames count
   e567359ec by Arpit Porwal, refactor: Allow registering collect_list under 
arbitrary names (18132)
   ad186be92 by BRIJ RAJ KISHORE, feat(sparksql): Make unaryminus 
ANSI-compliant for integer overflow detection (18096)
   56a570229 by Sreeni Viswanadha, fix: Report the actual JSON element type in 
JSON cast error messages (18017)
   3637c0cf9 by James Lamb, build(ci): Update to zizmor 1.26.1
   6609329d8 by Deepak Majeti, fix(stats): Add more runtime metrics to 
aggregate per operator (18022)
   53a990726 by Zhenyuan Zhao, fix(dwrf): Bound PATCHED_BASE gap chain to avoid 
OOB read (18106)
   e1dd48045 by Hongze Zhang, refactor: Split `dwio/common/Statistics.h` to 
`.h` and `.cpp`
   5219cafbd by oerling, feat(torchwave): Free intermediate tensors at last use 
(18119)
   d8ceba11c by Wechar Yu, fix(spark): Support VARBINARY type in to_json 
function
   1861f170f by Krishna Pai, fix: Revert folly double/float formatting goldens 
(18116)
   ea4ae633b by Apurva Kumar, feat: Replace (union) an existing deletion vector 
on repeated mutation (18117)
   a3c8b8657 by Zhenyuan Zhao, fix: Fix test failure (18103)
   949c7309b by Deepak Majeti, feat(parquet): Add support for Timestamp type 
rowgroup filtering
   0b01ba0ad by Deepak Majeti, fix(build): FBThrift warnings for parquet.thrift
   1e7952ade by Apurva Kumar, test: NIMBLE Iceberg field-id end-to-end 
write+read test (18056)
   332b91f61 by PRASHANT GOLASH, perf: Use DirectBufferedInput for Nimble 
(enable loadQuantum) (17948)
   db15b91d3 by Liangcai Li, feat(cudf): Add Spark substring support
   fe6aae45d by Reema, fix(cudf): Fix multi column `hash_with_seed` on GPU
   049fe62d6 by Philo He, fix: Enable ANSI-compliant cast from VARCHAR to 
DECIMAL
   c59294974 by Amit Dutta, fix(window): Fail gracefully on oversized window 
partitions (18102)
   efb391ea4 by Richard Barnes, Update test goldens for folly double/float 
string formatting change
   d12dbd60e by jkhaliqi, feat: Add CompressionKind_LZ4_HADOOP
   7f6858340 by Daniel Bauer, fix(cudf): Support ms/s timestamp scalars on GPU
   4991078e0 by Ping Liu, docs: Scope BMI/BMI2/F16C requirements to x86
   4920e8bf3 by Ismaël Mejía, perf(dwio): Cross-platform bit-unpacking and 
level conversion
   4cf8f85db by Ismaël Mejía, fix(hive): Wire 
file-handle-expiration-duration-ms to SimpleLRUCache (18072)
   c1287a247 by meta-codesync[bot], misc: Add Thrift AllowLegacyMissingUris 
annotation to Velox files (18012)
   c759d48ad by Raymond Lin, fix(dwio): Correct flat-map prefetch eligibility 
(18038)
   039614eda by Ismaël Mejía, fix(dwio): Advance pointers in unpackNaive after 
decoding loop
   a8c470526 by Durvesh Pilankar, docs: Fix duplicated words and typos
   affccb3dc by Kyle Hubert, fix(ucx): Preserve shared exchange client on 
operator close (17752)
   ```
   
   velox_branch: https://github.com/IBM/velox/commits/dft-2026_07_19
   


-- 
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]

Reply via email to