Eventual-Inc
Daft
Blog
Docs
Changelog
Blog
Docs
Changelog
Overview
Branches
Benchmarks
Runs
Performance History
Latest Results
fix(flotilla): wire min_cpu_per_task into TaskResourceRequest The min_cpu_per_task execution config field had no readers in the distributed scheduler: TaskResourceRequest::num_cpus() returned a hardcoded 1.0 when the plan's ResourceRequest had no num_cpus. This was wired up in #4506 for the legacy ray runner only; #5375 removed that runner along with its lone reader, and the flotilla scheduler was never wired in. Result: setting min_cpu_per_task via daft.set_execution_config or DAFT_MIN_CPU_PER_TASK had zero effect on autoscaler bundle requests. Plumb the value from DaftExecutionConfig (already in scope at SwordfishTaskBuilder::build) into TaskResourceRequest, and use it as the fallback in num_cpus() instead of the literal 1.0. Closes #7123
XiaoHongbo-Hope:fix/min-cpu-per-task-wiring
6 minutes ago
Merge branch 'main' into fix-actor-udf-blocking-sink-deadlock
XiaoHongbo-Hope:fix-actor-udf-blocking-sink-deadlock
46 minutes ago
Merge branch 'main' into fix-actor-udf-blocking-sink-deadlock
XiaoHongbo-Hope:fix-actor-udf-blocking-sink-deadlock
3 hours ago
Merge branch 'main' into fix-actor-udf-blocking-sink-deadlock
XiaoHongbo-Hope:fix-actor-udf-blocking-sink-deadlock
4 hours ago
Merge branch 'main' into fix-actor-udf-blocking-sink-deadlock
XiaoHongbo-Hope:fix-actor-udf-blocking-sink-deadlock
5 hours ago
Merge branch 'main' into fix-actor-concurrency-hang
XiaoHongbo-Hope:fix-actor-concurrency-hang
6 hours ago
Merge remote-tracking branch 'origin/main' into perf/inline-agg-product # Conflicts: # src/daft-recordbatch/src/ops/inline_agg.rs
BABTUNA:perf/inline-agg-product
8 hours ago
Address review feedback for bround/greatest/least - numeric.py: use raw docstring for unhex() to satisfy ruff D301 (escape sequence \x0f in a regular docstring) - bround.rs: avoid multiplying by non-exact powers of 10 for negative precisions. precision >= 0 keeps multiply-then-divide; precision < 0 now divides-then-multiplies by 10^|p|, eliminating the second rounding step that biased ties (e.g. 250 with d=-2). Add regression test plus an Int32 decimals literal test for BRoundArgs<Option<i32>>. - greatest.rs: deduplicate Greatest/Least via a private GreatestLeastKind trait carrying KEEP_GREATER/NAME/DOCSTRING. Both ScalarUDF impls now share compare_inputs through impl_call::<Self>; public unit structs and typetag::serde tags are preserved for backwards compatibility.
XuQianJin-Stars:feat/spark-math-functions
9 hours ago
Latest Branches
CodSpeed Performance Gauge
0%
fix(flotilla): wire min_cpu_per_task into TaskResourceRequest
#7125
60 minutes ago
1d05985
XiaoHongbo-Hope:fix/min-cpu-per-task-wiring
CodSpeed Performance Gauge
0%
fix: actor UDF followed by BlockingSink deadlocks due to held channel sender
#7115
6 hours ago
879f206
XiaoHongbo-Hope:fix-actor-udf-blocking-sink-deadlock
CodSpeed Performance Gauge
0%
fix: detect unschedulable actor UDFs and warn on over-requested concurrency
#7113
6 hours ago
5d108c1
XiaoHongbo-Hope:fix-actor-concurrency-hang
© 2026 CodSpeed Technology
Home
Terms
Privacy
Docs