-
Notifications
You must be signed in to change notification settings - Fork 300
All issues
Issue creation is restricted in this repository
Issues
is:issue state:open
is:issue state:open
Search results
[FEA] Add phase attribution for threaded shuffle and asynchronous file reads
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestperformanceA performance related task/issueA performance related task/issueStatus: Open.#15865 In NVIDIA/cudf-spark;[FEA] Define a safe bounded lifecycle for progress-aware GCS hedged reads
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15864 In NVIDIA/cudf-spark;[FEA] Reduce GCS task tail latency from large single range requests
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15863 In NVIDIA/cudf-spark;[BUG] Large ORC host-stream reads can create oversized temporary JVM allocations
? - Needs TriageNeed team to review and classifyNeed team to review and classifybugSomething isn't workingSomething isn't workingperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15862 In NVIDIA/cudf-spark;[BUG] Empty shuffle partitions can produce inconsistent Kudo framing
? - Needs TriageNeed team to review and classifyNeed team to review and classifybugSomething isn't workingSomething isn't workingreliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15861 In NVIDIA/cudf-spark;[BUG] RoundRobin partitioning can produce incorrect Kudo split boundaries
? - Needs TriageNeed team to review and classifyNeed team to review and classifybugSomething isn't workingSomething isn't workingreliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15860 In NVIDIA/cudf-spark;[FEA] Bound executor-wide resource amplification from threaded shuffle readers
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15859 In NVIDIA/cudf-spark;[FEA] Balance CPU backlog and GPU pressure during shuffle compression
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15858 In NVIDIA/cudf-spark;[BUG] A single oversized shuffle record can exceed the writer memory budget
? - Needs TriageNeed team to review and classifyNeed team to review and classifybugSomething isn't workingSomething isn't workingreliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15857 In NVIDIA/cudf-spark;[BUG] Concurrent multithreaded shuffle writers can amplify executor heap usage
? - Needs TriageNeed team to review and classifyNeed team to review and classifybugSomething isn't workingSomething isn't workingperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15856 In NVIDIA/cudf-spark;[EPIC] Improve cuDF-Spark performance and reliability on GCP Dataproc Serverless
epicIssue that encompasses a significant feature or body of workIssue that encompasses a significant feature or body of workperformanceA performance related task/issueA performance related task/issuereliabilityFeatures to improve reliability or bugs that severly impact the reliability of the pluginFeatures to improve reliability or bugs that severly impact the reliability of the pluginStatus: Open.#15855 In NVIDIA/cudf-spark;[FEA] Follow-up: Re-evaluate AST JIT multi-output group limits
? - Needs TriageNeed team to review and classifyNeed team to review and classifyfeature requestNew feature or requestNew feature or requestStatus: Open.#15853 In NVIDIA/cudf-spark;