Repository navigation

All pull requests

WIP: Simplified memory reservation telemetry
#1320TomAugspurger opened Nov 9, 2026Contributor
7
Use modularized CCCL headers
#1319miscco opened Nov 9, 2026
3
Explicit lifetime for the streaming collectives
#1316madsbk opened Nov 5, 2026Member
5
Fix signal handling in rrun
#1314madsbk opened Nov 5, 2026Member
3
Fix SocketServer destructor hang when a rank is still waiting
#1313madsbk opened Nov 5, 2026Member
7
Handle host NUMA nodes without local visible GPUs
#13080z5a opened Nov 4, 2026Contributor
14
[C++] MemoryReserveOrWait waits on spills for any addressable memory type
#1306nirandaperera opened Nov 3, 2026Contributor
1
feat: add Python 3.15 to testing matrix
#1305gforsyth opened Nov 3, 2026Contributor
1
Test on WSL in CI
#1304bdice opened Nov 3, 2026Contributor
13
[C++] Spill containers from any addressable memory type and make headroom per type
#1298nirandaperera opened Nov 2, 2026Contributor
32
Report progress function errors to an error callback
#1260pentschev opened Oct 14, 2026Member
2
Never let exceptions escape PausableThreadLoop
#1257pentschev opened Oct 13, 2026Member
16
[C++] Fix get_current_numa_nodes() for non-MPOL_BIND memory policies
#1256nirandaperera opened Oct 9, 2026Contributor
11
Evaluate per-rank/per-partition send ordering in the shuffler
#1247nirandaperera opened Oct 4, 2026Contributor
1
Use CCCL memory resource aliases directly
#1238bdice opened Sep 29, 2026Contributor
3
Break UCXX communicator callback ownership cycles
#1235pentschev opened Sep 28, 2026Member
2
Include rmm/cuda_stream_view.hpp in the CUDA stream tests.
#1218miscco opened Sep 15, 2026
2
Add a pre-partitioned shuffle benchmark
#1197nirandaperera opened Sep 3, 2026Contributor
2
Adding reservation aware resource adapter
#1142nirandaperera opened Jul 25, 2026Contributor
10
[DO NO MERGE] rework bulk-ray shuffle example
#1112quasiben opened Jun 25, 2026Member
Wait for UCXX listener addresses before distribution
#1103fallintoplace opened Jun 15, 2026Contributor
9
Use std::span for cudf::split/cudf::concatenate calls
#1077joannamooon opened Jun 1, 2026
3
Adding coderabbit artifacts
#1052nirandaperera opened May 18, 2026Contributor
9
Enabling concurrent spilling
#1048nirandaperera opened May 16, 2026Contributor
7
Allow shuffle extraction before protocol drain
#1019pentschev opened May 8, 2026Member
1

Navigation Menu

Repository navigation

All pull requests

WIP: Simplified memory reservation telemetry

Use modularized CCCL headers

Explicit lifetime for the streaming collectives

Fix signal handling in `rrun`

Fix SocketServer destructor hang when a rank is still waiting

Handle host NUMA nodes without local visible GPUs

[C++] MemoryReserveOrWait waits on spills for any addressable memory type

feat: add Python 3.15 to testing matrix

Test on WSL in CI

[C++] Spill containers from any addressable memory type and make headroom per type

Report progress function errors to an error callback

Never let exceptions escape `PausableThreadLoop`

[C++] Fix get_current_numa_nodes() for non-MPOL_BIND memory policies

Evaluate per-rank/per-partition send ordering in the shuffler

Use CCCL memory resource aliases directly

Break UCXX communicator callback ownership cycles

Include rmm/cuda_stream_view.hpp in the CUDA stream tests.

Add a pre-partitioned shuffle benchmark

Adding reservation aware resource adapter

[DO NO MERGE] rework bulk-ray shuffle example

Wait for UCXX listener addresses before distribution

Use std::span for cudf::split/cudf::concatenate calls

Adding coderabbit artifacts

Enabling concurrent spilling

Allow shuffle extraction before protocol drain

Pull requests