mmtk/mmtk-core

Issues

• High resident set size (RSS)

  #1220 opened 2 months ago by wks
  6

• Rewrite BlockPool using double-checked locking pattern

  #1234 opened 2 months ago by wks
  0

• Introduce WorkerLocal

  #1233 opened 2 months ago by wks
  0

• Rethink about whether creating Address from usize should be unsafe

  #1227 opened 2 months ago by wks
  4

• Replace InitializedOnce with OnceLock

  #1231 opened 2 months ago by wks
  0

• Introduce NonZeroAddress and write more idiomatic Rust code

  #1223 opened 2 months ago by wks
  2

• Show logs in release build, and stop disabling logs by level.

  #1222 opened 2 months ago by wks
  2

• Regression tests cannot run some benchmarks in DaCapo 2006 on OpenJDK

  #1219 opened 2 months ago by wks
  0

• Forwarding bits are set to `FORWARDED` before VO bits are set.

  #1213 opened 3 months ago by wks
  1

• Performance regression on 20240802 for luindex

  #1200 opened 3 months ago by qinsoon
  13

• Return whether a GC was ran or not for `handle_user_collection_request`

  #1204 opened 4 months ago by k-sareen
  0

• Immix performs faster if defrag is enabled for every GC?

  #1203 opened 4 months ago by qinsoon
  0

• Stress test for `ImmixAllocator` should not change limits of the allocator

  #1201 opened 4 months ago by k-sareen
  1

• `ChunkMap` is broken when using multiple spaces that access it

  #1197 opened 4 months ago by k-sareen
  1

• Require ObjectReference to be inside an object

  #1170 opened 4 months ago by wks
  0

• A more optimal implementation for finding base references

  #1193 opened 4 months ago by qinsoon
  0

• Don't abuse the file name `global.rs`

  #1190 opened 4 months ago by wks
  2

• Expose whether _default_ space for Plan can pin objects or not

  #1187 opened 5 months ago by k-sareen
  2

• Precise heap traversal API

  #1186 opened 5 months ago by qinsoon
  2

• Immix has larger minheap than MarkCompact, LXR, and G1 on h2

  #1183 opened 5 months ago by wenyuzhao
  0

• Emit stat counters as tracepoints

  #1179 opened 5 months ago by caizixian
  0

• Out-of-memory handling in NoGC

  #1177 opened 6 months ago by wks
  1

• Better support for optional spaces

  #1173 opened 6 months ago by k-sareen
  2

• Stale forwarding bits in nursery GC in StickyImmix

  #1118 opened 6 months ago by wks
  4

• Public API check is not always correct

  #1167 opened 6 months ago by qinsoon
  1

• Guarantee old objects don't move, or add a query API for it

  #1163 opened 6 months ago by wks
  12

• Mark sweep free list allocates high address cells first

  #1164 opened 6 months ago by qinsoon
  0

• OS scheduling and designated work

  #1160 opened 6 months ago by wks
  1

• Native MarkSweep: bad load balancing with single-threaded workloads

  #1146 opened 6 months ago by wks
  2

• No need to atomically set VO bit in post_alloc?

  #1156 opened 7 months ago by qinsoon
  1

• Expose APIs for bindings to check if (metadata) memory has been mapped or not

  #1152 opened 7 months ago by k-sareen
  2

• Use vector for BlockList in MarkSweepSpace

  #1153 opened 7 months ago by wks
  0

• MarkSweepSpace should use BlockPageResourcre

  #1145 opened 7 months ago by wks
  0

• Marksweep is slower in lusearch and xalan after #1112

  #1123 opened 7 months ago by qinsoon
  1

• PrepareChunkMap (mark-sweep) unnecessarily clearing mark bits of all chunks

  #1144 opened 7 months ago by wks
  1

• Do not do any significant amount of work in `Release`

  #1147 opened 7 months ago by wks
  1

• Prolific types

  #1143 opened 7 months ago by wks
  1

• `harness_begin` shouldn't trigger GC

  #1140 opened 8 months ago by wks
  1

• Custom tracing unit/packet

  #1137 opened 8 months ago by wks
  0

• Watch dog

  #1114 opened 9 months ago by wks
  2

• Marksweep is slower in most benchmarks after #953

  #1135 opened 8 months ago by qinsoon
  1

• API migration guide

  #1131 opened 8 months ago by qinsoon
  1

• Kill off-heap `ObjectReference`, or make a safer `to_address` method

  #1120 opened 9 months ago by wks
  21

• Run tests against M1 macOS runner

  #1115 opened 9 months ago by qinsoon
  0

• Improper but "silently succeeding" use of post write barrier

  #1129 opened 8 months ago by wks
  1

• ReferenceProcessor does not trace transitively discovered SoftReference instances

  #1125 opened 9 months ago by wks
  4

• Block cell size is sometimes zero in native MarkSweep

  #1103 opened 9 months ago by wks
  7

• Any plan to support C++?

  #1113 opened 9 months ago by harry75369
  7

• Support ASAN

  #1111 opened 9 months ago by qinsoon
  1

• Rustup warns against installing i686 toolchains on x86_64

  #1107 opened 9 months ago by wks
  0