Releases: JuliaGPU/NVTX.jl
Releases · JuliaGPU/NVTX.jl
v0.3.5
v0.3.4
NVTX v0.3.4
Merged pull requests:
- Update to Documenter v1 (#40) (@carstenbauer)
- Bump actions/checkout from 3 to 4 (#41) (@dependabot[bot])
- Bump actions/upload-artifact from 3 to 4 (#42) (@dependabot[bot])
- Add hook for inference (#44) (@simonbyrne)
v0.3.3
NVTX v0.3.3
Merged pull requests:
- enable dependabot for GitHub actions (#35) (@ranocha)
- Bump actions/checkout from 2 to 3 (#36) (@dependabot[bot])
- Add expression by default to NVTX.range (#39) (@charleskawczynski)
Closed issues:
- Use expression as default for message (#38)
v0.3.2
NVTX v0.3.2
Closed issues:
- Naming threads: include MPI rank information (#6)
- support nsys launch (#8)
- Should NVTX.jl always be active? (#32)
Merged pull requests:
- set up docs (#30) (@simonbyrne)
- add category macro for defining categories (#31) (@simonbyrne)
- Add an API for activating NVTX from CUDA.jl. (#34) (@maleadt)
v0.3.1
NVTX v0.3.1
Closed issues:
Merged pull requests:
- don't introduce new scope with
@range
(#29) (@simonbyrne)
v0.3.0
NVTX v0.3.0
Closed issues:
- GC callbacks not working on Julia 1.9 (#16)
Merged pull requests:
- Use small C library for callbacks (#18) (@simonbyrne)
- upload artifacts (#19) (@simonbyrne)
- Include header in header (#20) (@simonbyrne)
- switch to JLLs (#21) (@simonbyrne)
- fix size field (#22) (@simonbyrne)
- add some better tests (#23) (@simonbyrne)
- enable options by default (#24) (@simonbyrne)
- Add annotate macro (#25) (@simonbyrne)
- update for callback library changes (#26) (@simonbyrne)
v0.2.0
NVTX v0.2.0
Closed issues:
- Convenience macros (as in CUDA.jl) (#5)
Merged pull requests:
- Create CI.yml (#1) (@simonbyrne)
- Make
NSYS_ACTIVE
reference aconst
(#2) (@charleskawczynski) - add hooks for GC (#4) (@simonbyrne)
- Add image to README (#7) (@carstenbauer)
- Better GC handling (#9) (@simonbyrne)
- callback on alloc and free (#10) (@simonbyrne)
- add basic macro (#11) (@simonbyrne)
- add more tests (#12) (@simonbyrne)
- add function to get default domain for a module (#13) (@simonbyrne)