Update changelog for v5
This commit is contained in:
parent
01ef17e8b4
commit
165db1749c
11
CHANGELOG.md
11
CHANGELOG.md
@ -7,21 +7,26 @@ All notable changes to this project will be documented in this file.
|
|||||||
- Thrust managed memory.
|
- Thrust managed memory.
|
||||||
- HIP managed memory.
|
- HIP managed memory.
|
||||||
- New implementation using SYCL2020 USM (sycl2020-acc) and renamed original `sycl2020` to `sycl2020-acc`.
|
- New implementation using SYCL2020 USM (sycl2020-acc) and renamed original `sycl2020` to `sycl2020-acc`.
|
||||||
|
- New implementation in [Futhark](https://futhark-lang.org/)
|
||||||
- Data initialisation and read-back timing for all models, including Java, Scala, Julia, and Rust
|
- Data initialisation and read-back timing for all models, including Java, Scala, Julia, and Rust
|
||||||
- Add support for the latest Aparapi (3.0.0) and TornadoVM (0.15.x) for Java
|
- Add support for the latest Aparapi (3.0.0) and TornadoVM (0.15.x) for Java
|
||||||
|
- JuliaStream.jl published to registry (pending #113)
|
||||||
|
|
||||||
### Changed
|
### Changed
|
||||||
|
- Fix std-data/std-indices compatibility with oneDPL, NVHPC, and AdaptiveCpp (a.k.a. hipSYCL).
|
||||||
- RAJA CUDA CMake build issues resolved.
|
- RAJA CUDA CMake build issues resolved.
|
||||||
|
- Kokkos build updates (CXX version upgraded to C++17).
|
||||||
- Fix CUDA memory limit check.
|
- Fix CUDA memory limit check.
|
||||||
|
- Fix CUDA CMake options for `-DMEM` and `-DCMAKE_CUDA_FLAGS`.
|
||||||
- Use long double for `check_solution` in case of large problem size.
|
- Use long double for `check_solution` in case of large problem size.
|
||||||
- OneAPI DPCPP compiler is deprecated in favour of ICPX, so added new build option to SYCL 2020 version.
|
- OneAPI DPCPP compiler is deprecated in favour of ICPX, so added new build option to SYCL 2020 version.
|
||||||
- Updates to the HIP kernels and API usage.
|
- Updates to the HIP kernels and API usage.
|
||||||
- Number of thread-blocks in CUDA dot kernel implementation changed to 1024.
|
- Number of thread-blocks in CUDA dot kernel implementation changed to 1024.
|
||||||
- Fix compatibility of `sycl2020` (now `sycl2020-acc`) with hipSYCL.
|
- Fix compatibility of `sycl2020` (now `sycl2020-acc`) with AdaptiveCpp.
|
||||||
- Bumped Julia compat to 1.9
|
- Bumped Julia compat to 1.9
|
||||||
- Bumped Scala to 3.3.1
|
- Bumped Scala to 3.3.1
|
||||||
- Bumped Rust to 1.74.0-nightly (13e6f24b9 2023-09-23)
|
- Bumped Rust to 1.74.0-nightly (13e6f24b9 2023-09-23)
|
||||||
|
- Upgrade CI to Ubuntu 22.04
|
||||||
|
|
||||||
## [v4.0] - 2021-12-22
|
## [v4.0] - 2021-12-22
|
||||||
|
|
||||||
|
|||||||
Loading…
Reference in New Issue
Block a user