Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Remove CUDA < 9.2, add CUDA 11.4 #887

Merged
merged 4 commits into from
Oct 5, 2021
Merged

Remove CUDA < 9.2, add CUDA 11.4 #887

merged 4 commits into from
Oct 5, 2021

Conversation

upsj
Copy link
Member

@upsj upsj commented Sep 14, 2021

pretty self-explanatory :)

@upsj upsj added this to the Ginkgo 1.5.0 milestone Sep 14, 2021
@upsj upsj self-assigned this Sep 14, 2021
@upsj upsj requested a review from a team September 14, 2021 15:10
@ginkgo-bot ginkgo-bot added mod:all This touches all Ginkgo modules. reg:ci-cd This is related to the continuous integration system. reg:testing This is related to testing. type:solver This is related to the solvers labels Sep 14, 2021
@tcojean
Copy link
Member

tcojean commented Sep 15, 2021

Maybe we can also skip 10.1 or some other intermediate version and instead add 11.2 as well? In order to spread out somewhat evenly the CUDA versions which we actually test.

@upsj
Copy link
Member Author

upsj commented Sep 15, 2021

Looks like gtest is broken with GCC 11 (google/googletest#3514), nvcc doesn't play nice with clang-12 as claimed in the documentation, and let's not even talk about icpc :)

@upsj
Copy link
Member Author

upsj commented Sep 19, 2021

As @tcojean and I discussed, 16 GB for the intel-basekit isn't feasible, so we will most likely leave icpx out of this, since icpc has some issues again (https://community.intel.com/t5/Intel-C-Compiler/undefined-reference-to-std-failed-assertion-with-icpc-on-Fedora/td-p/1278857)

@Slaedr
Copy link
Contributor

Slaedr commented Sep 21, 2021

Is the server running out of storage? Perhaps we could even consider leaving icpc out and only testing with icpx, since the former now officially has a sort of maintenance-only status anyway.

@tcojean
Copy link
Member

tcojean commented Sep 21, 2021

It's not only for the server but also that our containers are uploaded to dockerhub.

Other arguments for not bothering with this for now:

  • We have 8 CUDA containers currently and 2 CPUs, which would potentially be 160 GB of Intel software. I think at most we should include Intel in one or two containers overall (since we already have the oneAPI container anyway).
  • The newer Intel compilers seem to be fairly fragile at the moment, and we are not in a hurry to combine DPC++ and CUDA backend compilation.

I'm personally fine if Intel isn't added here, it can be a separate pull request whenever things are more stable. Another option, is to add the Intel support to an older container if that fixes some issues.

@upsj
Copy link
Member Author

upsj commented Sep 21, 2021

clang-11 + libstdc++ (gcc-10) also doesn't work: https://gitlab.com/ginkgo-project/ginkgo-public-ci/-/jobs/1609337828, so I'll have to go down to gcc-9

@upsj
Copy link
Member Author

upsj commented Sep 21, 2021 •
edited

Are you freaking kidding me? Even nvcc 11.4 + clang 9 + libstdc++ gcc-9 is broken. I give up, no clang for you. https://gitlab.com/ginkgo-project/ginkgo-public-ci/-/jobs/1609871929
NVIDIA confirmed the issue (or a related issue) though, and are working on a fix

tcojean
tcojean approved these changes Oct 4, 2021
View reviewed changes
Copy link
Member

@tcojean tcojean left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

yhmtsai
yhmtsai approved these changes Oct 4, 2021
View reviewed changes
cuda/solver/common_trs_kernels.cuh
Comment on lines 61 to 63
struct SolveStruct {
virtual void dummy() {}
virtual ~SolveStruct() {}
};
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

might not need this anymore?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right, I think in this case we don't even need a vtable, but I wanted to use a more canonical way to make sure we still emit one.

@upsj
remove support for cuda < 9.2 and add 11.4 CI job
8403906
@upsj
fix compilation issues
580739a 
* CUDA 11.4 prints many spurious errors
* clang/icpc and CUDA 11.4 don't play nice together
* GCC 11 needs a workaround for GTest -Werror
@upsj
use published image
8f57642
@upsj upsj added the 1:ST:ready-to-merge This PR is ready to merge. label Oct 4, 2021
pratikvn
pratikvn reviewed Oct 4, 2021
View reviewed changes
cuda/base/cusparse_bindings.hpp
@@ -850,10 +850,6 @@ inline void destroy(cusparseSpMatDescr_t descr)
#endif


// CUDA versions 9.2 and above have csrsm2.
Copy link
Member

@pratikvn pratikvn Oct 4, 2021 •
edited

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So, we don't support CUDA <9.2 at all ? I thought this PR would stop using a CI job for <9.2 but we would still support <9.2 ? Maybe we can leave these #if defines, so if someone wants to compile with CUDA <9.2, they still can ?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That would very much complicate #764 (due to the support for multiple rhs/need for tranposition), so I would really prefer to just deprecate support for older versions. Compute capabilities down to 3.0 are still officially supported in CUDA 10, so we are not losing any hardware support by this.

Copy link
Member

@pratikvn pratikvn Oct 5, 2021 •
edited

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Okay, then that is fine with me. Please also update the README.md, INSTALL.md and other pages where we currently say we support CUDA 9.0+. I guess now it would be CUDA 10.0+ 9.2+

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

9.2 still works fine :)

@sonarcloud
Copy link

sonarcloud bot commented Oct 5, 2021

Kudos, SonarCloud Quality Gate passed!    Quality Gate passed

Bug A 0 Bugs
Vulnerability A 0 Vulnerabilities
Security Hotspot A 0 Security Hotspots
Code Smell A 0 Code Smells

No Coverage information No Coverage information
0.0% 0.0% Duplication

@upsj upsj merged commit b9013be into develop Oct 5, 2021
@upsj upsj deleted the cuda_114 branch October 5, 2021 18:09
tcojean added a commit that referenced this pull request Nov 12, 2022
@tcojean
Merge Release 1.5.0 to develop
7c9f530 
Advertise release 1.5.0 and last changes

+ Add changelog,
+ Update third party libraries
+ A small fix to a CMake file

See PR: #1195

The Ginkgo team is proud to announce the new Ginkgo minor release 1.5.0. This release brings many important new features such as:
- MPI-based multi-node support for all matrix formats and most solvers;
- full DPC++/SYCL support,
- functionality and interface for GPU-resident sparse direct solvers,
- an interface for wrapping solvers with scaling and reordering applied,
- a new algebraic Multigrid solver/preconditioner,
- improved mixed-precision support,
- support for device matrix assembly,

and much more.

If you face an issue, please first check our [known issues page](https://github.com/ginkgo-project/ginkgo/wiki/Known-Issues) and the [open issues list](https://github.com/ginkgo-project/ginkgo/issues) and if you do not find a solution, feel free to [open a new issue](https://github.com/ginkgo-project/ginkgo/issues/new/choose) or ask a question using the [github discussions](https://github.com/ginkgo-project/ginkgo/discussions).

Supported systems and requirements:
+ For all platforms, CMake 3.13+
+ C++14 compliant compiler
+ Linux and macOS
  + GCC: 5.5+
  + clang: 3.9+
  + Intel compiler: 2018+
  + Apple LLVM: 8.0+
  + NVHPC: 22.7+
  + Cray Compiler: 14.0.1+
  + CUDA module: CUDA 9.2+ or NVHPC 22.7+
  + HIP module: ROCm 4.0+
  + DPC++ module: Intel OneAPI 2021.3 with oneMKL and oneDPL. Set the CXX compiler to `dpcpp`.
+ Windows
  + MinGW and Cygwin: GCC 5.5+
  + Microsoft Visual Studio: VS 2019
  + CUDA module: CUDA 9.2+, Microsoft Visual Studio
  + OpenMP module: MinGW or Cygwin.


Algorithm and important feature additions:
+ Add MPI-based multi-node for all matrix formats and solvers (except GMRES and IDR). ([#676](#676), [#908](#908), [#909](#909), [#932](#932), [#951](#951), [#961](#961), [#971](#971), [#976](#976), [#985](#985), [#1007](#1007), [#1030](#1030), [#1054](#1054), [#1100](#1100), [#1148](#1148))
+ Porting the remaining algorithms (preconditioners like ISAI, Jacobi, Multigrid, ParILU(T) and ParIC(T)) to DPC++/SYCL, update to SYCL 2020, and improve support and performance ([#896](#896), [#924](#924), [#928](#928), [#929](#929), [#933](#933), [#943](#943), [#960](#960), [#1057](#1057), [#1110](#1110),  [#1142](#1142))
+ Add a Sparse Direct interface supporting GPU-resident numerical LU factorization, symbolic Cholesky factorization, improved triangular solvers, and more ([#957](#957), [#1058](#1058), [#1072](#1072), [#1082](#1082))
+ Add a ScaleReordered interface that can wrap solvers and automatically apply reorderings and scalings ([#1059](#1059))
+ Add a Multigrid solver and improve the aggregation based PGM coarsening scheme ([#542](#542), [#913](#913), [#980](#980), [#982](#982),  [#986](#986))
+ Add infrastructure for unified, lambda-based, backend agnostic, kernels and utilize it for some simple kernels ([#833](#833), [#910](#910), [#926](#926))
+ Merge different CUDA, HIP, DPC++ and OpenMP tests under a common interface ([#904](#904), [#973](#973), [#1044](#1044), [#1117](#1117))
+ Add a device_matrix_data type for device-side matrix assembly ([#886](#886), [#963](#963), [#965](#965))
+ Add support for mixed real/complex BLAS operations ([#864](#864))
+ Add a FFT LinOp for all but DPC++/SYCL ([#701](#701))
+ Add FBCSR support for NVIDIA and AMD GPUs and CPUs with OpenMP ([#775](#775))
+ Add CSR scaling ([#848](#848))
+ Add array::const_view and equivalent to create constant matrices from non-const data ([#890](#890))
+ Add a RowGatherer LinOp supporting mixed precision to gather dense matrix rows ([#901](#901))
+ Add mixed precision SparsityCsr SpMV support ([#970](#970))
+ Allow creating CSR submatrix including from (possibly discontinuous) index sets ([#885](#885), [#964](#964))
+ Add a scaled identity addition (M <- aI + bM) feature interface and impls for Csr and Dense ([#942](#942))


Deprecations and important changes:
+ Deprecate AmgxPgm in favor of the new Pgm name. ([#1149](#1149)).
+ Deprecate specialized residual norm classes in favor of a common `ResidualNorm` class ([#1101](#1101))
+ Deprecate CamelCase non-polymorphic types in favor of snake_case versions (like array, machine_topology, uninitialized_array, index_set) ([#1031](#1031), [#1052](#1052))
+ Bug fix: restrict gko::share to rvalue references (*possible interface break*) ([#1020](#1020))
+ Bug fix: when using cuSPARSE's triangular solvers, specifying the factory parameter `num_rhs` is now required when solving for more than one right-hand side, otherwise an exception is thrown ([#1184](#1184)).
+ Drop official support for old CUDA < 9.2 ([#887](#887))


Improved performance additions:
+ Reuse tmp storage in reductions in solvers and add a mutable workspace to all solvers ([#1013](#1013), [#1028](#1028))
+ Add HIP unsafe atomic option for AMD ([#1091](#1091))
+ Prefer vendor implementations for Dense dot, conj_dot and norm2 when available ([#967](#967)).
+ Tuned OpenMP SellP, COO, and ELL SpMV kernels for a small number of RHS ([#809](#809))


Fixes:
+ Fix various compilation warnings ([#1076](#1076), [#1183](#1183), [#1189](#1189))
+ Fix issues with hwloc-related tests ([#1074](#1074))
+ Fix include headers for GCC 12 ([#1071](#1071))
+ Fix for simple-solver-logging example ([#1066](#1066))
+ Fix for potential memory leak in Logger ([#1056](#1056))
+ Fix logging of mixin classes ([#1037](#1037))
+ Improve value semantics for LinOp types, like moved-from state in cross-executor copy/clones ([#753](#753))
+ Fix some matrix SpMV and conversion corner cases ([#905](#905), [#978](#978))
+ Fix uninitialized data ([#958](#958))
+ Fix CUDA version requirement for cusparseSpSM ([#953](#953))
+ Fix several issues within bash-script ([#1016](#1016))
+ Fixes for `NVHPC` compiler support ([#1194](#1194))


Other additions:
+ Simplify and properly name GMRES kernels ([#861](#861))
+ Improve pkg-config support for non-CMake libraries ([#923](#923), [#1109](#1109))
+ Improve gdb pretty printer ([#987](#987), [#1114](#1114))
+ Add a logger highlighting inefficient allocation and copy patterns ([#1035](#1035))
+ Improved and optimized test random matrix generation ([#954](#954), [#1032](#1032))
+ Better CSR strategy defaults ([#969](#969))
+ Add `move_from` to `PolymorphicObject` ([#997](#997))
+ Remove unnecessary device_guard usage ([#956](#956))
+ Improvements to the generic accessor for mixed-precision ([#727](#727))
+ Add a naive lower triangular solver implementation for CUDA ([#764](#764))
+ Add support for int64 indices from CUDA 11 onward with SpMV and SpGEMM ([#897](#897))
+ Add a L1 norm implementation ([#900](#900))
+ Add reduce_add for arrays ([#831](#831))
+ Add utility to simplify Dense View creation from an existing Dense vector ([#1136](#1136)).
+ Add a custom transpose implementation for Fbcsr and Csr transpose for unsupported vendor types ([#1123](#1123))
+ Make IDR random initilization deterministic ([#1116](#1116))
+ Move the algorithm choice for triangular solvers from Csr::strategy_type to a factory parameter ([#1088](#1088))
+ Update CUDA archCoresPerSM ([#1175](#1116))
+ Add kernels for Csr sparsity pattern lookup ([#994](#994))
+ Differentiate between structural and numerical zeros in Ell/Sellp ([#1027](#1027))
+ Add a binary IO format for matrix data ([#984](#984))
+ Add a tuple zip_iterator implementation ([#966](#966))
+ Simplify kernel stubs and declarations ([#888](#888))
+ Simplify GKO_REGISTER_OPERATION with lambdas ([#859](#859))
+ Simplify copy to device in tests and examples ([#863](#863))
+ More verbose output to array assertions ([#858](#858))
+ Allow parallel compilation for Jacobi kernels ([#871](#871))
+ Change clang-format pointer alignment to left ([#872](#872))
+ Various improvements and fixes to the benchmarking framework ([#750](#750), [#759](#759), [#870](#870), [#911](#911), [#1033](#1033), [#1137](#1137))
+ Various documentation improvements ([#892](#892), [#921](#921), [#950](#950), [#977](#977), [#1021](#1021), [#1068](#1068), [#1069](#1069), [#1080](#1080), [#1081](#1081), [#1108](#1108), [#1153](#1153), [#1154](#1154))
+ Various CI improvements ([#868](#868), [#874](#874), [#884](#884), [#889](#889), [#899](#899), [#903](#903),  [#922](#922), [#925](#925), [#930](#930), [#936](#936), [#937](#937), [#958](#958), [#882](#882), [#1011](#1011), [#1015](#1015), [#989](#989), [#1039](#1039), [#1042](#1042), [#1067](#1067), [#1073](#1073), [#1075](#1075), [#1083](#1083), [#1084](#1084), [#1085](#1085), [#1139](#1139), [#1178](#1178), [#1187](#1187))
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@pratikvn pratikvn pratikvn left review comments

@yhmtsai yhmtsai yhmtsai approved these changes

@tcojean tcojean tcojean approved these changes

Assignees

@upsj upsj

Labels
1:ST:ready-to-merge This PR is ready to merge. 1:ST:run-full-test mod:all This touches all Ginkgo modules. reg:ci-cd This is related to the continuous integration system. reg:testing This is related to testing. type:solver This is related to the solvers
Projects
None yet
Milestone
Ginkgo 1.5.0
Development

Successfully merging this pull request may close these issues.

None yet
6 participants
@upsj @tcojean @Slaedr @pratikvn @yhmtsai @ginkgo-bot