Objet : Developers list for StarPU
Archives de la liste
- From: Mirko Myllykoski <mirkom@cs.umu.se>
- To: Starpu Devel <starpu-devel@lists.gforge.inria.fr>
- Subject: [Starpu-devel] About concurrent parallel tasks and OpenMP
- Date: Tue, 28 Aug 2018 10:27:44 +0100
- Authentication-results: mail2-smtp-roc.national.inria.fr; spf=None smtp.pra=mirkom@cs.umu.se; spf=Pass smtp.mailfrom=mirkom@cs.umu.se; spf=None smtp.helo=postmaster@mail.cs.umu.se
- Ironport-phdr: 9a23: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
- List-archive: <http://lists.gforge.inria.fr/pipermail/starpu-devel/>
- List-id: "Developers list. For discussion of new features, code changes, etc." <starpu-devel.lists.gforge.inria.fr>
Hi,
I am slightly confused about the following statement in the StarPU handbook: “For instance, most OpenMP implementations (including the main ones) do not support concurrent pragma omp parallel statements without nesting them in another pragma omp parallel statement, but StarPU does not yet support creating its CPU workers by using such pragma.”
My interpretation is that a situation where two pthreads (StarPU worker threads) encounter parallel pragmas concurrently, then something goes wrong (thread safety? thread binding?). My understanding that the “clustering” is a some type of solution to this problem. Is this correct? Can I have concurrent OpenMP-based parallel tasks with clusters? If so, how does clustering fix the aforementioned problem? My initial guess was that clustering creates the worker threads using a parallel pragma but that appears not to be the case.
Best Regards,
Mirko Myllykoski
- [Starpu-devel] About concurrent parallel tasks and OpenMP, Mirko Myllykoski, 28/08/2018
- Re: [Starpu-devel] About concurrent parallel tasks and OpenMP, Samuel Thibault, 28/08/2018
- [Starpu-devel] question on CPU/GPU tasks distribution, philippe.helluy@unistra.fr, 31/08/2018
- Re: [Starpu-devel] question on CPU/GPU tasks distribution, Samuel Thibault, 31/08/2018
- Re: [Starpu-devel] question on CPU/GPU tasks distribution, philippe.helluy@unistra.fr, 31/08/2018
- Re: [Starpu-devel] question on CPU/GPU tasks distribution, Samuel Thibault, 31/08/2018
- Re: [Starpu-devel] question on CPU/GPU tasks distribution, philippe.helluy@unistra.fr, 31/08/2018
- Re: [Starpu-devel] question on CPU/GPU tasks distribution, Samuel Thibault, 31/08/2018
- [Starpu-devel] question on CPU/GPU tasks distribution, philippe.helluy@unistra.fr, 31/08/2018
- Re: [Starpu-devel] About concurrent parallel tasks and OpenMP, Samuel Thibault, 28/08/2018
Archives gérées par MHonArc 2.6.19+.