Skip to content

Conversation

@edgargabriel
Copy link
Member

This pr brings in support for external32 data representation in ompio. While there is still room for improvements, it is tested and proven to work for blocking and non-blocking individual I/O, shared file pointers, and blocking collective I/O operations. Non-blocking collective I/O operations (especially read operations) are a bit of a problem, but will be looking later into this.

in preparation for adding support for the external32 data
representation.

Signed-off-by: Edgar Gabriel <[email protected]>
the infrastructure put in place to manage cuda buffers is actually
a lot more generic than just for cuda buffers. Specifically, we ca
reuse much of the code to implement the external32 data representation.
This commit converts the code from common_ompio_cuda* to
common_ompio_buffer*. There are just very few places where we actually need to keep the OPAL_CUDA_SUPPORT ifdef in place.

Signed-off-by: Edgar Gabriel <[email protected]>
introduce separate convertors for memory vs. file representation. Adjust the interfaces for decode_datatype to provide the convertor to be used for that.

Signed-off-by: Edgar Gabriel <[email protected]>
both blocking and non-blocking. collective write and read operations not yet.

Signed-off-by: Edgar Gabriel <[email protected]>
external32 data representation is now support by ompio for everything
but non-blocking collective I/O operations. The support can further be improved
in a second step to limit the temporary buffer size (at least for blocking operations),
but it does work now for many scenarios.

Signed-off-by: Edgar Gabriel <[email protected]>
@edgargabriel edgargabriel added this to the v5.0.0 milestone May 20, 2019
@edgargabriel edgargabriel merged commit 88b3b24 into open-mpi:master May 21, 2019
@edgargabriel edgargabriel deleted the topic/external32-v2 branch March 3, 2023 14:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant