2D Convolution example using global and shared memory #2228
+429
−0
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
2 times, most recently
from
9c27efd to
aa3efcc
Compare
mehmetyusufoglu
changed the title
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
3 times, most recently
from
dc160f3 to
2b93807
Compare
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
from
80dad53 to
d16a616
Compare
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
from
bd30c0e to
a05dabc
Compare
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
9 times, most recently
from
ae50779 to
a8ac8e6
Compare
psychocoderHPC
requested changes
Feb 5, 2024
| // Allocate shared memory | ||
| auto* const sharedN = alpaka::getDynSharedMem<TElem>(acc); | ||
| // Fill shared memory of device so that tile items are accessed from shared memory | ||
| if(row < matrixHeight && col < matrixWidth && blockThreadIdx1D < blockThreadExtent.prod()) |
There was a problem hiding this comment.
On the host side you use getValidWorkDiv This means you will have one thread for some alpaka accelerators.
I know you wrote that the block size must be equal to the tile size but you do not enforce it e.g. with an ALPAKA_VERIFY
If you have only one thread in the block you can simply iterate over the shared memory to fill it.
There was a problem hiding this comment.
Yes but if there is one thread per block it means it is not a GPU (or it is not good to use GPU); so we dont know which level of memory is used ?
There was a problem hiding this comment.
If there is one block, whole block is loaded into the shared memory in the code.
psychocoderHPC
requested changes
Feb 21, 2024
psychocoderHPC
requested changes
Feb 21, 2024
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
from
91b7f51 to
4673202
Compare
mehmetyusufoglu
force-pushed
the
convolution2DExample
branch
from
4673202 to
06f9ace
Compare
psychocoderHPC
approved these changes
Feb 28, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
An example: A 2D Convolutional filter applied to a matrix. The values of filter-matrix were initially kept in constant memory at the first commit. But due to Gitlab pipeline error "The SYCL backend does not support global device constants"; in the second commit, constant memory usage has been removed.