Implemented multi-device sample#91
Conversation
|
|
Beanavil
force-pushed
the
multi-device-sample
branch
4 times, most recently
from
87a4154 to
419fe40
Compare
bashbaug
left a comment
bashbaug
left a comment
There was a problem hiding this comment.
Note: I only reviewed the last commit in this PR.
Beanavil
force-pushed
the
multi-device-sample
branch
from
11c0765 to
b818605
Compare
bashbaug
left a comment
bashbaug
left a comment
There was a problem hiding this comment.
Getting closer but a few issues remain, and there are a few new ones.
Beanavil
force-pushed
the
multi-device-sample
branch
from
b818605 to
b33c0e4
Compare
bashbaug
left a comment
bashbaug
left a comment
There was a problem hiding this comment.
Latest version LGTM, at least the last two commits (the ones for the multi-device sample specifically). Thanks!
You may want to keep just those two commits in this PR because the other infrastructure-related commits are still under review.
Beanavil
force-pushed
the
multi-device-sample
branch
from
b33c0e4 to
0babaf4
Compare
|
Rebased branch and removed unrelated commits. So I think this is ready to be merged :) |
|
Requested a review from @nikhiljnv, planning to merge tomorrow. |
nikhiljnv
left a comment
nikhiljnv
left a comment
There was a problem hiding this comment.
LGTM except a minor comment to fix the limit for cl_device_partition_property.
Beanavil
force-pushed
the
multi-device-sample
branch
from
0babaf4 to
b6475e4
Compare
|
Merging as agreed in the December 12th teleconference. |
5 participants
This PR adds the C/C++ samples showcasing a multi-device convolution computation that makes use of sub-devices and sub-buffers.
Note: only the last commit is relevant for the PR, all the other commits are from the StreamHPC:release-cd branch. After #88 is merged, this should be trivial to rebase/merge.