[OpenCL][Kernel] Add concat multi inputs kernel except channel is not aligned #6075
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
|
Thanks for your contribution! |
zhaoyang-star
added a commit
to zhaoyang-star/Paddle-Lite
that referenced
this pull request
Jun 13, 2021
zhaoyang-star
added a commit
to zhaoyang-star/Paddle-Lite
that referenced
this pull request
Jun 15, 2021
daming5432
pushed a commit
that referenced
this pull request
Jun 16, 2021
* [OpenCL] Fix select fp32 compile crash (#6006) * [Pass] Add opencl_kernel_place_correct_pass (#6037) * [OpenCL] Fix invalid arg size in instance_norm (#6064) * [OpenCL][Kernel] Add concat multi inputs kernel except channel is not aligned (#6075) * [OpenCL][Bugfix] Fix target choose in opencl_kernel_place_correct_pass (#6079) * [OpenCL] fix kernel select of concat (#6158) * [OpenCL] BindTargets KOpenCL for conv_conv_fuse_pass (#6125) * test=develop * [UTest] Loose abs_error for group_norm and instance_norm (#6188) * loose group_norm abs_err. test=develop
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Outdated suggestions cannot be applied.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.
【问题】
当大于 2 个输入且不对 channel 维度执行 concat 时，concat 执行流程为：
image2d->buffer, concat buffers, buffer->image2d。当输入 tensor 个数为N(N>2)时，会额外执行N次image2d->buffer和1次buffer->image2d，这些额外操作在 adreno gpu 上会引入较明显的不必要开销。【本PR内容】
增加一个通用型 concat kernel，可以支持 2-6 个输入 concat，使用限制条件为当对 channel 维度执行 concat 时所有输入 tensor 的 channel 必须满足 4 的整数倍，对其他维度执行 concat 则无限制。
kernel 中将重复代码放在宏中，可以实现较为简单的扩展和维护。
【效果】
ssd_mobilenetv3-large 模型中只有 2 个 conat，均为 8 输入，axis 均为 1：