Various maintenance for tiling and opencl#21075
Open
jenshannoschwalm wants to merge 3 commits into
Open
Conversation
jenshannoschwalm
added
scope: codebase
Collaborator
Author
|
@TurboGit still checking
BTW, Could we add a |
jenshannoschwalm
force-pushed
the
various_maintenance_2
branch
3 times, most recently
from
a648af3 to
fedbc45
Compare
Collaborator
Author
|
@TurboGit TIA i just added a third bug-fixing-commit. That's about fixing overestimations of tiling requirements when blending is involved! |
jenshannoschwalm
added
bugfix
1. Various functions are just internal helpers not used outside of the source scope and thus should not be exposed. dt_opencl_read_host_from_device_rowpitch() dt_opencl_write_host_to_device_rowpitch() 2. Some correct use of #defines and constify
1. dt_print_pipe() got a minor improvement, it now shows a leading `T` in the pipe name if in tiling mode. 2. use dt_print_pipe() variant where automatic display of roi is helping 3. overlap added to logs where helping 4. some logs are shown only in -d verbose mode 5. some logs are also relevant for -d pipe logs 6. added log for skipped tiles in ptp mode 7. added module instance in control logs 8. For extensive debug logs we now have compile time option DT_TILING_DEBUG 9. some constify, subtle simplifications and use of float where intended
If we blend a piece output there are additional requirements for memory to be checked. Some calculations lead to higher results than necessary if no feathering was involved.
jenshannoschwalm
force-pushed
the
various_maintenance_2
branch
from
b8f2872 to
8bb62f1
Compare
TurboGit
reviewed
May 20, 2026
| tiling->align = 1; | ||
|
|
||
| dt_develop_blend_params_t *const bldata = piece->blendop_data; | ||
| if(bldata) |
Member
There was a problem hiding this comment.
Not part of the description on this PR. Just to be sure before when bldata == NULL we were falling back to the end of the routine:
const float basic = 2.5f + outnorm; // in + out + (guide, tmp) + two quarter buffers for the mask
tiling->factor += basic;
tiling->factor_cl += basic;
So at least having 2.5f in factor, now we early return with 0.0f.
That's ok?
Collaborator
Author
There was a problem hiding this comment.
Intended :-) the more important point is, if we blend but do not feather we dont add the 2.5! That was the bad boy leading to larger tiles or even lead to tiling where not required.
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
A bit of OpenCL maintenance
dt_opencl_read_host_from_device_rowpitch()dt_opencl_write_host_to_device_rowpitch()Improved tiling related logs
dt_print_pipe()got a minor improvement, it now shows a leadingTin the pipe name if in tiling mode.dt_print_pipe()variant where automatic display of roi is helping-d verbosemode-d pipelogsDT_TILING_DEBUG