RPP - float, int & tensor support: required for RALI-SOW3#32
Merged
kiritigowda merged 894 commits intoROCm:masterfrom Aug 25, 2020
Merged
RPP - float, int & tensor support: required for RALI-SOW3#32kiritigowda merged 894 commits intoROCm:masterfrom
kiritigowda merged 894 commits intoROCm:masterfrom
Conversation
rrawther
reviewed
Aug 7, 2020
| dst_pixIdx += dst_inc[id_z]; | ||
| } | ||
| } else { | ||
| for (indextmp = 0; indextmp < channel; indextmp++) { |
Contributor
There was a problem hiding this comment.
consider loop unrolling and vector datatypes for better performance
rrawther
reviewed
Aug 7, 2020
| (id_x + id_y * max_dst_width[id_z]) * out_plnpkdind; | ||
| if ((id_x < dst_width[id_z]) && (id_y < dst_height[id_z])) { | ||
| for (indextmp = 0; indextmp < channel; indextmp++) { | ||
| output[dst_pixIdx] = (half)((input[src_pixIdx] - local_mean) / 255.0 * local_std_dev); |
Contributor
There was a problem hiding this comment.
avoid division for all constant divisors. Use multiply by inverse instead. Applicable to all kernels
rrawther
reviewed
Aug 7, 2020
|
|
||
| unsigned int pixId; | ||
| pixId = id_x + id_y * dest_width + id_z * dest_width * dest_height; | ||
| A = srcPtr[x + y * source_width + id_z * source_height * source_width]; |
Contributor
There was a problem hiding this comment.
consider doing more work by using vector datatypes
rrawther
reviewed
Aug 7, 2020
| const unsigned int dest_height, const unsigned int dest_width, | ||
| const unsigned int channel) { | ||
| int A, B, C, D, x, y, index, pixVal; | ||
| float x_ratio = ((float)(source_width - 1)) / dest_width; |
Contributor
There was a problem hiding this comment.
it is better is pass x_ratio and y_ratio instead of computing every time
rrawther
reviewed
Aug 7, 2020
| int id_y = get_global_id(1); | ||
| int id_z = get_global_id(2); | ||
|
|
||
| int xc = id_x - dest_width / 2; |
rrawther
reviewed
Aug 7, 2020
| color_twist_host(srcPtr, batch_srcSizeMax[batchCount], dstPtr, alpha, beta, hueShift, saturationFactor, chnFormat, channel); | ||
| color_twist_host(srcPtrImage, batch_srcSizeMax[batchCount], dstPtrImage, alpha, beta, hueShift, saturationFactor, chnFormat, channel); | ||
|
|
||
| if (outputFormatToggle == 1) |
Contributor
There was a problem hiding this comment.
this looks very inefficient. Need to revisit
rrawther
reviewed
Aug 7, 2020
| xG = _mm_loadu_ps(srcPtrTempG); | ||
| xB = _mm_loadu_ps(srcPtrTempB); | ||
|
|
||
| xR = _mm_div_ps(xR, pFactor); |
Contributor
There was a problem hiding this comment.
please use mulps instead. True for all constant divisors
kiritigowda
approved these changes
Aug 14, 2020
Collaborator
kiritigowda
left a comment
There was a problem hiding this comment.
@rrawther let me know this is good to merge. it LGTM.
Contributor
|
@kiritigowda : Pavel found some issues with GPU flow. Waiting for the status of that to merge |
Collaborator
|
Issues
======
+ Solved 1
- Added 6
Complexity increasing per file
==============================
- utilities/rpp-unittests/SOW3_HOST/tensorDifference.py 1
Clones added
============
- utilities/rpp-unittests/OCL/BatchPD_ocl_pkd3.cpp 63
- utilities/rpp-unittests/SOW3_OCL/BatchPD_ocl_pkd3.cpp 24
- src/modules/cl/cl_declarations.hpp 1
- utilities/rpp-unittests/SOW3_HOST/BatchPD_host_pkd3.cpp 22
- utilities/rpp-unittests/HOST/BatchPD_host_pkd3.cpp 101
- utilities/rpp-unittests/HOST/Single_host.cpp 4
- utilities/rpp-unittests/SOW3_HOST/BatchPD_host_pln1.cpp 22
- src/modules/cl/cl_fused_functions.cpp 3
- utilities/rpp-unittests/SOW3_OCL/BatchPD_ocl_pln1.cpp 23
- utilities/rpp-unittests/SOW3_HOST/BatchPD_host_pln3.cpp 24
- utilities/rpp-unittests/SOW3_OCL/BatchPD_ocl_pln3.cpp 25
- utilities/rpp-unittests/HIP/Single_hip.cpp 9
- src/include/cpu/rpp_cpu_common.hpp 12
- utilities/rpp-unittests/HOST/BatchPD_host_pln1.cpp 102
- utilities/rpp-unittests/HIP/BatchPD_hip.cpp 8
- src/modules/cl/cl_color_model_conversions.cpp 1
- utilities/rpp-unittests/OCL/Single_ocl.cpp 6
- include/rppi_fused_functions.h 2
- src/modules/cl/cl_geometry_transforms.cpp 19
See the complete overview on Codacy |
kiritigowda
reviewed
Aug 21, 2020
kiritigowda
reviewed
Aug 21, 2020
kiritigowda
reviewed
Aug 21, 2020
kiritigowda
reviewed
Aug 21, 2020
kiritigowda
reviewed
Aug 21, 2020
kiritigowda
reviewed
Aug 21, 2020
fiona-gladwin
pushed a commit
to fiona-gladwin/rpp
that referenced
this pull request
Dec 13, 2021
Resize Bilinear interpolation - Tensor support
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
7 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.
Major Work: