Skip to content

⚡ Optimize ComputeDerivativesFunction to reduce allocation churn #5

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ysdede wants to merge 1 commit into
base: master
Choose a base branch
from
Open

⚡ Optimize ComputeDerivativesFunction to reduce allocation churn #5

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 17 additions & 6 deletions app/src/main/java/net/sourceforge/opencamera/JavaImageFunctions.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2971,6 +2971,8 @@ static class ComputeDerivativesFunction implements JavaImageProcessing.ApplyFunc
private final Bitmap bitmap_in;
private final int width, height;
private JavaImageProcessing.FastAccessBitmap [] fast_bitmap_in;
private int [][] cache_bitmap_Ix;
private int [][] cache_bitmap_Iy;

ComputeDerivativesFunction(Bitmap bitmap_Ix, Bitmap bitmap_Iy, Bitmap bitmap_in) {
this.bitmap_Ix = bitmap_Ix;
Expand All @@ -2983,6 +2985,8 @@ static class ComputeDerivativesFunction implements JavaImageProcessing.ApplyFunc
@Override
public void init(int n_threads) {
fast_bitmap_in = new JavaImageProcessing.FastAccessBitmap[n_threads];
cache_bitmap_Ix = new int[n_threads][];
cache_bitmap_Iy = new int[n_threads][];

for(int i=0;i<n_threads;i++) {
fast_bitmap_in[i] = new JavaImageProcessing.FastAccessBitmap(bitmap_in);
Expand All @@ -2992,8 +2996,15 @@ public void init(int n_threads) {
@Override
public void apply(JavaImageProcessing.CachedBitmap output, int thread_index, int off_x, int off_y, int this_width, int this_height) {
// we could move these to class members for performance, remember we'd have to have a version per-thread
int [] cache_bitmap_Ix = new int[this_width*this_height];
int [] cache_bitmap_Iy = new int[this_width*this_height];
int required_size = this_width * this_height;
if (cache_bitmap_Ix[thread_index] == null || cache_bitmap_Ix[thread_index].length < required_size) {
cache_bitmap_Ix[thread_index] = new int[required_size];
}
if (cache_bitmap_Iy[thread_index] == null || cache_bitmap_Iy[thread_index].length < required_size) {
cache_bitmap_Iy[thread_index] = new int[required_size];
}
int [] current_cache_Ix = cache_bitmap_Ix[thread_index];
int [] current_cache_Iy = cache_bitmap_Iy[thread_index];

for(int y=off_y,c=0;y<off_y+this_height;y++) {

Expand Down Expand Up @@ -3041,13 +3052,13 @@ public void apply(JavaImageProcessing.CachedBitmap output, int thread_index, int

//bitmap_Ix.setPixel(x, y, Ix << 24);
//bitmap_Iy.setPixel(x, y, Iy << 24);
cache_bitmap_Ix[c] = Ix << 24;
cache_bitmap_Iy[c] = Iy << 24;
current_cache_Ix[c] = Ix << 24;
current_cache_Iy[c] = Iy << 24;
}
}

bitmap_Ix.setPixels(cache_bitmap_Ix, 0, this_width, off_x, off_y, this_width, this_height);
bitmap_Iy.setPixels(cache_bitmap_Iy, 0, this_width, off_x, off_y, this_width, this_height);
bitmap_Ix.setPixels(current_cache_Ix, 0, this_width, off_x, off_y, this_width, this_height);
bitmap_Iy.setPixels(current_cache_Iy, 0, this_width, off_x, off_y, this_width, this_height);
}

@Override
Expand Down