From ab9574bf1e1f8295a1bae0f9f520d897c65524ab Mon Sep 17 00:00:00 2001
From: Russel Arbore <russel.jma@gmail.com>
Date: Wed, 5 Mar 2025 09:59:03 -0600
Subject: [PATCH 1/2] edge opt
---
.../benches/edge_detection_bench.rs | 2 -
juno_samples/edge_detection/src/cpu.sch | 4 +-
.../edge_detection/src/edge_detection.jn | 40 +++++++++----------
juno_samples/edge_detection/src/lib.rs | 2 -
4 files changed, 23 insertions(+), 25 deletions(-)
diff --git a/juno_samples/edge_detection/benches/edge_detection_bench.rs b/juno_samples/edge_detection/benches/edge_detection_bench.rs
index 76035275..49e67a29 100644
--- a/juno_samples/edge_detection/benches/edge_detection_bench.rs
+++ b/juno_samples/edge_detection/benches/edge_detection_bench.rs
@@ -88,8 +88,6 @@ fn edge_detection_bench(c: &mut Criterion) {
height as u64,
width as u64,
gs as u64,
- sz as u64,
- sb as u64,
input_h.to(),
gaussian_filter_h,
structure_h,
diff --git a/juno_samples/edge_detection/src/cpu.sch b/juno_samples/edge_detection/src/cpu.sch
index 64fee6b6..34431711 100644
--- a/juno_samples/edge_detection/src/cpu.sch
+++ b/juno_samples/edge_detection/src/cpu.sch
@@ -134,7 +134,9 @@ if !feature("seq") {
reject_zero_crossings = reject_zero_crossings_body;
}
-async-call(edge_detection@le, edge_detection@zc);
+if !feature("seq") {
+ async-call(edge_detection@le, edge_detection@zc);
+}
fork-split(gaussian_smoothing, laplacian_estimate, zero_crossings, gradient, reject_zero_crossings);
unforkify(gaussian_smoothing, laplacian_estimate, zero_crossings, gradient, reject_zero_crossings);
diff --git a/juno_samples/edge_detection/src/edge_detection.jn b/juno_samples/edge_detection/src/edge_detection.jn
index 3e49cb36..da1977b6 100644
--- a/juno_samples/edge_detection/src/edge_detection.jn
+++ b/juno_samples/edge_detection/src/edge_detection.jn
@@ -13,12 +13,9 @@ fn gaussian_smoothing<n, m, gs : usize>(
@filter_loop for i = 0 to gs {
for j = 0 to gs {
- let val = input[if row + i < gr then 0
- else if row + i - gr > n - 1 then n - 1
- else row + i - gr,
- if col + j < gr then 0
- else if col + j - gr > m - 1 then m - 1
- else col + j - gr];
+ let br = min!(max!(row + i, gr) - gr, n - 1);
+ let bc = min!(max!(col + j, gr) - gr, m - 1);
+ let val = input[br, bc];
smoothed += val * filter[i, j];
}
}
@@ -33,10 +30,11 @@ fn gaussian_smoothing<n, m, gs : usize>(
const MIN_BR : f32 = 0;
const MAX_BR : f32 = 1;
-fn laplacian_estimate<n, m, sz: usize>(
+fn laplacian_estimate<n, m : usize>(
input: f32[n, m],
- structure: f32[sz, sz],
+ structure: f32[3, 3],
) -> f32[n, m] {
+ const sz = 3;
const r = sz / 2;
@res let result : f32[n, m];
@@ -77,10 +75,11 @@ fn laplacian_estimate<n, m, sz: usize>(
return result;
}
-fn zero_crossings<n, m, sz: usize>(
+fn zero_crossings<n, m : usize>(
input: f32[n, m],
- structure: f32[sz, sz],
+ structure: f32[3, 3],
) -> f32[n, m] {
+ const sz = 3;
const r = sz / 2;
@res let result : f32[n, m];
@@ -123,11 +122,12 @@ fn zero_crossings<n, m, sz: usize>(
return result;
}
-fn gradient<n, m, sb: usize>(
+fn gradient<n, m : usize>(
input: f32[n, m],
- sx: f32[sb, sb],
- sy: f32[sb, sb],
+ sx: f32[3, 3],
+ sy: f32[3, 3],
) -> f32[n, m] {
+ const sb = 3;
const sbr = sb / 2;
@res let result : f32[n, m];
@@ -191,18 +191,18 @@ fn reject_zero_crossings<n, m: usize>(
}
#[entry]
-fn edge_detection<n, m, gs, sz, sb: usize>(
+fn edge_detection<n, m, gs: usize>(
input: f32[n, m],
gaussian_filter: f32[gs, gs],
- structure: f32[sz, sz],
- sx: f32[sb, sb],
- sy: f32[sb, sb],
+ structure: f32[3, 3],
+ sx: f32[3, 3],
+ sy: f32[3, 3],
theta: f32,
) -> f32[n, m] {
let smoothed = gaussian_smoothing::<n, m, gs>(input, gaussian_filter);
- @le let laplacian = laplacian_estimate::<n, m, sz>(smoothed, structure);
- @zc let zcs = zero_crossings::<n, m, sz>(laplacian, structure);
- let gradient = gradient::<n, m, sb>(smoothed, sx, sy);
+ @le let laplacian = laplacian_estimate::<n, m>(smoothed, structure);
+ @zc let zcs = zero_crossings::<n, m>(laplacian, structure);
+ let gradient = gradient::<n, m>(smoothed, sx, sy);
let maxgrad = max_gradient::<n, m>(gradient);
return reject_zero_crossings::<n, m>(zcs, gradient, maxgrad, theta);
}
diff --git a/juno_samples/edge_detection/src/lib.rs b/juno_samples/edge_detection/src/lib.rs
index aa44e2e7..3eaf38dc 100644
--- a/juno_samples/edge_detection/src/lib.rs
+++ b/juno_samples/edge_detection/src/lib.rs
@@ -195,8 +195,6 @@ pub fn edge_detection_harness(args: EdgeDetectionInputs) {
height as u64,
width as u64,
gs as u64,
- sz as u64,
- sb as u64,
input_h.to(),
gaussian_filter_h.to(),
structure_h.to(),
--
GitLab
From ba7cfb6a8872ca7975f36cf2f4aa3be317cc5cfa Mon Sep 17 00:00:00 2001
From: Russel Arbore <russel.jma@gmail.com>
Date: Wed, 5 Mar 2025 10:01:44 -0600
Subject: [PATCH 2/2] Inline gaussian size
---
.../benches/edge_detection_bench.rs | 1 -
juno_samples/edge_detection/src/edge_detection.jn | 15 ++++++++-------
juno_samples/edge_detection/src/lib.rs | 1 -
3 files changed, 8 insertions(+), 9 deletions(-)
diff --git a/juno_samples/edge_detection/benches/edge_detection_bench.rs b/juno_samples/edge_detection/benches/edge_detection_bench.rs
index 49e67a29..01ae418a 100644
--- a/juno_samples/edge_detection/benches/edge_detection_bench.rs
+++ b/juno_samples/edge_detection/benches/edge_detection_bench.rs
@@ -87,7 +87,6 @@ fn edge_detection_bench(c: &mut Criterion) {
r.run(
height as u64,
width as u64,
- gs as u64,
input_h.to(),
gaussian_filter_h,
structure_h,
diff --git a/juno_samples/edge_detection/src/edge_detection.jn b/juno_samples/edge_detection/src/edge_detection.jn
index da1977b6..be987812 100644
--- a/juno_samples/edge_detection/src/edge_detection.jn
+++ b/juno_samples/edge_detection/src/edge_detection.jn
@@ -1,6 +1,10 @@
-fn gaussian_smoothing<n, m, gs : usize>(
+const gs : usize = 7;
+const sz : usize = 3;
+const sb : usize = 3;
+
+fn gaussian_smoothing<n, m : usize>(
input: f32[n, m],
- filter: f32[gs, gs],
+ filter: f32[7, 7],
) -> f32[n, m] {
@res let result : f32[n, m];
@@ -34,7 +38,6 @@ fn laplacian_estimate<n, m : usize>(
input: f32[n, m],
structure: f32[3, 3],
) -> f32[n, m] {
- const sz = 3;
const r = sz / 2;
@res let result : f32[n, m];
@@ -79,7 +82,6 @@ fn zero_crossings<n, m : usize>(
input: f32[n, m],
structure: f32[3, 3],
) -> f32[n, m] {
- const sz = 3;
const r = sz / 2;
@res let result : f32[n, m];
@@ -127,7 +129,6 @@ fn gradient<n, m : usize>(
sx: f32[3, 3],
sy: f32[3, 3],
) -> f32[n, m] {
- const sb = 3;
const sbr = sb / 2;
@res let result : f32[n, m];
@@ -191,7 +192,7 @@ fn reject_zero_crossings<n, m: usize>(
}
#[entry]
-fn edge_detection<n, m, gs: usize>(
+fn edge_detection<n, m : usize>(
input: f32[n, m],
gaussian_filter: f32[gs, gs],
structure: f32[3, 3],
@@ -199,7 +200,7 @@ fn edge_detection<n, m, gs: usize>(
sy: f32[3, 3],
theta: f32,
) -> f32[n, m] {
- let smoothed = gaussian_smoothing::<n, m, gs>(input, gaussian_filter);
+ let smoothed = gaussian_smoothing::<n, m>(input, gaussian_filter);
@le let laplacian = laplacian_estimate::<n, m>(smoothed, structure);
@zc let zcs = zero_crossings::<n, m>(laplacian, structure);
let gradient = gradient::<n, m>(smoothed, sx, sy);
diff --git a/juno_samples/edge_detection/src/lib.rs b/juno_samples/edge_detection/src/lib.rs
index 3eaf38dc..c1d04b0f 100644
--- a/juno_samples/edge_detection/src/lib.rs
+++ b/juno_samples/edge_detection/src/lib.rs
@@ -194,7 +194,6 @@ pub fn edge_detection_harness(args: EdgeDetectionInputs) {
r.run(
height as u64,
width as u64,
- gs as u64,
input_h.to(),
gaussian_filter_h.to(),
structure_h.to(),
--
GitLab