From: Matthew Fernandez <matthew.fernandez@gmail.com>
Date: Sun, 18 Sep 2022 16:41:12 +0000 (-0700)
Subject: topfish fill_neighbors_vec: use 'fabsf' instead of casting to float
X-Git-Tag: 6.0.2~32^2~10
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=99b4a5cd2be269df355194dc96613d8e9261cbd0;p=graphviz

topfish fill_neighbors_vec: use 'fabsf' instead of casting to float

This is simpler and more efficient when both the input and the desired output
type are floats.
---

diff --git a/lib/topfish/hierarchy.c b/lib/topfish/hierarchy.c
index b4fd1fd62..7ae51bdf9 100644
--- a/lib/topfish/hierarchy.c
+++ b/lib/topfish/hierarchy.c
@@ -60,7 +60,7 @@ static void fill_neighbors_vec(v_data *graph, int vtx, float *vtx_vec) {
     int j;
     if (graph[0].ewgts != NULL) {
 	for (j = 0; j < graph[vtx].nedges; j++) {
-	    vtx_vec[graph[vtx].edges[j]] = (float) fabs(graph[vtx].ewgts[j]);	// use fabs for the self loop
+	    vtx_vec[graph[vtx].edges[j]] = fabsf(graph[vtx].ewgts[j]);	// use fabsf for the self loop
 	}
     } else {
 	for (j = 0; j < graph[vtx].nedges; j++) {