From: Chris Duncan <chris@zoso.dev>
Date: Sat, 25 Jan 2025 07:29:26 +0000 (-0800)
Subject: Revert to single dimensional workgroup size since that apparently does in fact matter... 
X-Git-Tag: v3.0.0~70
X-Git-Url: https://git.codecow.com/?a=commitdiff_plain;h=02305420ff138f3bd50979b52a9f400f6ad2628b;p=nano-pow.git

Revert to single dimensional workgroup size since that apparently does in fact matter in regards to wavefronts and not just number of threads per workgroup.
---

diff --git a/src/shaders/compute.wgsl b/src/shaders/compute.wgsl
index e797a22..5c2e222 100644
--- a/src/shaders/compute.wgsl
+++ b/src/shaders/compute.wgsl
@@ -33,7 +33,7 @@ const ROTATE_31 = vec2(31u, 31u);
 * Search compute function
 * Calls main with a workgroup size of 64 which has been tested as optimal
 */
-@compute @workgroup_size(8,8)
+@compute @workgroup_size(64)
 fn search(@builtin(global_invocation_id) global_id: vec3<u32>) {
 	if (atomicLoad(&work.found) != 0u) { return; }
 	main(global_id);