From: Chris Duncan Date: Fri, 7 Feb 2025 22:53:30 +0000 (-0800) Subject: Test workgroup size of 32 since Nvidia card is able to take advantage of L1 cache... X-Git-Tag: v3.0.0~21 X-Git-Url: https://zoso.dev/?a=commitdiff_plain;h=042d1d9fa50d70ca8ffbdba463d6b7b2e8b9effb;p=nano-pow.git Test workgroup size of 32 since Nvidia card is able to take advantage of L1 cache this way. --- diff --git a/src/shaders/compute.wgsl b/src/shaders/compute.wgsl index 8dd90cd..5a828dd 100644 --- a/src/shaders/compute.wgsl +++ b/src/shaders/compute.wgsl @@ -49,7 +49,7 @@ var found: bool; * workgroup exits immediately if a nonce was already found by a previous * workgroup. */ -@compute @workgroup_size(64) +@compute @workgroup_size(32) fn search(@builtin(global_invocation_id) global_id: vec3, @builtin(local_invocation_id) local_id: vec3) { // found = (local_id.x == 0u && atomicLoad(&work.found) != 0u); // workgroupBarrier();