From 69cbb7fa99643bab19ad70842da1eb4a92496a57 Mon Sep 17 00:00:00 2001
From: rob thijssen <rthijssen@gmail.com>
Date: Fri, 12 Sep 2025 16:24:45 +0300
Subject: [PATCH] remove 1024 thread cap in CUDA kernel launches

---
 crates/engine-gpu-cuda/src/lib.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/crates/engine-gpu-cuda/src/lib.rs b/crates/engine-gpu-cuda/src/lib.rs
index 1d98cf9..d9aac14 100644
--- a/crates/engine-gpu-cuda/src/lib.rs
+++ b/crates/engine-gpu-cuda/src/lib.rs
@@ -305,7 +305,7 @@ impl CudaEngine {
             // Determine a small launch window (cover next nonces via GPU)
             // coverage bounded by remaining is enforced during consumption
 
-            let num_threads = threads.min(1024);
+            let num_threads = threads;
             // Use many iterations per thread to reduce host exponentiations per launch
             let grid_dim = ((num_threads + block_dim - 1) / block_dim).max(1);