diff --git a/csrc/kernel_v1.cu b/csrc/kernel_v1.cu
index a89d15e..bf47c1d 100644
--- a/csrc/kernel_v1.cu
+++ b/csrc/kernel_v1.cu
@@ -25,7 +25,7 @@ void simulate_v1_noninteractive_simple_scan_cuda(
   usize num_cycles,
   usize state_size,
   u32 *states_noninteractive,
-  int arrival_state_offset
+  i32 arrival_state_offset
   )
 {
   const u32 *timing_constraints = nullptr;
@@ -59,7 +59,7 @@ void simulate_v1_noninteractive_timed_cuda(
   u32 *states_noninteractive,
   const u32 *timing_constraints,
   u8 *event_buffer,
-  int arrival_state_offset
+  i32 arrival_state_offset
   )
 {
   void *arg_ptrs[12] = {
diff --git a/csrc/kernel_v1.hip.cpp b/csrc/kernel_v1.hip.cpp
index a3fa933..a941a94 100644
--- a/csrc/kernel_v1.hip.cpp
+++ b/csrc/kernel_v1.hip.cpp
@@ -42,7 +42,7 @@ void simulate_v1_noninteractive_simple_scan_hip(
   usize num_cycles,
   usize state_size,
   u32 *states_noninteractive,
-  int arrival_state_offset
+  i32 arrival_state_offset
   )
 {
   validate_warp_size();
@@ -79,7 +79,7 @@ void simulate_v1_noninteractive_timed_hip(
   u32 *states_noninteractive,
   const u32 *timing_constraints,
   u8 *event_buffer,
-  int arrival_state_offset
+  i32 arrival_state_offset
   )
 {
   validate_warp_size();
diff --git a/csrc/kernel_v1_impl.cuh b/csrc/kernel_v1_impl.cuh
index 2a4dfb9..b9a26c1 100644
--- a/csrc/kernel_v1_impl.cuh
+++ b/csrc/kernel_v1_impl.cuh
@@ -52,7 +52,7 @@ __device__ void simulate_block_v1(
   u32 clock_period_ps,
   EventBuffer *__restrict__ event_buffer,
   u32 cycle_i,
-  int arrival_state_offset  // offset in output_state for arrival data (0 = disabled)
+  i32 arrival_state_offset  // offset in output_state for arrival data (0 = disabled)
   )
 {
   int script_pi = 0;
@@ -574,7 +574,7 @@ __global__ void simulate_v1_noninteractive_simple_scan(
   u32 *__restrict__ states_noninteractive,
   const u32 *__restrict__ timing_constraints,
   EventBuffer *__restrict__ event_buffer,
-  int arrival_state_offset
+  i32 arrival_state_offset
   )
 {
   assert(num_blocks == gridDim.x);