From 91b66eb6fe213528bd71c7345f3449c4a3c0f375 Mon Sep 17 00:00:00 2001
From: Adrian Kummerlaender
Date: Mon, 10 Jun 2019 15:38:52 +0200
Subject: Reduce thread block size

=> ~780 MLUPS
---
 codegen_lbm.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'codegen_lbm.py')

diff --git a/codegen_lbm.py b/codegen_lbm.py
index 29c2dad..235a548 100644
--- a/codegen_lbm.py
+++ b/codegen_lbm.py
@@ -173,10 +173,10 @@ class D2Q9_BGK_Lattice:
     def evolve(self):
         if self.tick:
             self.tick = False
-            self.program.collide_and_stream(self.queue, (self.nX,self.nY), (64,1), self.cl_pop_a, self.cl_pop_b, self.cl_moments, self.cl_material)
+            self.program.collide_and_stream(self.queue, (self.nX,self.nY), (32,1), self.cl_pop_a, self.cl_pop_b, self.cl_moments, self.cl_material)
         else:
             self.tick = True
-            self.program.collide_and_stream(self.queue, (self.nX,self.nY), (64,1), self.cl_pop_b, self.cl_pop_a, self.cl_moments, self.cl_material)
+            self.program.collide_and_stream(self.queue, (self.nX,self.nY), (32,1), self.cl_pop_b, self.cl_pop_a, self.cl_moments, self.cl_material)
 
     def sync(self):
         self.queue.finish()
-- 
cgit v1.2.3