From b6a322a78b6127091170471c534cf2806a4cf812 Mon Sep 17 00:00:00 2001
From: Yi Xu <xy_xuyi@foxmail.com>
Date: Mon, 25 Jul 2022 17:25:33 +0800
Subject: [PATCH] [example] Make implicit_fem fallback to CPU when CUDA is not
 available (#5512)

---
 python/taichi/examples/simulation/implicit_fem.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/python/taichi/examples/simulation/implicit_fem.py b/python/taichi/examples/simulation/implicit_fem.py
index 8d8520dbb3b31..780d93a175004 100644
--- a/python/taichi/examples/simulation/implicit_fem.py
+++ b/python/taichi/examples/simulation/implicit_fem.py
@@ -14,10 +14,10 @@
 parser.add_argument('place_holder', nargs='*')
 args = parser.parse_args()
 
-ti.init(arch=ti.gpu, dynamic_index=True)
+ti.init(arch=ti.cuda, dynamic_index=True)
 
 if args.gui == 'auto':
-    if _ti_core.GGUI_AVAILABLE:
+    if _ti_core.GGUI_AVAILABLE and ti.lang.impl.current_cfg().arch == ti.cuda:
         args.gui = 'ggui'
     else:
         args.gui = 'cpu'