[clfft] 11/74: fixing thread synchronisation with barriers that affected CPU functionality

Ghislain Vaillant ghisvail-guest at moszumanska.debian.org
Thu Jan 14 19:52:12 UTC 2016


This is an automated email from the git hooks/post-receive script.

ghisvail-guest pushed a commit to branch debian/sid
in repository clfft.

commit aeb196a14f9f42de9d6a16795d4f4c300b40196d
Author: bragadeesh <bragadeesh.natarajan at amd>
Date:   Thu Nov 12 16:57:31 2015 -0800

    fixing thread synchronisation with barriers that affected CPU functionality
---
 src/library/generator.stockham.cpp | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/src/library/generator.stockham.cpp b/src/library/generator.stockham.cpp
index 08f32d6..49a7ff4 100644
--- a/src/library/generator.stockham.cpp
+++ b/src/library/generator.stockham.cpp
@@ -2384,9 +2384,7 @@ namespace StockhamGenerator
 			}
 
 
-			if( (position != 0) && (!linearRegs) && (nextPass != NULL) )
-				passStr += "\n\n\tbarrier(CLK_LOCAL_MEM_FENCE);\n";
-
+			passStr += "\n\n\tbarrier(CLK_LOCAL_MEM_FENCE);\n";
 			passStr += "\n\n";
 
 			// 3-step twiddle multiplies
@@ -3925,6 +3923,7 @@ namespace StockhamGenerator
 							str += ldsArgs; str += ", ";
 							str += outBuf;
 							str += IterRegs("&"); str += ");\n";
+							if (!halfLds) { str += exTab; str += "\tbarrier(CLK_LOCAL_MEM_FENCE);\n"; }
 						}
 						else // intermediate pass
 						{

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/clfft.git



More information about the debian-science-commits mailing list