[clfft] 27/32: reducing tolerance to tighten margin, fixing batch bug

Ghislain Vaillant ghisvail-guest at moszumanska.debian.org
Tue Apr 26 08:34:11 UTC 2016


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

ghisvail-guest pushed a commit to branch master
in repository clfft.

commit 0a9b9115fb266bbba23a23471f84bb43432a0176
Author: bragadeesh <bragadeesh.natarajan at amd>
Date:   Mon Apr 11 11:15:18 2016 -0700

    reducing tolerance to tighten margin, fixing batch bug
---
 src/tests/buffer.h       | 2 --
 src/tests/gtest_main.cpp | 2 +-
 2 files changed, 1 insertion(+), 3 deletions(-)

diff --git a/src/tests/buffer.h b/src/tests/buffer.h
index beaad2c..bcaeb23 100644
--- a/src/tests/buffer.h
+++ b/src/tests/buffer.h
@@ -543,8 +543,6 @@ private:
 							", maximum allowed RMSE = " << std::dec << rmse_tolerance_this << std::endl;
 					return 1;
 				}
-				else
-					return 0;
 			}
 
 			return 0;
diff --git a/src/tests/gtest_main.cpp b/src/tests/gtest_main.cpp
index 09fb1eb..7bda880 100644
--- a/src/tests/gtest_main.cpp
+++ b/src/tests/gtest_main.cpp
@@ -123,7 +123,7 @@ int main( int argc, char **argv )
 
 	// this rmse_tolerance is not absolute; it is for a 4096-point single precision transform
 	// the actual rmse tolerance is this value times sqrt(problem-size/4096)
-	rmse_tolerance = 0.00005;
+	rmse_tolerance = 0.00002;
 
 
 	//	Parse the command line options, ignore unrecognized options and collect them into a vector of strings

-- 
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