[clinfo] 26/148: CL_FP_CORRECTLY_ROUNDED_DIVIDE_SQRT floating-point config flag

[Andreas Beckmann]

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

anbe pushed a commit to branch clinfo
in repository clinfo.

commit 94d94a4b1dfa4eab955f0f15cee7403514e4f8f3
Author: Giuseppe Bilotta <giuseppe.bilotta at gmail.com>
Date:   Fri Jun 7 20:31:50 2013 +0200

    CL_FP_CORRECTLY_ROUNDED_DIVIDE_SQRT floating-point config flag
---
 man/clinfo.1 | 6 ++++++
 src/clinfo.c | 1 +
 2 files changed, 7 insertions(+)

diff --git a/man/clinfo.1 b/man/clinfo.1
index 6f13205..2c5b84b 100644
--- a/man/clinfo.1
+++ b/man/clinfo.1
@@ -47,6 +47,12 @@ the core OpenCL 1.2 device partition feature;
 the preferred work-group size multiple matches the NVIDIA warp size (on
 NVIDIA devices) or the AMD wavefront width (on AMD devices).
 
+.P
+Some floating-point configuration flags may only be meaningful for
+specific precisions and/or specific OpenCL versions. For example,
+.BR CL_FP_CORRECTLY_ROUNDED_DIVIDE_SQRT ,
+is only relevant for single precision in OpenCL 1.2 devices.
+
 .SH BUGS
 The results of the following undocumented properties might be
 interpreted incorrectly:
diff --git a/src/clinfo.c b/src/clinfo.c
index dc0edfd..f7c6850 100644
--- a/src/clinfo.c
+++ b/src/clinfo.c
@@ -483,6 +483,7 @@ printDeviceInfo(cl_uint d)
 	SHOW_FP_FLAG("Round to zero", ROUND_TO_ZERO); \
 	SHOW_FP_FLAG("Round to infinity", ROUND_TO_INF); \
 	SHOW_FP_FLAG("IEEE754-2008 fused multiply-add", FMA); \
+	SHOW_FP_FLAG("Correctly-rounded divide and sqrt operations", CORRECTLY_ROUNDED_DIVIDE_SQRT); \
 	SHOW_FP_FLAG("Support is emulated in software", SOFT_FLOAT); \
 } while (0)
 

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/pkg-opencl/clinfo.git