Turn on OCL_STRICT_CONFORMANCE as default.

Submitted by Yang, Rong R on June 15, 2015, 8:35 a.m.

Details

Message ID 1434357300-1020-1-git-send-email-rong.r.yang@intel.com
State New
Headers show

Not browsing as part of any series.

Commit Message

Yang, Rong R June 15, 2015, 8:35 a.m.
Also update README.md.

Signed-off-by: Yang Rong <rong.r.yang@intel.com>
---
 backend/src/backend/program.cpp | 2 +-
 docs/Beignet.mdwn               | 6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

Patch hide | download patch | download mbox

diff --git a/backend/src/backend/program.cpp b/backend/src/backend/program.cpp
index 5d1ea23..e4cdeaa 100644
--- a/backend/src/backend/program.cpp
+++ b/backend/src/backend/program.cpp
@@ -113,7 +113,7 @@  namespace gbe {
 
 #ifdef GBE_COMPILER_AVAILABLE
   BVAR(OCL_OUTPUT_GEN_IR, false);
-  BVAR(OCL_STRICT_CONFORMANCE, false);
+  BVAR(OCL_STRICT_CONFORMANCE, true);
 
   bool Program::buildFromLLVMFile(const char *fileName, const void* module, std::string &error, int optLevel) {
     ir::Unit *unit = new ir::Unit();
diff --git a/docs/Beignet.mdwn b/docs/Beignet.mdwn
index c0650bb..840f028 100644
--- a/docs/Beignet.mdwn
+++ b/docs/Beignet.mdwn
@@ -195,11 +195,11 @@  Known Issues
 * Precision issue.
   Currently Gen does not provide native support of high precision math functions
   required by OpenCL. We provide a software version to achieve high precision,
-  which you can turn on through
+  which you can turn off through
 
-  `# export OCL_STRICT_CONFORMANCE=1`.
+  `# export OCL_STRICT_CONFORMANCE=0`.
 
-  But be careful, this would make your CL kernel run a little longer.
+  This would lost some precision but gain performance.
 
 * cl\_khr\_gl\_sharing.
   This extension highly depends on mesa support. It seems that mesa would not provide

Comments

I think this is good. Now we comply with spec and users can still make their own choice.
LGTM

> -----Original Message-----

> From: Beignet [mailto:beignet-bounces@lists.freedesktop.org] On Behalf Of

> Yang Rong

> Sent: Monday, June 15, 2015 4:35 PM

> To: beignet@lists.freedesktop.org

> Cc: Yang, Rong R

> Subject: [Beignet] [PATCH] Turn on OCL_STRICT_CONFORMANCE as default.

> 

> Also update README.md.

> 

> Signed-off-by: Yang Rong <rong.r.yang@intel.com>

> ---

>  backend/src/backend/program.cpp | 2 +-

>  docs/Beignet.mdwn               | 6 +++---

>  2 files changed, 4 insertions(+), 4 deletions(-)

> 

> diff --git a/backend/src/backend/program.cpp

> b/backend/src/backend/program.cpp index 5d1ea23..e4cdeaa 100644

> --- a/backend/src/backend/program.cpp

> +++ b/backend/src/backend/program.cpp

> @@ -113,7 +113,7 @@ namespace gbe {

> 

>  #ifdef GBE_COMPILER_AVAILABLE

>    BVAR(OCL_OUTPUT_GEN_IR, false);

> -  BVAR(OCL_STRICT_CONFORMANCE, false);

> +  BVAR(OCL_STRICT_CONFORMANCE, true);

> 

>    bool Program::buildFromLLVMFile(const char *fileName, const void*

> module, std::string &error, int optLevel) {

>      ir::Unit *unit = new ir::Unit();

> diff --git a/docs/Beignet.mdwn b/docs/Beignet.mdwn index

> c0650bb..840f028 100644

> --- a/docs/Beignet.mdwn

> +++ b/docs/Beignet.mdwn

> @@ -195,11 +195,11 @@ Known Issues

>  * Precision issue.

>    Currently Gen does not provide native support of high precision math

> functions

>    required by OpenCL. We provide a software version to achieve high

> precision,

> -  which you can turn on through

> +  which you can turn off through

> 

> -  `# export OCL_STRICT_CONFORMANCE=1`.

> +  `# export OCL_STRICT_CONFORMANCE=0`.

> 

> -  But be careful, this would make your CL kernel run a little longer.

> +  This would lost some precision but gain performance.

> 

>  * cl\_khr\_gl\_sharing.

>    This extension highly depends on mesa support. It seems that mesa

> would not provide

> --

> 1.8.3.2

> 

> _______________________________________________

> Beignet mailing list

> Beignet@lists.freedesktop.org

> http://lists.freedesktop.org/mailman/listinfo/beignet