Browse files

Mentioned VEXCL_SHOW_KERNELS in documentation

  • Loading branch information...
1 parent f7cad66 commit 16511945a1abe67093c40359238ace5cb77fdeef @ddemidov committed Jun 10, 2012
Showing with 6 additions and 2 deletions.
  1. +3 −1 README.md
  2. +3 −1 vexcl/vexcl.hpp
View
4 README.md
@@ -82,7 +82,9 @@ vex::vector<double> Z(ctx.queue(), n);
You can now use simple vector arithmetic with device vectors. For every
expression you use, appropriate kernel is compiled (first time it is
-encountered in your program) and called automagically.
+encountered in your program) and called automagically. If you want to see
+sources of the generated kernels on the standard output, define
+`VEXCL_SHOW_KERNELS` macro before including VexCL headers.
Vectors are processed in parallel across all devices they were allocated on:
```C++
View
4 vexcl/vexcl.hpp
@@ -111,7 +111,9 @@ vex::vector<double> Z(ctx.queue(), n);
You can now use simple vector arithmetic with device vector. For every
expression you use, appropriate kernel is compiled (first time it is
-encountered in your program) and called automagically.
+encountered in your program) and called automagically. If you want to see
+sources of the generated kernels on the standard output, define
+`VEXCL_SHOW_KERNELS` macro before including VexCL headers.
Vectors are processed in parallel across all devices they were allocated on:
\code

0 comments on commit 1651194

Please sign in to comment.