diff options
| -rw-r--r-- | libclc/generic/include/clc/async/async_work_group_copy.h | 15 | ||||
| -rw-r--r-- | libclc/generic/include/clc/async/async_work_group_copy.inc | 5 | ||||
| -rw-r--r-- | libclc/generic/include/clc/clc.h | 1 | ||||
| -rw-r--r-- | libclc/generic/lib/SOURCES | 1 | ||||
| -rw-r--r-- | libclc/generic/lib/async/async_work_group_copy.cl | 9 | ||||
| -rw-r--r-- | libclc/generic/lib/async/async_work_group_copy.inc | 17 |
6 files changed, 48 insertions, 0 deletions
diff --git a/libclc/generic/include/clc/async/async_work_group_copy.h b/libclc/generic/include/clc/async/async_work_group_copy.h new file mode 100644 index 00000000000..39c637b0e26 --- /dev/null +++ b/libclc/generic/include/clc/async/async_work_group_copy.h @@ -0,0 +1,15 @@ +#define __CLC_DST_ADDR_SPACE local +#define __CLC_SRC_ADDR_SPACE global +#define __CLC_BODY <clc/async/async_work_group_copy.inc> +#include <clc/async/gentype.inc> +#undef __CLC_DST_ADDR_SPACE +#undef __CLC_SRC_ADDR_SPACE +#undef __CLC_BODY + +#define __CLC_DST_ADDR_SPACE global +#define __CLC_SRC_ADDR_SPACE local +#define __CLC_BODY <clc/async/async_work_group_copy.inc> +#include <clc/async/gentype.inc> +#undef __CLC_DST_ADDR_SPACE +#undef __CLC_SRC_ADDR_SPACE +#undef __CLC_BODY diff --git a/libclc/generic/include/clc/async/async_work_group_copy.inc b/libclc/generic/include/clc/async/async_work_group_copy.inc new file mode 100644 index 00000000000..d85df6c8fad --- /dev/null +++ b/libclc/generic/include/clc/async/async_work_group_copy.inc @@ -0,0 +1,5 @@ +_CLC_OVERLOAD _CLC_DECL event_t async_work_group_copy( + __CLC_DST_ADDR_SPACE __CLC_GENTYPE *dst, + const __CLC_SRC_ADDR_SPACE __CLC_GENTYPE *src, + size_t num_gentypes, + event_t event); diff --git a/libclc/generic/include/clc/clc.h b/libclc/generic/include/clc/clc.h index 85bf0fa4286..2afbcfdaccd 100644 --- a/libclc/generic/include/clc/clc.h +++ b/libclc/generic/include/clc/clc.h @@ -137,6 +137,7 @@ #include <clc/synchronization/barrier.h> /* 6.11.10 Async Copy and Prefetch Functions */ +#include <clc/async/async_work_group_copy.h> #include <clc/async/async_work_group_strided_copy.h> #include <clc/async/prefetch.h> #include <clc/async/wait_group_events.h> diff --git a/libclc/generic/lib/SOURCES b/libclc/generic/lib/SOURCES index 1dd3ff717c3..f7b3adf6329 100644 --- a/libclc/generic/lib/SOURCES +++ b/libclc/generic/lib/SOURCES @@ -1,3 +1,4 @@ +async/async_work_group_copy.cl async/async_work_group_strided_copy.cl async/prefetch.cl async/wait_group_events.cl diff --git a/libclc/generic/lib/async/async_work_group_copy.cl b/libclc/generic/lib/async/async_work_group_copy.cl new file mode 100644 index 00000000000..fe20ecfd9bb --- /dev/null +++ b/libclc/generic/lib/async/async_work_group_copy.cl @@ -0,0 +1,9 @@ +#include <clc/clc.h> + +#ifdef cl_khr_fp64 +#pragma OPENCL EXTENSION cl_khr_fp64 : enable +#endif + +#define __CLC_BODY <async_work_group_copy.inc> +#include <clc/async/gentype.inc> +#undef __CLC_BODY diff --git a/libclc/generic/lib/async/async_work_group_copy.inc b/libclc/generic/lib/async/async_work_group_copy.inc new file mode 100644 index 00000000000..a143ddfb9f6 --- /dev/null +++ b/libclc/generic/lib/async/async_work_group_copy.inc @@ -0,0 +1,17 @@ +_CLC_OVERLOAD _CLC_DEF event_t async_work_group_copy( + local __CLC_GENTYPE *dst, + const global __CLC_GENTYPE *src, + size_t num_gentypes, + event_t event) { + + return async_work_group_strided_copy(dst, src, num_gentypes, 1, event); +} + +_CLC_OVERLOAD _CLC_DEF event_t async_work_group_copy( + global __CLC_GENTYPE *dst, + const local __CLC_GENTYPE *src, + size_t num_gentypes, + event_t event) { + + return async_work_group_strided_copy(dst, src, num_gentypes, 1, event); +} |
