AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[hip] add initial implementation for hipLaunchCooperativeKernel API (#1339)

Browse Source 
* [hip] add initial implementation for hipLaunchCooperativeKernel API

* [hip] use total number of work groups to initialize the GWS resource

* [hip] use only one argument for init_gws kernel

* [hip] use the device associated with the stream for checking the device properties


[ROCm/clr commit: 32ce882d6e]
This commit is contained in:
Aryan Salmanpour
2019-08-23 05:19:35 -04:00
committed by Maneesh Gupta
parent 14a510ee74
commit e8f69a8c36
2 changed files with 98 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/clr/hipamd/include/hip/hcc_detail/device_library_decls.h
+4
View File
@@ -59,6 +59,10 @@ extern "C" __device__ __attribute__((const)) float __ocml_trunc_f32(float);
extern "C" __device__ __attribute__((const)) float __ocml_fmin_f32(float, float);
extern "C" __device__ __attribute__((const)) float __ocml_fmax_f32(float, float);
extern "C" __device__ __attribute__((convergent)) void __ockl_gws_init(uint nwm1, uint rid);
extern "C" __device__ __attribute__((convergent)) void __ockl_gws_barrier(uint nwm1, uint rid);
// Introduce local address space
#define __local __attribute__((address_space(3)))