loopy_gbarrier_error.py

import numpy as np                                                                                                         
import numpy.linalg as la                                                                                                  
import matplotlib.pyplot as plt                                                                                            
from time import time                                                                                                      
import loopy as lp                                                                                                         
import pyopencl as cl                                                                                                      
import pyopencl.array                                                                                                      
import pyopencl.clrandom                                                                                                   
                                                                                                                           
                                                                                                                           
if __name__ == '__main__':                                                                                                 
    ctx = cl.create_some_context(interactive=True)                                                                         
    queue = cl.CommandQueue(ctx)                                                                                           
                                                                                                                           
    knl = lp.make_kernel(                                                                                                  
            "{[i, j, ii, jj]: 0<=i,j, ii, jj<n}",                                                                          
            """                                                                                                            
            out[j, i] = a[i, j]{id=transpose}                                                                              
            out[ii, jj] = 2*out[ii, jj]{id=double}                                                                         
            """)                                                                                                           
    a = np.random.randn(16, 16)                                                                                            
    knl = lp.set_options(knl, write_cl=True)                                                                               
    knl = lp.add_barrier(knl, "gb1", "id:transpose", "id:double")                                                          
                                                                                                                           
                                                                                                                           
    knl = lp.split_iname(knl, "i", 2, outer_tag="g.0", inner_tag="l.0")                                                    
    knl = lp.split_iname(knl, "j", 2, outer_tag="g.1", inner_tag="l.1")                                                    
    '''                                                                                                                    
    knl = lp.split_iname(knl, "ii", 2, outer_tag="g.2", inner_tag="l.2")                                                   
    knl = lp.split_iname(knl, "jj", 2, outer_tag="g.3", inner_tag="l.3")                                                   
    '''                                                                                                                    
                                                                                                                           
    evt, (out,) = knl(queue, a=a)                                                                                          
    print(la.norm(out-2*a.T))