#matrix multiply large array size for intel machine
source: gemm.c
procedure: gemm
format: rose
loop: 0

TI = 128
TJ = 8
TK = 512
UI = 2
UJ = 2

permute([3,1,2])
tile(0,2,TJ)
#print space
tile(0,2,TI)
#print space
tile(0,5,TK)
#print space

datacopy(0,3,a,false,1)
#print space

datacopy(0,4,b)
print
unroll(0,4,UI)#print space
print 
unroll(0,5,UJ)
#print space
print