max3 = nTx*nTy
IF ( nThreads .NE. nTx*nTy ) THEN
& ' nTx * nTy = ',nTx*nTy,' nThreads = ',nThreads
bxPerThread = nSx/nTx
IF ( bxPerThread*nTx .NE. nSx ) THEN
DO I=1,nTx
Thid = (J-1)*nTx+I
nThreads = nTx * nTy