max3 = nTx*nTy
IF ( nThreads .NE. nTx*nTy ) THEN
& ' nTx * nTy = ',nTx*nTy,' nThreads = ',nThreads
byPerThread = nSy/nTy
IF ( byPerThread*nTy .NE. nSy ) THEN
DO J=1,nTy
nThreads = nTx * nTy