SMLNUM = slamch ('s') BIGNUM = 1 / SMLNUM
- (global) The number of rows to be operated on, that is, the number of rows of the distributed matrix sub(A)(.m≥0)
- (global) The number of columns to be operated on, that is, the number of columns of the distributed matrix sub(A)(.n≥0)
- (local)Pointer into the local memory to an array of local size.lld_a*LOCc(ja+n-1)The arrayacontains the local pieces of them-by-ndistributed matrix whose equilibration factors are to be computed.
- (global) The row and column indices in the global matrixAindicating the first row and the first column of the matrix sub(A), respectively.
- (global and local) array of sizedlen_. The array descriptor for the distributed matrixA.
- (local)Arrays of sizesandLOCr(m_a), respectively.LOCc(n_a)If, orinfo= 0,info>ia+m-1the row scale factors for sub(r[i] containA)for.ia-1≤i<ia+m-1ris aligned with the distributed matrixA, and replicated across every process column.ris tied to the distributed matrixA.If,info= 0the column scale factors for sub(c[i] containA)for.ja-1≤i<ja+n-1cis aligned with the distributed matrixA, and replicated down every process row.cis tied to the distributed matrixA.
- (global)Iforinfo= 0,info>ia+m-1rowcndcontains the ratio of the smallestto the largestri(iria≤i≤ia+m-1). Ifandrowcnd≥0.1amaxis neither too large nor too small, it is not worth scaling by.riIf,info= 0colcndcontains the ratio of the smallestto the largestcj(cjja≤j≤ja+n-1).If, it is not worth scaling bycolcnd≥0.1.cj
- (global)Absolute value of the largest matrix element. Ifamaxis very close to overflow or very close to underflow, the matrix should be scaled.
- (global) If, the execution is successful.info=0:info< 0If thei-th argument is an array and thej-th entry, indexedhad an illegal value, thenj- 1,info= -(i*100+j); if thei-th argument is a scalar and had an illegal value, theninfo=-i.:info>0Ifinfo=iandi≤m, thei-th row of the distributed matrixsub(A) is exactly zero;i>m, the (i-m)-th column of the distributedmatrix sub(A) is exactly zero.