The CM - 2 Data Transposition

In a previous paper 6], we noted that the CM-2's natural data layout is not conducive to exchanging data with other machines. In order for CM-2 data to be understandable to a remote machine, a bitwise transpose must be performed on the data. Each bit in an n bit value must be transmitted to a diierent processor, requiring n send operations through the CM-2's global router network. The time required to transpose the data limits the eeective throughput of the I/O channel to a small fraction of its peak theoretical bandwidth. For example, when sending data to a remote supercomputer using a 100 MB/s HIPPI channel, an eeective throughput of only 4.9 MB/s can be achieved. In this paper, we describe the CM-2 transpose problem in detail and study ways to improve the performance of transposed data transmissions.