The CM-2's natural data layout is not conducive to exchanging data with other machines. Before CM-2 data is sent 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 different processor, requiring n send operations through the CM-2's global router network. The time required to transpose the data limits the effective 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 effective throughput of only 4.9 MB/s can be achieved. The authors describe the CM-2 transpose problem and study ways to improve the performance of transposed data transmissions.
|Original language||English (US)|
|Title of host publication||Proceedings of 7th International Parallel Processing Symposium, IPPS 1993|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Number of pages||6|
|ISBN (Electronic)||0818634421, 9780818634420|
|State||Published - 1993|
|Event||7th International Parallel Processing Symposium, IPPS 1993 - Newport, United States|
Duration: Apr 13 1993 → Apr 16 1993
|Name||Proceedings of 7th International Parallel Processing Symposium, IPPS 1993|
|Conference||7th International Parallel Processing Symposium, IPPS 1993|
|Period||4/13/93 → 4/16/93|
Bibliographical noteFunding Information:
t Research supported in part by the University of Minnesota Army High Performance Computing Research Center under DoD contract No. DAAL03-89-C-0038.
© 1993 IEEE.
Copyright 2019 Elsevier B.V., All rights reserved.