ODBC /performance

I am currently using SCO SQL retriever to access informix SE from a win95 client using a delphi written application. When the application is reading a large number of records 300,000 it is painfully slow, the Unix server HPUX10.20 performance dips only by 10% when retrieving the data, however the win95 cpu is 100% utilized during the process. I am suspecious of the SCO SQL retriever having a large over head. Can anyone help with any suggestions ie alternative connectivity software . Andrew Coaton

Andrew Coaton wrote: > > I am currently using SCO SQL retriever to access > informix SE from a win95 client using a delphi written application. > When the application is reading a large number of records 300,000 it is > painfully slow, the Unix server > HPUX10.20 performance dips only by 10% when retrieving the data, however the > win95 cpu is 100% utilized during the process. I am suspecious of the > SCO SQL retriever having a large over head. > Can anyone help with any suggestions ie alternative > connectivity software . Andrew ODBC was not originally conceived of to haul back vast amounts of data and it is not surprising that there is some performance degradation. I am a little (although not totally) surprised that the client CPU is 100% used during the fetch. Does the same happen in something like MS Access? Sometimes, this sort of overhead is a function of the application, or perhaps a function of the database's indexing structure. It is possible that some of the caching options on SQL-Retriever could be fine-tuned to make a difference. If you would like to provide some more details, I will be happy to have a look at the problem: Version of SQL-Retriever (pc & host) Version of Informix SE Version of Windows95 Could you also send you address & tel/fax details. Allan Gould SCO CID Support, Leeds, UK