Please remember that the performance of 3590's with multiple datasets
per cartridge had a problem.
As you said, the problem affected mainly small files. Data to tape is
just moved to the buffer and eventually written to the tape. But a
WRITE TAPE MARK at the end of a file was a synchronizing event, forcing
all buffered data to be written. Then two small records (EOFx) are
written, and another tape mark. So end of file process was slow, and if
the data file is not large, this could be a signficiant part of its
elapsed time.
For larger files, it was not so noticable.
The A60 was enhanced to introduce a non-syncrhonizing WRITE TAPE MARK.
It had to be requested by an option (in the DCBE??) so existing programs
worked as always and could be assured that their data was on the tape.
Programs writing mulitiple files could request the new function.
--
Bruce A. Black
Senior Software Developer for FDR
Innovation Data Processing 973-890-7300
personal: [EMAIL PROTECTED]
sales info: [EMAIL PROTECTED]
tech support: [EMAIL PROTECTED]
web: www.innovationdp.fdr.com
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
