HP OpenVMS Systems

Optimizing application performance?

» close window

The Question is:

 
My application has it own file copy utility.
We have just purchased a couple of DS10's with 30GB IDE drives. The application
 copy utility is much slower on this system than it is when running on a DS10
 with SCSI interface. However a normal VMS copy command is faster on the new
 IDE disk systems.
 
The application uses QIO function io$_readvblk and io$_writevblk to do the
 copying. It will read/write up to 1000 blocks in each I/O operation.
 
Can you please give me some advice as to how I should change the application
 file copy routine to work faster.
I would convert to using VMS copy but the application has a few extra features
 such as being able to copy only parts of the file.
 

The Answer is :

 
  Use the existing callable Convert API to perform the necessary file
  COPY operations, or use a lib$spawn or a client-server connection or
  similar and a DCL COPY command.
 
  As for suggestions for improving your existing application (assuming
  you cannot partition your data into separate files), you will have to
  profile the code via integrated support and/or via tools such as PCA,
  the process quota use and requirements, the I/O completion sequencing
  (particularly if using asynchronous completion), and other similar
  investigative tasks.
 
  Without far more detail, there can be no simple answer to the question.
 

  

     

     
answer written or last revised on ( 18-JUL-2001 )
     
» close window