TRT is NOT fast, generally speaking. For anything less than 256 bytes you are generally better off doing CLI's in a loop.
If you are looking for a number of different characters in a single pass, like blank, comma, period, etc, then TRT is better. If you are looking for anything other than a blank, a CLI loop will still probably be best. I have not tested CLCL for speed, but if it follows the pattern for MVCL, you need to be looking over 8K for it to be as fast as other ways. I would write a simple test routine and see what works better. As for the B/J/BR question, I think (if I correctly remember a conversation I had with the hardware designer) the hardware is optimized for JUMP. Chris Blaicher Phone: 512-340-6154 Mobile: 512-627-3803 -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of McKown, John Sent: Tuesday, February 16, 2010 12:10 PM To: [email protected] Subject: Fastest branch instruction Would an EX of a TRT be worth while? Second: once I have found the blank, what is the fastest way to find the next non-blank, again not going past the end of the buffer, which is in R9. I'm currently using CLCL. Where the registers are set up so that the "source" address is zeros, the length is the length of the remaining buffer, and the "pad" character is 0x40. This makes it scan for a blank without accessing the second memory address. This is on a z9BC, which is now purchased. So we ain't gonna be upgrading anytime soon. John McKown Systems Engineer IV IT ---------------------------------------------------------------------- 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
