> I would have thought the reducing CPU workload would be a downside for IBM.

One might think that.

The reality is that cycle times are not getting any faster. A z14 does not 
execute z10 machine instructions significantly (any?) faster than a z10. One 
way to get customers to buy a new box -- VERY important to IBM -- is to promise 
that their programs will run faster. That only happens if the compiler exploits 
the new instructions. There are "business problems" that a z14 can solve faster 
than a z10, given compiler support of new instructions.

This has been a HUGE area of emphasis of late for the COBOL compiler team. Less 
so I think for C/C++, but still an emphasis.

Charles


-----Original Message-----
From: IBM Mainframe Assembler List [mailto:[email protected]] On 
Behalf Of Jon Perryman
Sent: Tuesday, January 23, 2018 5:11 PM
To: [email protected]
Subject: Re: Fair comparison C vs HLASM

I love the structured programming macro's (SPM). They make the code so much 
more readable, logical and maintainable.  My only suggestion would be logical 
indentation (> and < in place of the label where >>> represents 3 indents).  

As for C optimization, do you think IBM changes it as new instructions are 
implemented? I would have thought the reducing CPU workload would be a downside 
for IBM.

Reply via email to