What setup are you talking about? PAUSE/RELEASE are functions that are
called by branching to them. You don't go through any interrupt handler
code. The only setup I'm aware of to use these functions is in acquiring
the Pause Element itself, but once acquired, a token is used to reference.
I don't say this to argue, just trying to understand what is going on.
----- Original Message -----
From: "Binyamin Dissen" <[EMAIL PROTECTED]>
Newsgroups: bit.listserv.ibm-main
To: <IBM-MAIN@BAMA.UA.EDU>
Sent: Tuesday, November 27, 2007 5:03 PM
Subject: Re: SUSPEND/RESUME is slower than WAIT/POST. PAUSE/RELEASE is
slower than both.
On Tue, 27 Nov 2007 16:45:27 -0600 David Day <[EMAIL PROTECTED]> wrote:
:>I don't use SUSPEND/RESUME in my code, but I do use PAUSE/RELEASE. I
:>intentionally coded PAUSE/RELEASE because I was under the impression it
had
:>a shorter path length and was more effecient than POST/WAIT. You say it
was
:>no surprise to you that the opposite was true. When PAUSE/RELEASE was
:>introduced some years back, I think it was 2.6 of os/390, it was touted
by
:>IBM as a superior performer to POST/WAIT. How is it that this was not a
:>surprise?
I would be shocked if pause/release was faster than wait post. There is a
bit
more overhead in the setup. And WAIT/POST are right to the point while
pause/release needs to go thru a level of abstraction, i.e., POST gives
the
direct address of the ECB and if the ECB is waiting, it contains the RB
address.
I am surprised that the SUSPEND/RESUME is slower than WAIT/POST since the
addresses are given as well. Perhaps SRB vs. task was a factor.
:> --Dave Day
:>----- Original Message -----
:>From: "Edward Jaffe" <[EMAIL PROTECTED]>
:>Newsgroups: bit.listserv.ibm-main
:>To: <IBM-MAIN@BAMA.UA.EDU>
:>Sent: Tuesday, November 27, 2007 3:35 PM
:>Subject: Re: SUSPEND/RESUME is slower than WAIT/POST. PAUSE/RELEASE is
:>slower than both.
:>
:>
:>> David Day wrote:
:>>> Ed,
:>>> On the face of it, that doesn't make sense to me. Why would
:>>> pause/release be slower than wait/post? Are you saying the code path
:>>> length is longer? Is there some built-in wait for something within
:>>> pause/release that isn't there in post/wait? Do you have some kind
of
:>>> measurement statistics to validate what you said? I'm not saying you
are
:>>> wrong, I'm just curios about the statement.
:>>>
:>>
:>> Yes. We measured this. The path length for PAUSE/RESUME is longer.
That
:>> was no surprise.
:>>
:>> What we *did* find surprising was that the path length for
SUSPEND/RESUME
:>> in an SRB is longer than simple WAIT/POST in a TCB! We were scratching
our
:>> heads for a couple of weeks trying to understand why some (E)JES code
used
:>> slightly more CPU running in enclave SRB mode redirected to zIIP than
it
:>> did running in TCB mode on ordinary CP. This measurable difference
(about
:>> 7%) turned out to be the reason.
--
Binyamin Dissen <[EMAIL PROTECTED]>
http://www.dissensoftware.com
Director, Dissen Software, Bar & Grill - Israel
Should you use the mailblocks package and expect a response from me,
you should preauthorize the dissensoftware.com domain.
I very rarely bother responding to challenge/response systems,
especially those from irresponsible companies.
----------------------------------------------------------------------
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
----------------------------------------------------------------------
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
