In
77142d37c0c3c34da0d7b1da7d7ca34308822...@nwt-s-mbx2.rocketsoftware.com,
on 02/22/2011
at 04:06 PM, Bill Fairchild bi...@mainstar.com said:
CONTIG and ROUND are/were independent, but they both addressed I/O
performance improvement. One purpose of ROUND was to reduce DASD
revolutions that
allocation mean cylinder boundary?
The two requirements--those of contiguity and rounding to a cylinder
boundary--are/were independent ones.
John Gilmore Ashland, MA 01721-1817 USA
--
For IBM
On Tue, 2011-02-22 at 11:06 -0500, Bill Fairchild wrote:
One purpose of ROUND was to reduce DASD revolutions that did not
transfer any data.
According to my fading memory, Back In The Day a limited set file mask
had to be imposed on CCW strings against non-cylinder-aligned extents.
Hence those
From: hobbitt druidl...@gmail.com
Date: Fri, 4 Feb 2011 08:43:48 -0800 (PST)
Local: Fri, Feb 4 2011 11:43 am
On Feb 1, 10:41 pm, jayare...@hotmail.com (J R) wrote:
The JCL User's Guide is more specific on this. It explicitly states:
if you code ROUND as the last subparameter in
Cylinder allocation (and therefore ROUND used with average block or average
record) allows faster input/output of sequential data sets than does track
allocation.
Unfortunately, since ECKD was introduced, this statement is inaccurate.
If the I/O is within a defined extent, then track/cylinder
You're quite, Ron. It dates back to the days of SLEDs.
Rick
---
Ron Hawkins wrote:
While it may be in the manual, I don't think this statement has been true
for several decades.
Cylinder allocation (and therefore ROUND used
David Andrews wrote:
| Um... comma, comma, contig, round
which appears to be a polite attempted correction of the litany I provided.
If it is, it is a misconceived one. The two requirements--those of contiguity
and rounding to a cylinder boundary--are/were independent ones.
John
ROUND dataset allocation mean cylinder
boundary?
You're quite, Ron. It dates back to the days of SLEDs.
Rick
---
Ron Hawkins wrote:
While it may be in the manual, I don't think this statement has been true
for several decades
On Wed, 2011-02-02 at 13:20 -0500, john gilmore wrote:
| Um... comma, comma, contig, round
which appears to be a polite attempted correction of the litany I provided.
No, you simply reminded me of a relic of the past that I can *still*
find in my proclibs. As you say, it is another problem.
I'm not disagreeing with or correcting anybody. I wanted to point out that I
was just helping out the hobbit with some explicit references to cylinder
allocation via use of the ROUND option.
As much as anything else, I was indicating that one should always read both JCL
volumes, the
I'm not disagreeing with or correcting anybody.
I wanted to point out that I was just helping out the hobbit with some
explicit references to cylinder allocation via use of the ROUND option.
As much as anything else, I was indicating that one should always read both
JCL volumes, the Reference
From: hobbitt druidl...@gmail.com
Date: Tue, 1 Feb 2011 15:37:27 -0800 (PST)
Local: Tues, Feb 1 2011 6:37 pm
If I code a dataset allocation using block size and count, and the ROUND
keyword:
(1024,(100),,,ROUND)
Does the dataset (or at least primary extent) begin on a cylinder
While it may be in the manual, I don't think this statement has been true
for several decades.
Cylinder allocation (and therefore ROUND used with average block or
average
record) allows faster input/output of sequential data sets than does track
allocation.
13 matches
Mail list logo