On Tue, 2010-08-17 at 13:27 -0500, Elpida Tzortzatos. wrote:
The goal of every platform is to move to transparent use of large pages but
that is a very high goal to achieve especially when the large page size is
1MB
(256 contiguous 4k Frames). Implementation of large pages is much simpler
Shane wrote:
On Tue, 2010-08-17 at 13:27 -0500, Elpida Tzortzatos. wrote:
The goal of every platform is to move to transparent use of large pages but
that is a very high goal to achieve especially when the large page size is
1MB
(256 contiguous 4k Frames). Implementation of large pages
I have been following this thread as an interested spectator, not having
had the opportunity to play. Even zLinux has had large pages for a
while now - pity about z/VM ... ahem.
Linux doesn't attempt to break up large pages under memory pressure, but
does allow the number of such pages to be
On Tue, 17 Aug 2010 20:19:04 +1000, Shane ibm-m...@tpg.com.au
wrote:
I have been following this thread as an interested spectator, not having
had the opportunity to play. Even zLinux has had large pages for a
while now - pity about z/VM ... ahem.
Linux doesn't attempt to break up large pages
Right now because of the bug I wrote about in a previous post, it just
takes a lot more
planning than it should have
What was the resolution of the Sev 1? BAD? I can understand something not
working correctly, but I can't understand it not working correctly forever.
Bob Shannon
Rocket
On Mon, 16 Aug 2010 11:27:09 +, Bob Shannon
bshan...@rocketsoftware.com wrote:
Right now because of the bug I wrote about in a previous post, it just
takes a lot more
planning than it should have
What was the resolution of the Sev 1? BAD? I can understand something not
working correctly,
The support to address the issue where 1MB pages were not used for 4K when
there was a real storage shortage for 4K frames was fixed in z/OS 1.12 and
IBM opened APAR OA31116 to ship this fix for z/OS 1.10 and z/OS 1.11.
Today we updated the APAR description to be clear and concise as to what
The impetus for large page exploitation is well documented. I have discussed
this at length at both SHARE and CMG conferences. Also you can reference
the papers below which describe specific success stories for JAVA and z/OS
http://www-
Bob,
As I posted in my previous append, APAR OA31116 and z/OS 1.12 base
support address this issue. Also the severity of this APAR has been raised to a
SEV 1.
I do want to emphasize that even with the APAR, capacity planning needs to
be done before selecting the LFAREA size. The optimal
In a message dated 8/16/2010 12:09:08 P.M. Central Daylight Time,
elp...@us.ibm.com writes:
On the flip side of the coin if you under specify the LFAREA size you may
be
giving up some large page performance benefits.
DOC APAR OA34024 has been opened to provide some guidance on how to
On Mon, 16 Aug 2010 12:51:46 -0500, Elpida Tzortzatos. elp...@us.ibm.com
wrote:
I am working with RMF to open an APAR that would provide at least some of
the information you described below. In addition I am looking at providing a
z/OS DISPLAY command to report on the LFAREA size, amount used,
Sounds like a future health check item.
-Original Message-
From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of
Elpida Tzortzatos.
Sent: Monday, August 16, 2010 12:52 PM
To: IBM-MAIN@bama.ua.edu
Subject: Re: [IBM-MAIN] Anyone Using z10 Large Pages in anger
I
I am working with RMF to open an APAR that would provide at least some of
the information you described below. In addition I am looking at providing a
z/OS DISPLAY command to report on the LFAREA size, amount used, amount
available, and how much has been converted to 4K. A high water mark of
Subject: Re: Anyone Using z10 Large Pages in anger
Sounds like a future health check item.
-Original Message-
From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf
Of Elpida Tzortzatos.
Sent: Monday, August 16, 2010 12:52 PM
To: IBM-MAIN@bama.ua.edu
Subject: Re: [IBM
Thanks, Sounds encouraging. Lots of tuning opportunities.
In a message dated 8/16/2010 12:51:54 P.M. Central Daylight Time,
elp...@us.ibm.com writes:
I am working with RMF to open an APAR that would provide at least some of
the information you described below. In addition I am looking
for
being proactive.
zNorman
-Original Message-
From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On
Behalf
Of Michael Wickman
Sent: Monday, August 16, 2010 Monday 11:05 AM
To: IBM-MAIN@bama.ua.edu
Subject: Re: Anyone Using z10 Large Pages in anger
Sounds like a future
On Mon, 16 Aug 2010 12:09:04 -0500, Elpida Tzortzatos. wrote:
I do want to emphasize that even with the APAR, capacity planning needs to
be done before selecting the LFAREA size. The optimal configuration is when
there is enough in the 4K memory pool to handle the 4K workload and enough
in the
In a message dated 8/16/2010 4:49:09 P.M. Central Daylight Time,
brian.peterson.ibm.m...@comcast.net writes:
WLM for example, to periodically set an LFAREA size appropriate for the
now-running workload. Asking me to predict how many large pages I will
need
for the life of my NEXT IPL is a
...@bama.ua.edu] On Behalf
Of Brian Peterson
Sent: Monday, August 16, 2010 Monday 2:49 PM
To: IBM-MAIN@bama.ua.edu
Subject: Re: Anyone Using z10 Large Pages in anger
On Mon, 16 Aug 2010 12:09:04 -0500, Elpida Tzortzatos. wrote:
I do want to emphasize that even with the APAR, capacity planning needs
On Mon, 16 Aug 2010 15:26:54 -0700, Norman Hollander on DesertWiz wrote:
Very much agree, Brian. You can never predict who will start up another
region or when.
Of course you may know when the phones start ringing with poor performance
to find
your system is paging excessively...
zNorman
I
On Mon, 16 Aug 2010 17:36:51 -0500, Brian Peterson
brian.peterson.ibm.m...@comcast.net wrote:
On Mon, 16 Aug 2010 15:26:54 -0700, Norman Hollander on DesertWiz wrote:
Very much agree, Brian. You can never predict who will start up another
region or when.
Of course you may know when the phones
: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of
Gerhard Adam
Sent: Saturday, August 14, 2010 Saturday 3:52 PM
To: IBM-MAIN@bama.ua.edu
Subject: Re: Anyone Using z10 Large Pages in anger?
Thus there is a hole in the RMF metrics (unless ALL large pages are being
used
- Original Message -
From: Mark Zelden mzel...@flash.net
Newsgroups: bit.listserv.ibm-main
Sent: Saturday, August 14, 2010 10:27 AM
Subject: Re: Anyone Using z10 Large Pages in anger?
On Sat, 14 Aug 2010 10:24:02 +0100, Graham Harris harris...@gmail.com
wrote:
Is anyone actively
We are still using large pages but have been doing it in steps over the
last 6 months (IPL with larger LFAREA, convert more WAS regions to 64-bit).
OK, I give up. Why? What is the benefit versus the cost? Even the original
literature suggested that it could cause performance degradation
On Sun, 15 Aug 2010 12:23:47 -0700, Gerhard Adam gada...@charter.net wrote:
We are still using large pages but have been doing it in steps over the
last 6 months (IPL with larger LFAREA, convert more WAS regions to 64-bit).
OK, I give up. Why? What is the benefit versus the cost? Even the
.
-Original Message-
From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of
Gerhard Adam
Sent: Sunday, August 15, 2010 Sunday 12:24 PM
To: IBM-MAIN@bama.ua.edu
Subject: Re: Anyone Using z10 Large Pages in anger
We are still using large pages but have been doing
Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf
Of Pinnacle
Sent: Sunday, August 15, 2010 Sunday 11:56 AM
To: IBM-MAIN@bama.ua.edu
Subject: Re: Anyone Using z10 Large Pages in anger?
- Original Message -
From: Mark Zelden mzel...@flash.net
Newsgroups: bit.listserv.ibm-main
Sent: Saturday
In theory, getting 1 Storage Obtain for 1M should be less overhead than 256 4K
versions. The virtual storage algorithms to get the data is supposed to be
less (need less levels of hashing).
It is my understanding that the primary benefit is to extend the reach of the
TLB so that effectively,
sometimes you just have to take IBM's word for certain recommendations as they
are very difficult to measure outside a lab. One example that comes quickly
to mind is CPENABLE.
True enough, however perhaps one of the reasons it's difficult to measure
outside the lab is that there is no
Is anyone actively using z10 (or z196!!) Large Pages in any significant way
yet?
As a fallout of some initial testing in our sandpit, I had previously raised
some issues with IBM regarding some missing RMF3/RMF1 metrics for large
pages, mainly in the area of unaccounted-for unused Large Page
On Sat, 14 Aug 2010 10:24:02 +0100, Graham Harris harris...@gmail.com wrote:
Is anyone actively using z10 (or z196!!) Large Pages in any significant way
yet?
Yes. On my clients WebSphere LPARs. First we used a small LFAREA and
migrated some WAS regions to 64-bit. All was fine. Then on a
My current main concern is that once large pages are deployed, a complete
memory map of real memory will no longer be possible until the missing
metric(s) are added (and at what level of z/OS that would take effect/be
retro-fitted). As things stand, this missing 'unused large pages' metric in
It turns out the support didn't work as documented and 1M pages weren't
really used for 4K when there was a shortage! After we opened a Sev 1
PMR with IBM during the problem IBM said something like well, this isn't
really supported.
Mark. Ouch! In my sandpit testing, I had tried to force the
Thus there is a hole in the RMF metrics (unless ALL large pages are being
used) covering unused large pages, and indeed, the total large pages present
on the system.
Actually my thought is also a so what. It seems that the other issues of 4K
to 1M split and coalesce are certainly legitimate
34 matches
Mail list logo
