DSPOPT parameter - Revision history

ELowell at 17:10, 23 June 2016

2016-06-23T17:10:51Z

← Older revision		Revision as of 17:10, 23 June 2016
Line 35:		Line 35:

	<tr><th align="right"><var>X'04'</var></th>		<tr><th align="right"><var>X'04'</var></th>
	<td>Use compression when swapping servers. (~~New~~ in Model 204 version 7.6)		<td>Use compression when swapping servers. (Introduced in Model 204 version 7.6.)
	<p>		<p>
	This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />		This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />

JAL: add link

2015-07-02T19:44:07Z

add link

← Older revision		Revision as of 19:44, 2 July 2015
Line 20:		Line 20:
	<tr class="head"><th>Settings</th><th>Meaning</th></tr>		<tr class="head"><th>Settings</th><th>Meaning</th></tr>
	<tr><th align="right"><var>X'01' </var></th>		<tr><th align="right"><var>X'01' </var></th>
	<td>Use ~~MVPG (~~Move 4K Page) when swapping users to and from CCASERVR. This option applies when using the CCASERVR in Storage feature. With this option enabled, CCASERVR is allocated as one contiguous block of storage, above the bar in 64-bit systems. The size of CCASERVR is calculated by NUSERS*(largest SERVSIZE) and it is logically divided into a collection of 4K pages.		<td>Use the Assembly Language Move Page (MVPG) facility to move a 4-K page when swapping users to and from CCASERVR. This option applies when using the [[Creating server data sets (CCASERVR)#CCASERVR in Storage feature\|CCASERVR in Storage]] feature. With this option enabled, CCASERVR is allocated as one contiguous block of storage, above the bar in 64-bit systems. The size of CCASERVR is calculated by <code>[[NUSERS parameter\|NUSERS]]*(largest [[SERVSIZE parameter\|SERVSIZE]])</code>, and it is logically divided into a collection of 4K pages.
	<p>		<p>
	When users are server swapped to or from CCASERVR, their server slot in CCASERVR might occupy multiple, non-contiguous blocks of 4K pages. This algorithm tends to “front load” CCASERVR, keeping frequently referenced pages at the beginning and results in reduced operating-system paging. This management strategy reduces the number of 4K page moves ~~and~~ keeps the virtual storage allocated to CCASERVR less fragmented and often requires fewer operating-system-page tables.		When users are server swapped to or from CCASERVR, their server slot in CCASERVR might occupy multiple, non-contiguous blocks of 4K pages. This algorithm tends to “front load” CCASERVR, keeping frequently referenced pages at the beginning, and it results in reduced operating-system paging. This management strategy reduces the number of 4K-page moves, keeps the virtual storage allocated to CCASERVR less fragmented, and often requires fewer operating-system-page tables.
	</p>		</p>
	<p>		<p>

JDamon: Clarified description of DSPOPT=1

2015-07-01T22:12:37Z

Clarified description of DSPOPT=1

← Older revision		Revision as of 22:12, 1 July 2015
Line 20:		Line 20:
	<tr class="head"><th>Settings</th><th>Meaning</th></tr>		<tr class="head"><th>Settings</th><th>Meaning</th></tr>
	<tr><th align="right"><var>X'01' </var></th>		<tr><th align="right"><var>X'01' </var></th>
	<td>~~Allocate space for servers~~ in ~~memory~~ in ~~chunks~~ of 4K pages, ~~not as a permanent~~ contiguous ~~area~~. This ~~allocation lets you move fewer 4K pages~~, keeping the virtual storage allocated ~~for servers in memory~~ less fragmented and ~~possibly using~~ fewer ~~paging~~ tables.		<td>Use MVPG (Move 4K Page) when swapping users to and from CCASERVR. This option applies when using the CCASERVR in Storage feature. With this option enabled, CCASERVR is allocated as one contiguous block of storage, above the bar in 64-bit systems. The size of CCASERVR is calculated by NUSERS*(largest SERVSIZE) and it is logically divided into a collection of 4K pages.
			<p>
			When users are server swapped to or from CCASERVR, their server slot in CCASERVR might occupy multiple, non-contiguous blocks of 4K pages. This algorithm tends to “front load” CCASERVR, keeping frequently referenced pages at the beginning and results in reduced operating-system paging. This management strategy reduces the number of 4K page moves and keeps the virtual storage allocated to CCASERVR less fragmented and often requires fewer operating-system-page tables.
			</p>
	<p>		<p>
	As of Model 204 version 7.5, this setting is the default. (Previously, X'00' was the default.)</p>		As of Model 204 version 7.5, this setting is the default. (Previously, X'00' was the default.)</p>

ELowell: /* Description */

2015-06-12T19:48:42Z

Description

← Older revision		Revision as of 19:48, 12 June 2015
Line 35:		Line 35:
	<p>		<p>
	This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />		This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />
	In addition, if ~~[[PAGEFIX parameter\|~~page fixing]] is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>		In addition, if page fixing is specified for the server swapping memory (the [[PAGEFIX parameter\|PAGEFIX=X’00000004’]] option), only the highwater mark of pages will be fixed.</p>
	</td></tr>		</td></tr>

ELowell: /* Description */

2015-06-12T19:37:05Z

Description

← Older revision		Revision as of 19:37, 12 June 2015
Line 35:		Line 35:
	<p>		<p>
	This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />		This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />
	In addition, if [[~~Performance monitoring and tuning#Page fixing~~\|page fixing]] is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>		In addition, if [[PAGEFIX parameter\|page fixing]] is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>
	</td></tr>		</td></tr>

ELowell: /* Description */

2015-06-05T22:11:14Z

Description

← Older revision		Revision as of 22:11, 5 June 2015
Line 34:		Line 34:
	<td>Use compression when swapping servers. (New in Model 204 version 7.6)		<td>Use compression when swapping servers. (New in Model 204 version 7.6)
	<p>		<p>
	This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.		This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage. <br />
	In addition, if page fixing is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>		In addition, if [[Performance monitoring and tuning#Page fixing\|page fixing]] is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>
	</td></tr>		</td></tr>

DVRICKP at 19:41, 27 January 2015

2015-01-27T19:41:31Z

← Older revision		Revision as of 19:41, 27 January 2015
Line 32:		Line 32:

	<tr><th align="right"><var>X'04'</var></th>		<tr><th align="right"><var>X'04'</var></th>
	<td>Use compression when swapping servers.		<td>Use compression when swapping servers. (New in Model 204 version 7.6)
	<p>		<p>
	This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.		This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.

JAL at 22:00, 25 November 2014

2014-11-25T22:00:13Z

← Older revision		Revision as of 22:00, 25 November 2014
Line 15:		Line 15:

	==Description==		==Description==
	<p>The bits have the following meanings:</p>		<p>
			The bits have the following meanings:</p>
	<table>		<table>
	<tr><th>Settings</th><th>Meaning</th></tr>		<tr class="head"><th>Settings</th><th>Meaning</th></tr>
	<tr><th align="right"><var>X'01' </var></th><td>Allocate space for servers in memory in chunks of 4K pages, not as a permanent contiguous area. This allocation lets you move fewer 4K pages, keeping the virtual storage allocated for servers in memory less fragmented and possibly using fewer paging tables. <p>As of Model 204 version 7.5, this setting is the default. (Previously X'00' was the default.)</p>		<tr><th align="right"><var>X'01' </var></th>
			<td>Allocate space for servers in memory in chunks of 4K pages, not as a permanent contiguous area. This allocation lets you move fewer 4K pages, keeping the virtual storage allocated for servers in memory less fragmented and possibly using fewer paging tables.
			<p>
			As of Model 204 version 7.5, this setting is the default. (Previously, X'00' was the default.)</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'02'</var></th><td>Put servers in a hiperspace instead of a dataspace.
	<p>This setting is applicable only for real dataspaces for non-64-bit systems.</p>		<tr><th align="right"><var>X'02'</var></th>
			<td>Put servers in a hiperspace instead of a dataspace.
			<p>
			This setting is applicable only for real dataspaces for non-64-bit systems.</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'04'</var></th><td>Use compression when swapping servers.
	<p>This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.		<tr><th align="right"><var>X'04'</var></th>
	In addition, if page fixing is specified for the server swapping memory, only the ~~high water~~ mark of pages will be fixed.</p>		<td>Use compression when swapping servers.
			<p>
			This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.
			In addition, if page fixing is specified for the server swapping memory, only the highwater mark of pages will be fixed.</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'20' </var></th><td>Put ~~<var>~~[[Application_Subsystem_development\|APSY]]~~</var>~~ precompiled procedures in hiperspace instead of a dataspace. Hiperspace pages can be in ~~either~~ real or expanded storage.
	<p>This option is not valid on IBM z/VSE and requires that <var class="product">Model 204</var> be APF-authorized.</p>		<tr><th align="right"><var>X'20' </var></th>
			<td>Put [[Application_Subsystem_development\|APSY]] precompiled procedures in hiperspace instead of a dataspace. Hiperspace pages can be in real or expanded storage.
			<p>
			This option is not valid on IBM z/VSE, and it requires that <var class="product">Model 204</var> be APF-authorized.</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'40' </var></th><td>
	<p>Put APSY precompiled procedures in cache hiperspace instead of a scrolling hiperspace. </p>		<tr><th align="right"><var>X'40' </var></th>
	Cache hiperspace pages can reside in ~~only~~ expanded storage, so they are sometimes called ESO (for Expanded Storage Only) hiperspaces.		<td>Put APSY precompiled procedures in cache hiperspace instead of a scrolling hiperspace.
	<p>Also, cache hiperspace pages never result in operating system paging. Pages stolen by the operating system are simply discarded and <var class="product">Model 204</var> is responsible for replacing the pages from the copy of the precompiled procedures in CCATEMP. </p>		<p>
	<p>This option is not valid on IBM z/VSE and requires that <var class="product">Model 204</var> be APF-authorized.</p>		Cache hiperspace pages can reside only in expanded storage, so they are sometimes called <code>ESO</code> (for Expanded Storage Only) hiperspaces. </p>
			<p>
			Also, cache hiperspace pages never result in operating system paging. Pages stolen by the operating system are simply discarded, and <var class="product">Model 204</var> is responsible for replacing the pages from the copy of the precompiled procedures in CCATEMP. </p>
			<p>
			This option is not valid on IBM z/VSE, and it requires that <var class="product">Model 204</var> be APF-authorized.</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'80'</var></th><td>
	<p>Keep all APSY saved pages in CCATEMP.</p>		<tr><th align="right"><var>X'80'</var></th>
	<p>~~Set~~ DSPOPT to X'80' only when <var>[[TEMPPAGE_parameter\|TEMPPAGE]]</var>=0, and then only in unusual circumstances.</p>		<td>Keep all APSY saved pages in CCATEMP.
			<p>
			Set <var>DSPOPT</var> to X'80' only when <var>[[TEMPPAGE_parameter\|TEMPPAGE]]</var> is 0, and then only in unusual circumstances.</p>
	</td></tr>		</td></tr>
	</table>		</table>

	[[Category:System parameters]]		[[Category:System parameters]]
	[[Category:Parameters]]		[[Category:Parameters]]

DVRICKP at 21:07, 25 November 2014

2014-11-25T21:07:31Z

← Older revision		Revision as of 21:07, 25 November 2014
Line 22:		Line 22:
	<tr><th align="right"><var>X'02'</var></th><td>Put servers in a hiperspace instead of a dataspace.		<tr><th align="right"><var>X'02'</var></th><td>Put servers in a hiperspace instead of a dataspace.
	<p>This setting is applicable only for real dataspaces for non-64-bit systems.</p>		<p>This setting is applicable only for real dataspaces for non-64-bit systems.</p>
			</td></tr>
			<tr><th align="right"><var>X'04'</var></th><td>Use compression when swapping servers.
			<p>This setting is applicable only for 64-bit in-storage server swapping. When enabled, each table in the server is individually moved based on its usage.
			In addition, if page fixing is specified for the server swapping memory, only the high water mark of pages will be fixed.</p>
	</td></tr>		</td></tr>
	<tr><th align="right"><var>X'20' </var></th><td>Put <var>[[Application_Subsystem_development\|APSY]]</var> precompiled procedures in hiperspace instead of a dataspace. Hiperspace pages can be in either real or expanded storage.		<tr><th align="right"><var>X'20' </var></th><td>Put <var>[[Application_Subsystem_development\|APSY]]</var> precompiled procedures in hiperspace instead of a dataspace. Hiperspace pages can be in either real or expanded storage.

DCameron: /* Description */

2014-10-16T14:37:17Z

Description

← Older revision		Revision as of 14:37, 16 October 2014
Line 34:		Line 34:
	<tr><th align="right"><var>X'80'</var></th><td>		<tr><th align="right"><var>X'80'</var></th><td>
	<p>Keep all APSY saved pages in CCATEMP.</p>		<p>Keep all APSY saved pages in CCATEMP.</p>
	<p>Set DSPOPT to X'80' only when TEMPPAGE=0, and then only in unusual circumstances.</p>		<p>Set DSPOPT to X'80' only when <var>[[TEMPPAGE_parameter\|TEMPPAGE]]</var>=0, and then only in unusual circumstances.</p>
	</td></tr>		</td></tr>
	</table>		</table>
	[[Category:System parameters]]		[[Category:System parameters]]
	[[Category:Parameters]]		[[Category:Parameters]]