Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
<?xml version="1.0" encoding="utf-8" ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
< html xmlns = "http://www.w3.org/1999/xhtml" xml:lang = "en" lang = "en" >
< head >
< meta http-equiv = "Content-Type" content = "text/html; charset=utf-8" / >
< meta name = "generator" content = "Docutils 0.6: http://docutils.sourceforge.net/" / >
< title > < / title >
2014-09-26 15:55:55 -05:00
< meta name = "date" content = "26, september 2014" / >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< meta name = "authors" content = "Jean-Paul Chaput" / >
2014-09-26 15:55:55 -05:00
< style type = "text/css" >
html, body, th, td, tr, p, li, h1, h2, h3, h4, h5, h6 {
font-size: 96%;
font-family: "Open Sans", verdana, sans-serif;
}
p, li {
text-align: justify;
}
.sc {
font-variant: small-caps;
font-size: 120%;
}
h1, h2, h3, h4, h5, h6 {
font-family: "Open Sans", verdana, sans-serif;
}
h1 { text-align: center;
border-top: 2px solid #09550b;
border-bottom: 2px solid #09550b;
padding-top: 7pt;
padding-bottom: 7pt;
}
h2, h3, h4, h5, h6 { text-align: left; }
h1, h2, h3 { font-family: "Open Sans";
}
h1 { font-weight: normal; font-size: 170%; padding-top: 7pt; margin-top: 25pt; }
h2 { font-weight: normal; font-size: 140%; padding-top: 7pt; margin-top: 25pt; }
h3 { font-weight: bold; font-size: 118%; padding-top: 7pt; margin-top: 25pt; }
h4 { font-weight: bold; font-size: 100%; }
h5 { font-style: italic; font-size: 100%; }
h6 { font-variant: small-caps; font-size: 100%; }
body {
color: black;
background: white;
/*
background: #09550B;
background-color: white;
*/
background-position: top left;
background-attachment: fixed;
background-repeat: no-repeat;
margin: 0 0 0 0;
padding: 20pt;
width: 550pt;
margin-right: auto;
margin-left: auto;
margin-top: 20pt;
margin-bottom: 20pt;
-moz-box-shadow: 4px 4px 5px 3px #ccc;
-webkit-box-shadow: 4px 4px 5px 3px #ccc;
box-shadow: 4px 4px 5px 3px #ccc;
}
hr {
color: #09550b;
border: 1px dotted #09550b;
border-style: none none dotted;
padding-top: 10pt;
padding-bottom: 10pt;
}
div#contents {
margin: 30pt;
padding: 2pt 10pt;
background-color: #fff676;
-moz-box-shadow: 4px 4px 5px 2px #ccc;
-webkit-box-shadow: 4px 4px 5px 2px #ccc;
box-shadow: 4px 4px 5px 2px #ccc;
/* Shadow explanation:
* The shadow is a rectangle the same size as the box. It is then shifted
* blurred according to the following parameters.
*
* 1. The horizontal offset of the shadow, positive means the shadow will be
* on the right of the box, a negative offset will put the shadow on the
* left of the box.
* 2. The vertical offset of the shadow, a negative one means the box-shadow
* will be above the box, a positive one means the shadow will be below
* the box.
* 3. The blur radius (optional), if set to 0 the shadow will be sharp,
* the higher the number, the more blurred it will be.
* 4. The spread radius (optional), positive values increase the size of the
* shadow, negative values decrease the size. Default is 0 (the shadow is
* same size as blur).
* 5. Color
*/
}
div#centered {
margin-left: auto;
margin-right: auto;
text-align: center;
}
pre, tt, code {
font-family: "courrier", "andale mono", monospace;
font-size: 100%;
white-space: pre;
}
tt {
color: #09550b;
}
pre.wiki, div.code, pre.literal-block {
font-size: 90%;
padding: 5pt;
margin-left: 4%;
margin-right: 4%;
border: dashed;
border-width: thin;
border-color: #FC8676;
background-color: #FCFCE1;
}
a:link, a:active {
font-weight: normal;
text-decoration: none;
color: #09550b;
border-bottom: 1px dotted #09550b;
}
a:hover, a:focus, a:visited {
font-weight: normal;
font-style: italic;
text-decoration: none;
/*
color: #A40010;
border-bottom: 1px dotted #A40010;
*/
color: #09550b;
border-bottom: 1px dotted #09550b;
}
h1 a:link {
border-bottom: 0px;
}
p.credit {
margin-left: 10%;
margin-right: 10%;
font-size: 110%;
}
p.credit span.left {
float: left;
white-space: nowrap;
}
p.credit span.right {
float: right;
white-space: nowrap;
}
img.addborder {
border: 1px solid black;
}
div#header {
margin: 0px;
padding: 0pt;
background-color: white;
display: inline-block;
width: 100%;
}
div#header_logo {
margin: 0px;
padding: 10px 0px 10px 12pt;
background-color: white;
width: 40%;
float: left;
}
div#header_menus {
background-color: white;
width: 55%;
float: right;
padding-top: 60pt;
padding-right: 10pt;
text-align: right;
font-size: 80%;
}
div#header_menus ul {
padding-top: 45pt;
list-style: none;
text-align: right;
font-size: 80%;
}
div#header_menus li {
padding: 0pt;
margin: 0pt;
display: inline;
white-space: nowrap;
}
/*
div#header_menus a {
border-left: 1px solid #d7d7d7;
padding: 0 .75em;
}
div#header_menus a.first {
border-left: none;
}
*/
div#header a:link, div#header a:active, div#header a:visited {
margin: 0pt;
padding: 0pt 5pt;
font-weight: normal;
color: black;
text-decoration: none;
border-bottom: 1px solid black;
border-left: 0px;
border-right: 0px;
}
div#header a:hover, div#header a:focus {
margin: 0pt;
padding: 0pt 5pt;
font-weight: normal;
color: black;
text-decoration: none;
border-bottom: 4px solid #09550b;
border-left: 0px;
border-right: 0px;
}
div#header a.current:link, div#header a.current:active, div#header a.current:visited {
margin: 0pt;
padding: 0pt 5pt;
font-weight: bold;
font-style: normal;
font-size: 120%;
color: white;
text-decoration: none;
border-bottom: 4px solid #09550b;
border-left: 0px;
border-right: 0px;
background-color: #09550b;
}
div#header a.current:hover, div#header a.current:focus {
margin: 0pt;
padding: 0pt 5pt;
font-weight: bold;
font-style: normal;
font-size: 120%;
color: white;
text-decoration: none;
border-bottom: 4px solid #09550b;
border-left: 0px;
border-right: 0px;
background-color: #09550b;
}
div#header_ancestors {
padding: 4px 0px 4px 12pt;
background-color: #09550B;
color: white;
}
div#header_ancestors ul, div#header_ancestors * li {
display: inline;
list-style-type: none;
padding: 0px 0px 0px 0pt;
}
div#header_ancestors a:link, div#header_ancestors a:active, div#header_ancestors a:visited {
font-weight: bold;
color: white;
text-decoration: none;
border-bottom: 0px;
}
div#header_ancestors a:hover, div#header_ancestors a:focus {
font-weight: bold;
color: white;
text-decoration: underline;
}
div#footer {
margin: 0px;
padding: 0px;
border-top: 1px dotted #09550b;
background-color: white;
display: inline-block;
width: 100%;
text-align: right;
}
div#searchform {
width: 80%;
background-color: #ccffcd;
padding: 15pt 10pt 15pt 10pt;
margin-top: 50pt;
margin-bottom: 50pt;
margin-left: auto;
margin-right: auto;
text-align: center;
}
div#searchform input#id_q {
background-color: white;
border: 1px solid #09550b;
padding: 2pt;
width: 80%;
font-size: 110%;
font-weight: bold;
}
span.queryref {
font-weight: bold;
}
div#searchform ul {
list-style: none;
}
div#searchform li {
display: inline;
}
hr#search_vs_results {
color: #09550b;
border: 2px dotted #09550b;
border-style: none none dotted;
margin-top: 0pt;
margin-bottom: 20pt;
}
div#search_results {
width: 85%;
margin: auto;
}
div#sidebar hr#separator {
color: white;
border: 0px;
margin-top: 0pt;
margin-bottom: 20pt;
}
img.footer-logo {
height: 24px;
padding: 0px 2px;
}
hr#site_vs_page {
color: white;
border: 3px dotted white;
border-style: none none dotted;
margin-top: 20pt;
margin-bottom: 20pt;
}
div#sidebar {
/*
background: #09550B;
background: #ccffcd;
*/
background: white;
}
div#sidebar div#sitemenu, div#sidebar div#pagemenu {
/*
background: white;
*/
background: #09550b;
width: 85%;
margin: auto;
padding: 5pt 10pt;
}
div#sidebar * li {
text-align: left;
}
div#sidebar * ul {
list-style-type: square;
padding-left: 12pt;
}
div#sitemenu ul {
list-style-type: none;
padding-left: 0pt;
}
div#sitemenu ul ul {
list-style-type: none;
padding-left: 0pt;
}
div#sitemenu ul ul ul {
list-style-type: square;
padding-left: 12pt;
}
div#sitemenu ul li ul li {
padding-top: 3pt;
padding-bottom: 5pt;
border-top: 1px dotted white;
}
div#sitemenu ul li ul li ul li {
border-top: none;
padding-top: 1pt;
padding-bottom: 1pt;
}
div#sitemenu ul li a:link,
div#sitemenu ul li a:active,
div#sitemenu ul lu a:visited
{
font-size: 140%;
font-weight: bold;
border-bottom: none;
}
div#sitemenu ul li a:focus,
div#sitemenu ul lu a:hover
{
font-size: 140%;
font-weight: bold;
font-style: italic;
border-bottom: none;
}
div#sitemenu ul ul li a:link,
div#sitemenu ul ul li a:active,
div#sitemenu ul ul lu a:visited,
div#sitemenu ul ul li a:focus,
div#sitemenu ul ul lu a:hover
{
font-size: 90%;
font-weight: normal;
border-bottom: none;
/*border-bottom: 1px dotted white;*/
}
div#pagemenu ul {
list-style-type: none;
padding-left: 0pt;
}
div#pagemenu ul ul {
list-style-type: none;
padding-left: 0pt;
}
div#pagemenu ul ul ul {
padding-left: 12pt;
}
div#pagemenu ul li ul li {
padding-top: 3pt;
padding-bottom: 5pt;
border-top: 1px dotted white;
}
div#pagemenu ul li ul li ul li {
border-top: none;
padding-top: 1pt;
padding-bottom: 1pt;
}
div#pagemenu ul li a:link,
div#pagemenu ul li a:active,
div#pagemenu ul lu a:visited
{
font-size: 120%;
font-weight: bold;
border-bottom: none;
}
div#pagemenu ul li a:focus,
div#pagemenu ul lu a:hover
{
font-size: 120%;
font-weight: bold;
font-style: italic;
border-bottom: none;
}
div#pagemenu ul ul li a:link,
div#pagemenu ul ul li a:active,
div#pagemenu ul ul lu a:visited,
div#pagemenu ul ul li a:focus,
div#pagemenu ul ul lu a:hover
{
font-size: 90%;
font-weight: normal;
border-bottom: none;
/*
border-bottom: 1px dotted white;
*/
}
div#sidebar ul.ancestor * li {
padding-top: 0pt;
}
div#sidebar ul {
padding-bottom: 8pt;
}
div#sidebar a:link, div#sidebar a:active, div#sidebar a:visited {
/*
font-weight: normal;
*/
color: white;
text-align: left;
text-decoration: none;
border-bottom: 1px dotted white;
}
div#sidebar a:hover, div#sidebar a:focus {
/*
font-weight: normal;
*/
color: white;
text-align: left;
text-decoration: none;
border-bottom: 1px dotted white;
}
div#main {
border-left: 1px solid #09550b;
}
div#main ul#summary {
list-style-type: square;
padding-left: 12pt;
font-size: 14pt;
}
div#main ul#summary * ul {
padding-left: 12pt;
padding-bottom: 14pt;
font-size: 95%;
}
div.code * {
background-color: #FCFCE1;
}
div.note {
margin: 8px 2% 0px 2%;
border: 1px none #fff01c;
border-left-width: 4px;
border-left-style: solid;
padding: 1px 10pt 1px 55px;
background: #fff676 url('./images/clipboard.png') no-repeat 0% 50%;;
font-size: 90%
}
div.error {
margin: 8px 2% 0px 2%;
border: 1px none #dd0000;
border-left-width: 4px;
border-left-style: solid;
padding: 1px 10pt 1px 55px;
background: #ffddcc url('./images/i-core.png') no-repeat 0% 50%;;
font-size: 90%
}
p.admonition-title {
font-weight: bold;
}
div.topic {
margin: 5pt;
padding: 2pt 10pt;
background-color: fff676;
}
div.topic p.first {
font-weight: bold;
}
table.wiki th, table th {
color: black;
background: #FFFFCC;
}
table.docutils {
margin-left: 10%;
margin-right: 10%;
}
table.wiki, table.wiki th, table.wiki td { border: 1px solid black; }
table.wiki th * p { text-align: center; }
table.wiki * p { margin: 0pt; }
table.wiki * blockquote { margin: 0pt; }
table { border-collapse: collapse; }
table th, table td { border: 1px solid black;
padding: 2px 10px 2px 10px; }
table.docinfo {
margin-top: 10pt;
margin-left: auto;
margin-right: 0pt;
border: 10px solid #303030;
border-collapse: collapse;
background: #303030;
font-size: 90%;
font-family: sans-serif;
}
table.docinfo tr {
border-bottom: 1px dotted white;
}
th.docinfo-name,
table.docinfo td,
table.docinfo td a:link,
table.docinfo td a:active,
table.docinfo td a:visited,
table.docinfo td a:focus,
table.docinfo td a:hover
{
border: 0px solid white;
background: #303030;
color: white;
text-align: left;
font-weight: bold;
}
th.docinfo-name {
font-weight: normal;
}
table.docinfo td {
font-weight: bold;
}
span.ul {
text-decoration: underline;
}
* span.smallcaps {
/*font-variant: "small-caps";*/
text-transform: "uppercase";
font-size: "smaller";
}
span.cb {
font-family: "andale mono", monospace;
font-weight: bold;
white-space: pre;
}
span.fboxtt {
border: 1px solid black;
padding: 0px 4px;
font-family: "andale mono", monospace;
font-weight: bold;
white-space: pre;
}
#notice.system-message, .notice.system-message {
color: black;
background: #DDFFDD;
padding-top: 5pt;
padding-bottom: 5pt;
border: 1px none #55BB55;
border-top-width: 4px;
border-top-style: solid;
}
#content.error .message, div.system-message {
color: #550000;
background: #ffddcc;
border: 2px none #dd0000;
border-top-width: 4px;
border-top-style: solid;
padding: .5em;
margin: 1em 0;
}
#main {
float: right;
width: 70%;
padding: 0pt;
margin: 0pt;
min-height: 700px;
background: white;
}
div#main h1 {
border-bottom: 2px solid #09550b;
}
div#main div.section h1 {
border-bottom: none;
}
#cmscontent {
padding: 0pt 4% 10pt 4%;
margin: 0pt;
}
div#htmlerrorcontents {
padding: 10pt 4% 10pt 4%;
margin: 0pt;
}
div#htmlerrorcontents span.cs {
font-size: 80%;
font-family: "andale mono", monospace;
white-space: pre;
}
div#htmlerrorcontents hr.lang_separator {
border: 1px dotted black;
border-style: none none dotted;
margin-top: 20pt;
margin-bottom: 10pt;
}
#embedcontent {
border: 0pt;
padding: 0pt;
margin: 0pt;
}
#sidebar {
float: left;
width: 29.9%;
padding: 0 0 0 0;
margin: 0 0 0 0;
color: white;
background: #09550B;
/*
min-height: 300px;
background: #f2f2f2;
border-right: 1px solid #ccc;
padding: 0 0 0 10px;
*/
}
< / style >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< / head >
< body >
< div class = "document" >
< table class = "docinfo" frame = "void" rules = "none" >
< col class = "docinfo-name" / >
< col class = "docinfo-content" / >
< tbody valign = "top" >
< tr > < th class = "docinfo-name" > Date:< / th >
2014-09-26 15:55:55 -05:00
< td > 26, september 2014< / td > < / tr >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< tr > < th class = "docinfo-name" > Authors:< / th >
< td > Jean-Paul Chaput< / td > < / tr >
< tr > < th class = "docinfo-name" > Contact:< / th >
< td > < < a class = "reference external" href = "mailto:alliance-users@soc.lip6.fr" > alliance-users@ soc.lip6.fr< / a > > < / td > < / tr >
< tr > < th class = "docinfo-name" > Version:< / th >
2014-09-26 15:55:55 -05:00
< td > 0.2< / td > < / tr >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< / tbody >
< / table >
<!-- - * - Mode: rst - * - -->
<!-- - * - Mode: rst - * - -->
<!-- URLs that changes between the various backends. -->
<!-- For HTML backend -->
<!-- Stand - alone images. -->
<!-- Direct LaTeX commands encapsulation. -->
<!-- - * - Mode: rst - * - -->
<!-- Acronyms & names. -->
<!-- Tools -->
<!-- RDS file syntax. -->
< p > < span class = "raw-html" > < br > < / span > < / p >
< p > < strong > Disclaimer:< / strong > This document is still far from complete.< / p >
< p > < span class = "raw-html" > < br > < / span > < / p >
2014-09-26 15:55:55 -05:00
< div class = "section" id = "symbolic-to-real-conversion-in-alliance" >
< h1 > < a class = "toc-backref" href = "#id1" > Symbolic to Real Conversion In Alliance< / a > < / h1 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< div class = "contents topic" id = "contents" >
< p class = "topic-title first" > Contents< / p >
< ul class = "simple" >
2014-09-26 15:55:55 -05:00
< li > < a class = "reference internal" href = "#symbolic-to-real-conversion-in-alliance" id = "id1" > Symbolic to Real Conversion In Alliance< / a > < ul >
< li > < a class = "reference internal" href = "#symbolic-layout" id = "id2" > Symbolic Layout< / a > < ul >
< li > < a class = "reference internal" href = "#symbolic-components" id = "id3" > Symbolic Components< / a > < / li >
< li > < a class = "reference internal" href = "#symbolic-segments" id = "id4" > Symbolic Segments< / a > < / li >
< / ul >
< / li >
< li > < a class = "reference internal" href = "#the-rds-file" id = "id5" > The RDS File< / a > < ul >
< li > < a class = "reference internal" href = "#physical-grid-lambda-value" id = "id6" > Physical Grid & Lambda Value< / a > < / li >
< li > < a class = "reference internal" href = "#the-mbk-to-rds-segment-table" id = "id7" > The < tt class = "docutils literal" > MBK_TO_RDS_SEGMENT< / tt > table< / a > < / li >
< li > < a class = "reference internal" href = "#the-mbk-to-rds-via-table" id = "id8" > The < tt class = "docutils literal" > MBK_TO_RDS_VIA< / tt > table< / a > < / li >
< li > < a class = "reference internal" href = "#the-mbk-to-rds-bigvia-hole-table" id = "id9" > The < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_HOLE< / tt > table< / a > < / li >
< li > < a class = "reference internal" href = "#the-mbk-to-rds-bigvia-metal-table" id = "id10" > The < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_METAL< / tt > table< / a > < / li >
< li > < a class = "reference internal" href = "#the-mbk-wiresetting-table" id = "id11" > The < tt class = "docutils literal" > MBK_WIRESETTING< / tt > table< / a > < / li >
< / ul >
< / li >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< / ul >
< / li >
< / ul >
< / div >
< p > < span class = "raw-html" > < br > < / span > < / p >
< p > < span class = "raw-html" > < br > < / span > < / p >
2014-09-26 15:55:55 -05:00
< div class = "section" id = "symbolic-layout" >
< h2 > < a class = "toc-backref" href = "#id2" > Symbolic Layout< / a > < / h2 >
< div class = "section" id = "symbolic-components" >
< h3 > < a class = "toc-backref" href = "#id3" > Symbolic Components< / a > < / h3 >
< p > A symbolic layout is, in practice, made of only of three objects:< / p >
< table border = "1" class = "docutils" >
< colgroup >
< col width = "30%" / >
< col width = "13%" / >
< col width = "57%" / >
< / colgroup >
< thead valign = "bottom" >
< tr > < th class = "head" > Object< / th >
< th class = "head" > < span class = "sc" > mbk< / span > < / th >
< th class = "head" > Explanation< / th >
< / tr >
< / thead >
< tbody valign = "top" >
< tr > < td > Segments< / td >
< td > < tt class = "docutils literal" > phseg< / tt > < / td >
< td > Oriented segments with a width and an orientation.< / td >
< / tr >
< tr > < td > VIAs & contacts< / td >
< td > < tt class = "docutils literal" > phvia< / tt > < / td >
< td > Boils down to just a point.< / td >
< / tr >
< tr > < td > Big VIAs & Big Contacts< / td >
< td > < tt class = "docutils literal" > phvia< / tt > < / td >
< td > Point with a width and a height
That is a rectangle of width by height centered
on the VIA coordinates.< / td >
< / tr >
< / tbody >
< / table >
< p > Each of thoses objects is associated to a < em > symbolic layer< / em > which will
control how the object is translated in many < em > real rectangles< / em > .< / p >
< table border = "1" class = "docutils" >
< colgroup >
< col width = "11%" / >
< col width = "19%" / >
< col width = "16%" / >
< col width = "54%" / >
< / colgroup >
< thead valign = "bottom" >
< tr > < th class = "head" > < span class = "sc" > mbk< / span > < / th >
< th class = "head" > Layer Name< / th >
< th class = "head" > Usable By< / th >
< th class = "head" > Usage< / th >
< / tr >
< / thead >
< tbody valign = "top" >
< tr > < td rowspan = "12" > < tt class = "docutils literal" > phseg< / tt > < / td >
< td > < tt class = "docutils literal" > NWELL< / tt > < / td >
< td > Segment< / td >
< td > N Well< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > PWELL< / tt > < / td >
< td > Segment< / td >
< td > P Well< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > NDIF< / tt > < / td >
< td > Segment< / td >
< td > N Diffusion< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > PDIF< / tt > < / td >
< td > Segment< / td >
< td > P Diffusion< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > NTIE< / tt > < / td >
< td > Segment< / td >
< td > N Tie< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > PTIE< / tt > < / td >
< td > Segment< / td >
< td > P Tie< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > NTRANS< / tt > < / td >
< td > Segment< / td >
< td > N transistor, in < span class = "sc" > Alliance< / span > , a transistor
is represented as a segment (it's grid).< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > PTRANS< / tt > < / td >
< td > Segment< / td >
< td > P transistor< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > POLY< / tt > < / td >
< td > Segment< / td >
< td > Polysilicium< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > ALUx< / tt > < / td >
< td > Segment< / td >
< td > Metal level < em > x< / em > < / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CALUx< / tt > < / td >
< td > Segment< / td >
< td > Metal level < em > x< / em > , that can be used by the
upper hierarchical level as a connector.
From the layout point of view it is the
same as < tt class = "docutils literal" > ALUx< / tt > .< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > TALUx< / tt > < / td >
< td > Segment< / td >
< td > Blockage for metal level < em > x< / em > . Will
diseappear in the real layout as it is an
information for the P& R tools only.< / td >
< / tr >
< tr > < td rowspan = "9" > < tt class = "docutils literal" > phvia< / tt > < / td >
< td > < tt class = "docutils literal" > CONT_BODY_N< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact to N Well< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_BODY_P< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact to P Well< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_DIF_N< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact to N Diffusion< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_DIF_P< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact to P Diffusion< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_POLY< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact to polysilicium< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_VIA< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact between metal1 and metal2< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > CONT_VIAx< / tt > < / td >
< td > VIA, BIGVIA< / td >
< td > Contact between metal < em > x< / em > and metal < em > x+1< / em > .
The index is the the one of the bottom
metal of the VIA.< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > C_X_N< / tt > < / td >
< td > VIA< / td >
< td > N transistor corner, to build transistor
bend. Not used anymore in recent technos< / td >
< / tr >
< tr > < td > < tt class = "docutils literal" > C_X_P< / tt > < / td >
< td > VIA< / td >
< td > P transistor corner, to build transistor
bend. Not used anymore in recent technos< / td >
< / tr >
< / tbody >
< / table >
< div class = "note" >
< p class = "first admonition-title" > Note< / p >
< p class = "last" > Not all association of object and symbolic layers are meaningful.
For instance you cannot associate a contact to a < tt class = "docutils literal" > NTRANS< / tt > layer.< / p >
< / div >
< div class = "note" >
< p class = "first admonition-title" > Note< / p >
< p class = "last" > The symbolic layer associated with blockages is prefixed by a < tt class = "docutils literal" > T< / tt > ,
for < em > transparency< / em > , which may seems silly. It is for historical reasons,
it started as a true transparency, but at some point we had to invert
the meaning (blockage) with the rise of over-the-cell routing, but the
name stuck...< / p >
< / div >
< / div >
< div class = "section" id = "symbolic-segments" >
< h3 > < a class = "toc-backref" href = "#id4" > Symbolic Segments< / a > < / h3 >
< p > In < span class = "sc" > Alliance< / span > , segments are oriented (up, down, left, right). This disambiguate
the left or right side when using the < tt class = "docutils literal" > LCW< / tt > and < tt class = "docutils literal" > RCW< / tt > rules in the < span class = "sc" > rds< / span > file.
It allows to generate, if needed, asymetric object in the real layout file.< / p >
< p > < span class = "raw-html" > < center > < img src = "./images/SegmentOrientation.png" alt = "Symbolic Segment Orientations" > < / center > < / span > < / p >
< / div >
< / div >
< div class = "section" id = "the-rds-file" >
< h2 > < a class = "toc-backref" href = "#id5" > The RDS File< / a > < / h2 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > The RDS file control how a symbolic layout is transformed into it's real
conterpart.< / p >
< div class = "note" >
< p class = "first admonition-title" > Note< / p >
< p class = "last" > < strong > Unit used inside the RDS file:< / strong > all units are expressed in micrometers.< / p >
< / div >
< p > Alliance tools relying on the RDS file, and what layers are active for them:< / p >
< table border = "1" class = "docutils" >
< colgroup >
< col width = "47%" / >
< col width = "16%" / >
< col width = "37%" / >
< / colgroup >
< thead valign = "bottom" >
< tr > < th class = "head" > Tool< / th >
< th class = "head" > Name< / th >
< th class = "head" > RDS Flags< / th >
< / tr >
< / thead >
< tbody valign = "top" >
< tr > < td > Layout editor< / td >
< td > < tt class = "docutils literal" > graal< / tt > < / td >
< td > < tt class = "docutils literal" > ALL< / tt > < / td >
< / tr >
< tr > < td > Design Rule Checker< / td >
< td > < tt class = "docutils literal" > druc< / tt > < / td >
< td > < tt class = "docutils literal" > ALL< / tt > , < tt class = "docutils literal" > DRC< / tt > < / td >
< / tr >
< tr > < td > Electrical extractor< / td >
< td > < tt class = "docutils literal" > cougar< / tt > < / td >
< td > < tt class = "docutils literal" > ALL< / tt > , < tt class = "docutils literal" > EXT< / tt > < / td >
< / tr >
< tr > < td > The symbolic to real layout translator< / td >
< td > < tt class = "docutils literal" > s2r< / tt > < / td >
< td > < tt class = "docutils literal" > ALL< / tt > < / td >
< / tr >
< / tbody >
< / table >
< div class = "section" id = "physical-grid-lambda-value" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id6" > Physical Grid & Lambda Value< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > RDS file:< / p >
< pre class = "literal-block" >
DEFINE PHYSICAL_GRID 0.005
DEFINE LAMBDA 0.09
< / pre >
< p > Tells that the physical grid (founder grid) step is 0.005µm and the lambda has
a value of 0.09µm. That is, one lambda is 18 grid steps.< / p >
< p > We can distinguish two kind of < span class = "sc" > rds< / span > files:< / p >
< ul class = "simple" >
< li > The < em > 1µm< / em > kind, odd segment widths and coordinates are allowed, but the < tt class = "docutils literal" > LAMBDA< / tt >
value < strong > must< / strong > represent an < em > even< / em > number of foundry grid step.< / li >
< li > The < em > 2µm< / em > kind, segments widths and coordinates must all be even. And in that case
the < tt class = "docutils literal" > LAMBDA< / tt > value can be any multiple of the foundry grid.< / li >
< / ul >
< / div >
< div class = "section" id = "the-mbk-to-rds-segment-table" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id7" > The < tt class = "docutils literal" > MBK_TO_RDS_SEGMENT< / tt > table< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > The < tt class = "docutils literal" > MBK_TO_RDS_SEGMENT< / tt > table control the way segments are translated into
real rectangles. Be aware that we are translating < em > segments< / em > and not < em > rectangles< / em > .
Segments are defined by their axis (source & target points) and their width.
The geometrical transformations are described according to that model.
Obviously, they are either horizontal or vertical.< / p >
< p > The translation method of a symbolic segment is as follow:< / p >
< ol class = "arabic" >
< li > < p class = "first" > The segment is translated into one or more physical rectangles.
The generated rectangles depends on the tool which is actually
using < span class = "sc" > rds< / span > and the flag for the considered real layer.
For instance, real layers flagged with < tt class = "docutils literal" > DRC< / tt > will be generated
for < tt class = "docutils literal" > s2r< / tt > (for the < tt class = "docutils literal" > cif< / tt > or < tt class = "docutils literal" > gds< / tt > ) and < tt class = "docutils literal" > druc< / tt > , but will not
be shown under < tt class = "docutils literal" > graal< / tt > .< / p >
< / li >
< li > < p class = "first" > Translation into one real layer. < em > First< / em > the source & target coordinates and width
of the symbolic segment are multiplied by the < tt class = "docutils literal" > LAMBDA< / tt > value to obtain a real
segment. < em > Then< / em > one of the < tt class = "docutils literal" > VW< / tt > , < tt class = "docutils literal" > LCW< / tt > or < tt class = "docutils literal" > RCW< / tt > transformation is applied to
that segment to get the final real rectangle.< / p >
< ul >
< li > < p class = "first" > < tt class = "docutils literal" > VW< / tt > for Variable Width, expand the real layer staying centered from the
original one. In those rules, the third number is not used, it is only here
to make the life easier for the parser...< / p >
< p > < span class = "raw-html" > < center > < img src = "./images/RDS_VW.png" alt = "RDS Variable Width Rule" > < / center > < / span > < / p >
< / li >
< li > < p class = "first" > < tt class = "docutils literal" > LCW< / tt > or < tt class = "docutils literal" > RCW< / tt > for Left/Right Constant Width, create an off-center rectangle
of fixed width relatively to the real segment. Note that the < tt class = "docutils literal" > SP< / tt > number
is the distance < em > between the edge< / em > of the real segment and the edge of the
generated real rectangle (< em > not< / em > from the axis). It is often zero.< / p >
< p > < span class = "raw-html" > < center > < img src = "./images/RDS_LCW.png" alt = "RDS Left Constant Width Rule" > < / center > < / span > < / p >
< / li >
< / ul >
< / li >
< / ol >
< p > < span class = "raw-html" > < br > < / span > < / p >
< p > Examples:< / p >
< pre class = "literal-block" >
TABLE MBK_TO_RDS_SEGMENT
# (Case 1)
ALU1 RDS_ALU1 VW 0.18 0.09 0.0 ALL
# (Case 2)
NDIF RDS_NDIF VW 0.18 0.0 0.0 ALL \
RDS_ACTIV VW 0.18 0.0 0.0 DRC \
RDS_NIMP VW 0.36 0.36 0.0 DRC
# (Case 3)
NTRANS RDS_POLY VW 0.27 0.00 0.0 ALL \
RDS_GATE VW 0.27 0.00 0.0 DRC \
RDS_NDIF LCW 0.0 0.27 0.0 EXT \
RDS_NDIF RCW 0.0 0.27 0.0 EXT \
RDS_NDIF VW 0.0 0.72 0.0 DRC \
RDS_ACTIV VW 0.0 0.72 0.0 ALL \
RDS_NIMP VW 0.18 1.26 0.0 DRC
END
< / pre >
< p > < span class = "fboxtt" > Case 1< / span > the < tt class = "docutils literal" > ALU1< / tt > is translated in exacltly one real rectangle of
< tt class = "docutils literal" > RDS_ALU1< / tt > , both ends are extended by 0.18µm and it's width is increased
by 0.09µm.< / p >
< p > < span class = "fboxtt" > Case 2< / span > the < tt class = "docutils literal" > NDIF< / tt > will be translated into only one segment
under < tt class = "docutils literal" > graal< / tt > , for symbolic visualization. And into three real rectangles
for < tt class = "docutils literal" > s2r< / tt > and < tt class = "docutils literal" > druc< / tt > .< / p >
< p > < span class = "fboxtt" > Case 3< / span > the < tt class = "docutils literal" > NTRANS< / tt > , associated to a transistor is a little bit
more complex, the generated shapes are different for the extractor < tt class = "docutils literal" > cougar< / tt >
in one hand, and for both < tt class = "docutils literal" > druc< / tt > & < tt class = "docutils literal" > s2r< / tt > in the other hand.< / p >
< ul >
< li > < p class = "first" > For the extractor (< tt class = "docutils literal" > EXT< / tt > & < tt class = "docutils literal" > ALL< / tt > flags) there will be four rectangles
generateds:< / p >
< ol class = "arabic simple" >
< li > The gate (< tt class = "docutils literal" > RDS_GATE< / tt > )< / li >
< li > The left diffusion of the transistor (source or drain) (< tt class = "docutils literal" > RDS_NDIF< / tt > ).< / li >
< li > The right diffusion of the transistor (drain or source) (< tt class = "docutils literal" > RDS_NDIF< / tt > ).< / li >
< li > The active area (< tt class = "docutils literal" > RDS_ACTIV< / tt > ).< / li >
< / ol >
< p > As the extractor must kept separate the source and the drain of the transistor,
they are generated as two offset rectangles, using the < tt class = "docutils literal" > LCW< / tt > and < tt class = "docutils literal" > RCW< / tt > directives.< / p >
< / li >
< li > < p class = "first" > For < tt class = "docutils literal" > s2r< / tt > and < tt class = "docutils literal" > druc< / tt > (< tt class = "docutils literal" > DRC< / tt > and < tt class = "docutils literal" > ALL< / tt > ), five rectangles are generateds:< / p >
< ol class = "arabic simple" >
< li > The poly (< tt class = "docutils literal" > RDS_POLY< / tt > ).< / li >
< li > The gate (< tt class = "docutils literal" > RDS_GATE< / tt > ).< / li >
< li > The diffusion, as one rectangle that covers both the < tt class = "docutils literal" > LCW< / tt > and the < tt class = "docutils literal" > RCW< / tt > (< tt class = "docutils literal" > RDS_NDIF< / tt > ).< / li >
< li > The active area (< tt class = "docutils literal" > RDS_ACTIV< / tt > ).< / li >
< li > The N implantation (< tt class = "docutils literal" > RDS_NIMP< / tt > ).< / li >
< / ol >
< p > In the layout send to the foundry, the source & drain are draw as one rectangle
across the gate area (the transistor being defined by the intersection of both
rectangles).< / p >
< / li >
< / ul >
< p > < span class = "raw-html" > < br > < / span > < / p >
< / div >
< div class = "section" id = "the-mbk-to-rds-via-table" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id8" > The < tt class = "docutils literal" > MBK_TO_RDS_VIA< / tt > table< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > This table is to translate < em > default< / em > VIAs into real via. In the symbolic layout
the default VIA is simply a point and a set of layers. All layers are converted
in squares shapes centered on the VIA coordinate. The one dimension given is the
size of the side of that square.< / p >
< p > Note that although we are refering to VIAs, which for the purists are between two
metal layers, this table also describe < em > contacts< / em > .< / p >
< p > Example:< / p >
< pre class = "literal-block" >
TABLE MBK_TO_RDS_VIA
CONT_DIF_P RDS_PDIF 0.54 ALL \
RDS_CONT 0.18 ALL \
RDS_ALU1 0.36 ALL \
RDS_ACTIV 0.54 DRC \
RDS_PIMP 0.90 DRC
CONT_POLY RDS_POLY 0.54 ALL \
RDS_CONT 0.18 ALL \
RDS_ALU1 0.36 ALL
CONT_VIA RDS_ALU1 0.45 ALL \
RDS_VIA1 0.27 ALL \
RDS_ALU2 0.45 ALL
END
< / pre >
< div class = "note" >
< p class = "first admonition-title" > Note< / p >
< p class = "last" > < strong > In CONT_DIF_P< / strong > you may see that only three layers will be shown under
< tt class = "docutils literal" > graal< / tt > , but five will be generated in the < tt class = "docutils literal" > gds< / tt > layout.< / p >
< / div >
< / div >
< div class = "section" id = "the-mbk-to-rds-bigvia-hole-table" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id9" > The < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_HOLE< / tt > table< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > In < tt class = "docutils literal" > s2r< / tt > , when generating BIGVIAs, the matrix of holes they contains is
not draw relative to the position of the BIGVIA itself, but on a grid which
is common througout all the design real layout. This is to allow overlap
between two BIGVIA without risking the holes matrix to be not exactly overlapping.
As a consequence, when visualizing the < tt class = "docutils literal" > gds< / tt > file, the holes may not be centerend
inside one individual BIGVIA.< / p >
< p > The < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_HOLE< / tt > table define the global hole matrix for the whole
design. The first number is the individual hole side and the second the grid step
2014-09-26 15:55:55 -05:00
(edge to edge). The figure below show the hole generation.< / p >
< p > < span class = "raw-html" > < center > < img src = "./images/bigvia-1.png" alt = "BIGVIA holes" > < / center > < / span > < / p >
< p > Example of BIGVIA overlap:< / p >
< p > < span class = "raw-html" > < center > < img src = "./images/bigvia-2.png" alt = "BIGVIA holes overlap" > < / center > < / span > < / p >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > Example:< / p >
< pre class = "literal-block" >
TABLE MBK_TO_RDS_BIGVIA_HOLE
CONT_VIA RDS_VIA1 0.27 0.27 ALL
CONT_VIA2 RDS_VIA2 0.27 0.27 ALL
CONT_VIA3 RDS_VIA3 0.27 0.27 ALL
CONT_VIA4 RDS_VIA4 0.27 0.27 ALL
CONT_VIA5 RDS_VIA5 0.36 0.36 ALL
END
< / pre >
2014-09-26 15:55:55 -05:00
< div class = "note" >
< p class = "first admonition-title" > Note< / p >
< p class = "last" > < strong > BIGVIA demotion.< / strong > If the size of the bigvia is too small, there is
a possibility that no hole from the global matrix will be under it.
To avoid that case, if the either side of the BIGVIA is less than
< tt class = "docutils literal" > 1.5 * step< / tt > , the BIGVIA is demoted to a simple VIA.< / p >
< / div >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > < span class = "raw-html" > < br > < / span > < / p >
< / div >
< div class = "section" id = "the-mbk-to-rds-bigvia-metal-table" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id10" > The < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_METAL< / tt > table< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > This table describe how the metal part of a BIGVIA is expanded (for the hole
part, see the previous table < tt class = "docutils literal" > MBK_TO_RDS_BIGVIA_HOLE< / tt > ). The rule give for each
metal:< / p >
< ol class = "arabic simple" >
< li > The < em > delta-with< / em > (have to ask Franck).< / li >
< li > The < em > overhang< / em > , the length the real rectangle is expanded on each side from
the symbolic rectange.< / li >
< / ol >
< p > Example:< / p >
< pre class = "literal-block" >
TABLE MBK_TO_RDS_BIGVIA_METAL
CONT_VIA RDS_ALU1 0.0 0.09 ALL \
RDS_ALU2 0.0 0.09 ALL
CONT_VIA2 RDS_ALU2 0.0 0.09 ALL \
RDS_ALU3 0.0 0.09 ALL
CONT_VIA3 RDS_ALU3 0.0 0.09 ALL \
RDS_ALU4 0.0 0.09 ALL
CONT_VIA4 RDS_ALU4 0.0 0.09 ALL \
RDS_ALU5 0.0 0.09 ALL
CONT_VIA5 RDS_ALU5 0.0 0.09 ALL \
RDS_ALU6 0.0 0.18 ALL
END
< / pre >
< p > < span class = "raw-html" > < br > < / span > < / p >
< / div >
< div class = "section" id = "the-mbk-wiresetting-table" >
2014-09-26 15:55:55 -05:00
< h3 > < a class = "toc-backref" href = "#id11" > The < tt class = "docutils literal" > MBK_WIRESETTING< / tt > table< / a > < / h3 >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< p > From a strict standpoint this table shouldn't be here but put in a separate
configuration file, because it contains informations only used by the symbolic
layout tools (< tt class = "docutils literal" > ocp< / tt > , < tt class = "docutils literal" > nero< / tt > , < tt class = "docutils literal" > ring< / tt > ).< / p >
< p > This table defines the cell gauge the routing pitch and minimal (symbolic)
wire width and minimal spacing for the routers. They are patly redundant.< / p >
< p > Example:< / p >
< pre class = "literal-block" >
TABLE MBK_WIRESETTING
X_GRID 10
Y_GRID 10
Y_SLICE 100
WIDTH_VDD 12
WIDTH_VSS 12
TRACK_WIDTH_ALU8 0
TRACK_WIDTH_ALU7 4
TRACK_WIDTH_ALU6 4
TRACK_WIDTH_ALU5 4
TRACK_WIDTH_ALU4 3
TRACK_WIDTH_ALU3 3
TRACK_WIDTH_ALU2 3
TRACK_WIDTH_ALU1 3
TRACK_SPACING_ALU8 0
TRACK_SPACING_ALU7 4
TRACK_SPACING_ALU6 4
TRACK_SPACING_ALU5 4
TRACK_SPACING_ALU4 4
TRACK_SPACING_ALU3 4
TRACK_SPACING_ALU2 4
TRACK_SPACING_ALU1 3
END
< / pre >
< / div >
< / div >
< / div >
2014-09-26 15:55:55 -05:00
< / div >
Bug fixes for MOSIS SCMOS_DEEP support.
* Change: In Hurricane, the NetRoutingProperty is moved into Hurricane
from Katabatic. Needed for Knik to be able to access thoses
informations.
* Change: In Hurricane, in RoutingPad::setOnBestComponent(), now in
case of identical area, select the component of lowest id.
This should not be needed if the component ordering was fully
deterministic as it should be (will investigate later).
This is to ensure that the choosen component is always the
same, especially between save/load of a global routing.
* Bug: In Katabatic, in AutoContactHTee::updateTopology(), invalidate
the segments only if the topology is valid (no NULL in the
cached segments).
* Bug: In Katabatic, in GCellTopology::construct(), throw an error
if the topology is bad instead of trying to continue (and core
dump later... ).
* Bug: In Kite, in BuildPowerRails, distinguish the name of the master
net in the pad (for vddi, vssi, vdde, vsse, ck, cki & cko) and the
name of the net in the *chip* netlist. Must use the later to make
comparison as they may differs.
* Change: In Knik, in save/load solution, exclude nets that are not
globally routed by Knik. That is which NetRoutingProperty is not
*Automatic*.
* Bug: In Cumulus, in chip.BlockPower take account of the layer
width extention to sligthy shrink the connector thus avoiding a
notch with standart cell in some cases.
* Change: In Cumulus, in chip.ClockTree disable the use of fixed Steiner
trees for the leaf clocks, as it seems overconstrained for the
router. First move was to lower them in M2/M3 (instead of M3/M4)
but that was not sufficent.
* New: In Cumulus, RSavePlugin for recursively saving a physical
hierarchy.
* New: In documentation, first embryo for RDS file. Should have been
in Alliance git, but I prefer to keep newest doc in Coriolis.
2014-09-21 09:44:37 -05:00
< / body >
< / html >
