coriolis/documentation/content/pages/python-tutorial/CellNetComponent.rst

339 lines
12 KiB
ReStructuredText
Raw Normal View History

Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
.. -*- Mode: rst -*-
.. include:: ../etc/definitions.rst
.. include:: ./definitions.rst
3. Making a Standard Cell -- Layout
===================================
In this part, we will show how to create and save a terminal Cell_,
that is, a cell without instances (the end point of a hierarchical
design). To illustrate the case we will draw the layout of a
standard cell.
We will introduce the following classes : Cell_, Net_, Component_
and its derived classes.
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
3.1 The AllianceFramework (CRL Core)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The |Hurricane| database only manages objects in memory. To load or save
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
something from the outside, we need to use a *framework*. As of today, only
one is available : the Alliance framework. It allows |Coriolis| to handle
|Alliance| libraries and cells in the exact same way.
.. note:: To see how the AllianceFramework_ is configured for your
installation, please have a look to ``alliance.conf`` in the
``etc/coriolis2`` directory. It must contains the same settings
as the various ``MBK_`` variables used for |Alliance|.
3.2 Session Mechanism (Hurricane)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In the |Hurricane| database, all modifications must take place inside
an UpdateSession_. At the closing of a session, created or modificated
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
objects are fully inserted in the database. This is especially true for
the visualisation, a created component will be visible *only* only after
the session close.
.. note:: See ``QuadTree`` and ``Query``.
3.3 Creating a new Cell (CRL Core)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The creation of a new Cell_ occurs through the AllianceFramework_,
and, as stated above, inside a UpdateSession_. The AllianceFramework_
environment is provided by the |CRL| module.
.. code-block:: Python
from Hurricane import *
from CRL import *
af = AllianceFramework.get()
UpdateSession.open()
cell = af.createCell( 'my_inv' )
# Build then save the Cell.
UpdateSession.close()
This is the simplest call to ``createCell()``, and in that case, the newly
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
created Cell_ will be saved in the *working library* (usually, the current
directory). You may supply a second argument telling into which library
you want the Cell_ to be created.
In the |Hurricane| Cell_ object, there is no concept of *view*, it contains
completly fused logical (netlist) and physical (layout) views.
3.4 The DbU Measurement Unit
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
All coordinates in the |Hurricane| database are expressed in DbU_
(for *Database Unit*) which are integer numbers of foundry grid.
To be more precise, they are fixed points numbers expressed in
hundredth of foundry grid (to allow transient non-integer
computation).
To work with symbolic layout, that is, using lambda based lengths,
two conversion functions are provided:
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
* ``unit = DbU.fromLambda( lbd )`` convert a lambda :cb:`lbd` into a ``DbU``.
* ``lbd = DbU.toLambda( unit )`` convert a ``DbU`` into a lambda :cb:`lbd`.
In the weakly typed |Python| world, :cb:`lbd` is *float* while :cb:`unit`
is *integer*.
3.5 Setting up the Abutment Box
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To setup the abutment box, we use a Box_ which defines a box from
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
the coordinates of the lower left corner ``(x1,y1)`` and upper left
corner ``(x2,y2)``.
.. code-block:: Python
b = Box( DbU.fromLambda( 0.0) # x1
, DbU.fromLambda( 0.0) # y1
, DbU.fromLambda(15.0) # x2
, DbU.fromLambda(50.0) ) # y2
cell.setAbutmentBox( b )
Or more simply:
.. code-block:: Python
cell.setAbutmentBox( Box( DbU.fromLambda( 0.0)
, DbU.fromLambda( 0.0)
, DbU.fromLambda(15.0)
, DbU.fromLambda(50.0) ) )
3.6 Adding Nets and Components
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In the |Hurricane| terminology, a **component** is any kind of physical object
among:
* Contact_
* Pad_
* RoutingPad_
* Horizontal_
* Vertical_
* Plug_ is the only exception and will be detailed later (see ??).
Components_ cannot be created *alone*. They must be part of a Net_.
3.6.1 Getting a Layer
---------------------
As physical elements, Components_ are created using a Layer_. So prior to
their creation, we must get one from the database. Layers_ are stored in the
Technology_, which in turn, is stored in the DataBase_. So, to get a
Layer_:
.. code-block:: Python
layer = DataBase.getDB().getTechnology().getLayer( 'METAL1' )
.. note:: **Convention for layer names.** As the database can manage both real layers
and symbolic ones we adopt the following convention:
* **Real layers** are named in lowercase (``metal1``, ``nwell``).
* **Symbolic layers** are named in uppercase (``METAL1``, ``NWELL``).
3.6.2 Creating a Net
--------------------
As said above, prior to creating any Component_, we must create the Net_ it
will belong to. In that example we also make it an *external* net, that is,
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
a part of the interface. Do not mistake the name of the net given as a string
argument :cb:`'i'` and the name of the *variable* :cb:`i` holding the Net_ object.
For the sake of clarity we try to give the variable a close name, but this is
not mandatory.
.. code-block:: Python
i = Net.create( cell, 'i' )
i.setExternal( True )
.. note:: Unlike some other database models, in |Hurricane|,
**there is no explicit terminal object**, you only need to make the
net external. For more information about how to connect to an
external net, see `6.2 Creating Nets and connecting to Instances`_.
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
3.6.3 Creating a Component
--------------------------
Finally, we get ready to create a Component_, we will make a Vertical_ segment
of ``METAL1``.
.. code-block:: Python
segment = Vertical.create( i # The owner Net.
, layer # The layer.
, DbU.fromLambda( 5.0 ) # The X coordinate.
, DbU.fromLambda( 2.0 ) # The width.
, DbU.fromLambda( 10.0 ) # The Y source coordinate.
, DbU.fromLambda( 40.0 ) ) # The Y target coordinate.
With this overload of the ``Vertical.create()`` function the segment is created at an
absolute position. There is a second overload for creating a relatively placed
segment, see *articulated layout*.
If the net is external, that is, part of the interface of the cell, you may have
to declare some of its components as physical connectors usable by the router.
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
This is done by calling the NetExternalComponents_ class:
.. code-block:: Python
NetExternalComponents.setExternal( segment )
3.7 Saving to Disk (CRL Core)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Once you have finished to build your cell, you have to save it on disk.
Validating channel routing mode (two metals) on SNX. * New: In Hurricane::Entity, add an id counter limit and a memory size limit. The two limits are checked only when a new Entity object is created. This should help avoiding massive memory links. * New: In CRL Core, add a "crlcore.groundName" and a "crlcore.powerName" parameter to specify the name of the ground/power signals to be created if they are missing in a Cell. For Alliance libraries it would be "vss" & "vdd" (default values), but for real technologies, it is often "gnd!" & "vdd!". The Blif parser is modificated to make use of it. * Bug: In AnabaticEngine::unify(), set the resulting unified segment in the center of the GCells common side. Gcells under a segment are found by using the edge that cover the segment axis. When we have a "bend" GCell stack and the axis is wrong, they could be ommited. This was causing deleted segments to be not removed from some Edges, then core dump. * Change: In Anabatic::AutoSegment::create(), smarter choosing of the reference contact, select the fixed or terminal one instead of always the source one. * New: In Anabatic::Edge::isEnding(), new function to check if a segment going through an Edge is starting/ending in either source or target GCell of the edge (active only when running in channel mode). * New: In Anabatic::Edge::add(), a segment takes part in the occupancy only if it is not ending in either source or target (channel mode only). The occupancy due to terminal is pre-computed in Katana. * New: In Anabatic::Edge::ripup(), in channel mode, never ripup a segment which is ending in either source or target (we *have* to access this edge to connect to the terminal). * Bug: In Anabatic::GCell::hcut() and vcut(), force the update of the Edge which is on the side that will get splitted by the cut. It's capacity will be reduced to it must be updated. * Change: In Anabatic::GCell::updateGContacts() add a flag to conditionnally update horizontals or verticals only. We may require only a partial update when resizing the GCell in only one direction. This, again, related to the fact that we compute the GCells under a segment thanks to it's axis position, so we need to be very careful when modificating axis. * Change: In Katana::Block::resizeChannels(), only update GContact vertical position. Do not disturb X positions of segments. * Bug: In Katana::GlobalRoute::DigitalDistance, in channel mode, some Edges can have a zero capacity, but still be reachable if the net has a terminal in either source or target. Look for this case and return a distance of zero instead of "unreachable". This was causing the global routing not to complete in channel mode. For computing the edge distance, makes the vertical edges much more long (10 times) than the horizontal ones as the vertical capacity is very limited. Hard coded for now, should make it a parameter in the future. * Change: In KatanaEngine::annotateGlobalGraph(), decrease the capacity of edges with reserveCapacity for each terminal inside a GCell. Both north and south edges are decreased as we a terminal will block both north and south edges. As a counterpart, the Edge capacity is not decreased when the global router connect to a terminal. * Change: In Katana::RoutingEvent::revalidate(), when in repair stage, do not expand the slack for horizontal segments in channel mode. So they may not overlap the standard cell row. * Bug: In Stratus documentation, do not use the french option in babel, the documentation is in english! * New: In Documentation, added Hurricane/Python tutorial, part for drawing layout.
2018-03-16 10:20:04 -05:00
Using the AllianceFramework_ you can save it as a pair of file:
========================= =================================== =======================
View Flag File extension
========================= =================================== =======================
Logical / Netlist ``Catalog.State.Logical`` ``.vst``
Physical / Layout ``Catalog.State.Physical`` ``.ap``
========================= =================================== =======================
To save both views, use the ``Catalog.State.Views`` flag. The files
will be written in the |Alliance| ``WORK_DIR``.
.. code-block:: Python
af.saveCell( cell, Catalog.State.Views )
3.8 The Complete Example File
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The example files can be found in the ``share/doc/coriolis2/examples/scripts/``
directory (under the the root of the |Coriolis| installation).
The code needed to run it through the |cgt| viewer has been added. For the
explanation of that part of the code, refer to `5. Make a script runnable through cgt`_.
.. code-block:: Python
#!/usr/bin/python
import sys
from Hurricane import *
from CRL import *
def toDbU ( l ): return DbU.fromLambda(l)
def doBreak ( level, message ):
UpdateSession.close()
Breakpoint.stop( level, message )
UpdateSession.open()
def buildInvertor ( editor ):
UpdateSession.open()
cell = AllianceFramework.get().createCell( 'invertor' )
cell.setTerminal( True )
cell.setAbutmentBox( Box( toDbU(0.0), toDbU(0.0), toDbU(15.0), toDbU(50.0) ) )
if editor:
UpdateSession.close()
editor.setCell( cell )
editor.fit()
UpdateSession.open()
technology = DataBase.getDB().getTechnology()
metal1 = technology.getLayer( "METAL1" )
poly = technology.getLayer( "POLY" )
ptrans = technology.getLayer( "PTRANS" )
ntrans = technology.getLayer( "NTRANS" )
pdif = technology.getLayer( "PDIF" )
ndif = technology.getLayer( "NDIF" )
contdifn = technology.getLayer( "CONT_DIF_N" )
contdifp = technology.getLayer( "CONT_DIF_P" )
nwell = technology.getLayer( "NWELL" )
contpoly = technology.getLayer( "CONT_POLY" )
ntie = technology.getLayer( "NTIE" )
net = Net.create( cell, "nwell" )
Vertical.create( net, nwell, toDbU(7.5), toDbU(15.0), toDbU(27.0), toDbU(51.0) )
vdd = Net.create( cell, "vdd" )
vdd.setExternal( True )
vdd.setGlobal ( True )
h = Horizontal.create(vdd, metal1, toDbU(47.0), toDbU(6.0), toDbU(0.0), toDbU(15.0) )
NetExternalComponents.setExternal( h )
Contact.create ( vdd, contdifn, toDbU(10.0), toDbU(47.0), toDbU( 1.0), toDbU( 1.0) )
Contact.create ( vdd, contdifp, toDbU( 4.0), toDbU(45.0), toDbU( 1.0), toDbU( 1.0) )
Vertical.create( vdd, pdif , toDbU( 3.5), toDbU( 4.0), toDbU(28.0), toDbU(46.0) )
Vertical.create( vdd, ntie , toDbU(10.0), toDbU( 3.0), toDbU(43.0), toDbU(48.0) )
doBreak( 1, 'Done building vdd.' )
vss = Net.create( cell, "vss" )
vss.setExternal( True )
vss.setGlobal ( True )
h = Horizontal.create(vss, metal1, toDbU(3.0), toDbU(6.0), toDbU(0.0), toDbU(15.0))
NetExternalComponents.setExternal( h )
Vertical.create( vss, ndif , toDbU(3.5), toDbU(4.0), toDbU(4.0), toDbU(12.0) )
Contact.create ( vss, contdifn, toDbU(4.0), toDbU(5.0), toDbU(1.0), toDbU( 1.0) )
doBreak( 1, 'Done building vss.' )
i = Net.create( cell, "i" )
i.setExternal( True )
v = Vertical.create ( i, metal1, toDbU(5.0), toDbU(2.0), toDbU(10.0), toDbU(40.0) )
NetExternalComponents.setExternal( v )
Vertical.create ( i, ptrans , toDbU( 7.0), toDbU( 1.0), toDbU(26.0), toDbU(39.0) )
Vertical.create ( i, ntrans , toDbU( 7.0), toDbU( 1.0), toDbU( 6.0), toDbU(14.0) )
Vertical.create ( i, poly , toDbU( 7.0), toDbU( 1.0), toDbU(14.0), toDbU(26.0) )
Horizontal.create( i, poly , toDbU(20.0), toDbU( 3.0), toDbU( 4.0), toDbU( 7.0) )
Contact.create ( i, contpoly, toDbU( 5.0), toDbU(20.0), toDbU( 1.0), toDbU( 1.0) )
doBreak( 1, 'Done building i.' )
nq = Net.create ( cell, "nq" )
nq.setExternal( True )
v = Vertical.create( nq, metal1, toDbU(10.0), toDbU(2.0), toDbU(10.0), toDbU(40.0) )
NetExternalComponents.setExternal( v )
Vertical.create( nq, pdif , toDbU(10.0), toDbU( 3.0), toDbU(28.0), toDbU(37.0) )
Vertical.create( nq, ndif , toDbU(10.0), toDbU( 3.0), toDbU( 8.0), toDbU(12.0) )
Contact.create ( nq, contdifp, toDbU(10.0), toDbU(35.0), toDbU( 1.0), toDbU( 1.0) )
Contact.create ( nq, contdifp, toDbU(10.0), toDbU(30.5), toDbU( 1.0), toDbU( 1.0) )
Contact.create ( nq, contdifn, toDbU(10.0), toDbU(10.0), toDbU( 1.0), toDbU( 1.0) )
doBreak( 1, 'Done building q.' )
UpdateSession.close()
AllianceFramework.get().saveCell( cell, Catalog.State.Views )
return
def ScriptMain ( **kw ):
editor = None
if kw.has_key('editor') and kw['editor']:
editor = kw['editor']
buildInvertor( editor )
return True