Skip to content

Conversation

@clavin-xlnx
Copy link
Member

@clavin-xlnx clavin-xlnx commented Jul 13, 2024

Splits up the logical and physical netlist Interchange writers/readers to use multiple messages. This is to avoid out of memory issues and also to avoid hitting fundamental limits with Cap'n Proto for very large designs. This builds on #1017.

The logical netlist that is reused inside the DeviceResources message is preserved as not multi-message.

clavin-xlnx and others added 17 commits June 12, 2024 12:24
Signed-off-by: Chris Lavin <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
… device (#1026)

* Do not assume Y = 0 has Laguna tiles, since it could be HBM device

Signed-off-by: Eddie Hung <[email protected]>

* Add test for xcu50

Signed-off-by: Eddie Hung <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
* Do not NPE on encrypted netlists

Signed-off-by: Eddie Hung <[email protected]>

* Apply suggestions from code review

Co-authored-by: Chris Lavin <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>
Co-authored-by: Chris Lavin <[email protected]>
* Test for site routing from raw placed design

Signed-off-by: Chris Lavin <[email protected]>

* Update src/com/xilinx/rapidwright/util/VivadoTools.java

Co-authored-by: eddieh-xlnx <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>

* Refactoring

Signed-off-by: Chris Lavin <[email protected]>

* Switching to HEAD

Signed-off-by: Chris Lavin <[email protected]>

---------

Signed-off-by: Chris Lavin <[email protected]>
Co-authored-by: eddieh-xlnx <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
Write *_load.tcl into same dir as DCP rather than into current
working dir

Signed-off-by: Eddie Hung <[email protected]>
* [RWRoute] Do not assume Y = 0 has Laguna tiles, since it could be HBM device (#1026)

* Do not assume Y = 0 has Laguna tiles, since it could be HBM device

Signed-off-by: Eddie Hung <[email protected]>

* Add test for xcu50

Signed-off-by: Eddie Hung <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>

* [RWRoute] Do not NPE on encrypted netlists (#1025)

* Do not NPE on encrypted netlists

Signed-off-by: Eddie Hung <[email protected]>

* Apply suggestions from code review

Co-authored-by: Chris Lavin <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>
Co-authored-by: Chris Lavin <[email protected]>

* Add TestSite.testGetIntTile()

Signed-off-by: Eddie Hung <[email protected]>

* Update test/src/com/xilinx/rapidwright/device/TestSite.java

Co-authored-by: Chris Lavin <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>
Co-authored-by: Chris Lavin <[email protected]>
@clavin-xlnx clavin-xlnx requested a review from eddieh-xlnx July 13, 2024 00:05
clavin-xlnx and others added 12 commits July 12, 2024 20:27
Signed-off-by: Chris Lavin <[email protected]>
* [RWRoute] Do not assume Y = 0 has Laguna tiles, since it could be HBM device (#1026)

* Do not assume Y = 0 has Laguna tiles, since it could be HBM device

Signed-off-by: Eddie Hung <[email protected]>

* Add test for xcu50

Signed-off-by: Eddie Hung <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>

* [RWRoute] Do not NPE on encrypted netlists (#1025)

* Do not NPE on encrypted netlists

Signed-off-by: Eddie Hung <[email protected]>

* Apply suggestions from code review

Co-authored-by: Chris Lavin <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>
Co-authored-by: Chris Lavin <[email protected]>

* Test that PIP.getArbitraryPIP() sets the reversed flag correctly

Signed-off-by: Eddie Hung <[email protected]>

* Add TestNode.getAll{Uphill,Downhill}PIPsReversed() tests

Signed-off-by: Eddie Hung <[email protected]>

* Adding HDIOB types (#1028)

Signed-off-by: Chris Lavin <[email protected]>

* [EDIFTools] writeTclLoadScriptForPartialEncryptedDesigns abspath (#1029)

Write *_load.tcl into same dir as DCP rather than into current
working dir

Signed-off-by: Eddie Hung <[email protected]>

* Test that PIP.getArbitraryPIP() sets the reversed flag correctly

Signed-off-by: Eddie Hung <[email protected]>

* Add TestNode.getAll{Uphill,Downhill}PIPsReversed() tests

Signed-off-by: Eddie Hung <[email protected]>

---------

Signed-off-by: Eddie Hung <[email protected]>
Signed-off-by: eddieh-xlnx <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
Co-authored-by: Chris Lavin <[email protected]>
Signed-off-by: Chris Lavin <[email protected]>
@clavin-xlnx
Copy link
Member Author

Closing as not much interest for now.

@eddieh-xlnx eddieh-xlnx deleted the large_interchange_designs branch May 13, 2025 19:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants