Hugging Face's logo

Datasets:
The-OpenROAD-Project
/
ORAssistant_RAG_Dataset

Dataset card Files
xet
Community
1
espsluar commited on
Commit
fd15ddd
·
verified ·
1 Parent(s): 493461e

Delete gh_discussions/Feature Request

Browse files
Files changed (15) hide show
  1. gh_discussions/Feature Request/1871.md +0 -23
  2. gh_discussions/Feature Request/2784.md +0 -108
  3. gh_discussions/Feature Request/3331.md +0 -37
  4. gh_discussions/Feature Request/3359.md +0 -53
  5. gh_discussions/Feature Request/3531.md +0 -17
  6. gh_discussions/Feature Request/3619.md +0 -110
  7. gh_discussions/Feature Request/3788.md +0 -434
  8. gh_discussions/Feature Request/3812.md +0 -34
  9. gh_discussions/Feature Request/3856.md +0 -32
  10. gh_discussions/Feature Request/3939.md +0 -33
  11. gh_discussions/Feature Request/4420.md +0 -65
  12. gh_discussions/Feature Request/4422.md +0 -22
  13. gh_discussions/Feature Request/4819.md +0 -19
  14. gh_discussions/Feature Request/5041.md +0 -23
  15. gh_discussions/Feature Request/5420.md +0 -24
gh_discussions/Feature Request/1871.md DELETED
@@ -1,23 +0,0 @@
1
- # Support for MEOL layer?
2
-
3
- Tool: OpenDB
4
-
5
- Subcategory: Add MEOL layer support
6
-
7
- ## Conversation
8
-
9
- ### GuzTech
10
- Hello everyone!
11
-
12
- I'm in the process of bringing up a new platform which is going well. I am running into a problem with a local interconnect layer which is specified as a `MASTERSLICE` with `LEF58_TYPE` = `MEOL`, so it is not specified as a `ROUTING` layer as is the case with the sky130 PDK. OpenDB complains that there is a parse mismatch for this `lil` layer.
13
-
14
- If I add this `lil` layer as a routing layer to `make_tracks.tcl` then OpenDB complains that it is not a routing layer. If I don't add `lil` as a routing layer, then TritonRoute gives an error that it cannot find layer `lil` for viarule that connects `lil` to the first metal layer.
15
-
16
- Did anyone else run into a similar problem? If so, how did you solve it?
17
-
18
- ### maliberty
19
- Please include the odb message. You should not add tracks if it is not a routing layer.
20
-
21
- ### maliberty
22
- We need to add support for MEOL. You can workaround it my making it a masterslice layer temporarily.
23
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/2784.md DELETED
@@ -1,108 +0,0 @@
1
- # New DFT Tool: Starting with scan replace
2
-
3
- Tool: Design for Test
4
-
5
- Subcategory: New tool addition
6
-
7
- ## Conversation
8
-
9
- ### fgaray
10
- Hello OpenROAD devs,
11
-
12
- I am planning on contributing a new tool (following https://github.com/The-OpenROAD-Project/OpenROAD/blob/master/docs/contrib/DeveloperGuide.md, I will call this "dft") for DFT insertion into OpenROAD and I would like to discuss the user facing API.
13
-
14
- # User facing API
15
-
16
- I think we should create two new main commands for DFT: preview_dft and insert_dft:
17
-
18
- * preview_dft: will show a summary of what scan chains are we going to create (after scan architect) but will not apply any changes to the design. This is an immutable command and will not perform any modifications to the design or database. This is useful for debugging proposes and to help users to iterate any DFT option that could change the creation of the scan chains.
19
- * insert_dft: will run again from scratch the immutable part of preview_dft, including scan architect and perform scan stitching. The result of insert_dft should be the same that the one presented by preview_dft but applied to the design.
20
-
21
- I would say that it is better to configure the DFT compiler with commands like "set_dft_*" instead of receiving any argument in "preview_dft" or "insert_dft". For example: set_dft_no_scan (for cells that we don't want to scan replace), set_dft_chain_count (for the number of scan chains to be created) or set_dft_clock_mix (if we want or not to mix clock domains).
22
-
23
- Now, I am not sure if this is going to be something you expect to see in the Python interface, this is a very TCL oriented API and I would like to see if you have any suggestion on doing this more programmatically for Python scripting (maybe this is totally fine as we can have a different API for Python?).
24
-
25
- # Where in the flow is DFT insertion going to be?
26
-
27
- ![OpenROAD's flow](https://user-images.githubusercontent.com/1619948/213331834-76ea8bb9-a63b-4576-90ff-5e6eca6e5647.png)
28
-
29
- Looking into OpenROAD's flow, I think the best part to perform DFT insertion is between "Logic Synthesis" and "Floorplanning". That way we floorplanning will have all the info about DFT to perform optimizations in the design.
30
-
31
- Feel free to suggest something different here :).
32
-
33
- # What is the plan?
34
-
35
- For now, I will only be creating "insert_dft" and "set_no_scan" for scan replacement: I will replace sequential cells (excluding black boxes as we don't have support for [CTLs](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=966626) yet!) in the design with scan equivalents if they are available and nothing more inside insert_dft. I will keep of of the scope adding scan muxes or any other kind of scan stitching as this is my first approach to OpenROAD.
36
-
37
- Please, let me know if there is something I should be aware of or anything related to this idea or the user interface.
38
-
39
- Thanks!,
40
- Felipe
41
-
42
-
43
- ### maliberty
44
- I think it is best to do scan chain after global placement. You want to put the ff in the best placement for timing and only then construct the chain ordering based on neighbors.
45
-
46
- It would be helpful to know what commands you intend to implement to define the chains. A small complete example would clarify the use model.
47
-
48
- For python API please see the discussion in https://github.com/The-OpenROAD-Project/OpenROAD/discussions/2745
49
-
50
- ### maliberty
51
- @rovinski @msaligane @louiic @Colin-Holehouse any comments?
52
-
53
- ### rovinski
54
- Very cool! Always excited to see more functionality contributed 😃
55
-
56
- ### maliberty
57
- Would you say more about scandef and how you intend to generate it?
58
-
59
- ### maliberty
60
- Note that odb doesn't store any scan information from DEF today so you'll have to plan on enhancing it as part of this project.
61
-
62
- ### maliberty
63
- In case its useful: http://vlsicad.eecs.umich.edu/BK/Slots/cache/vlsicad.ucsd.edu/GSRC/Bookshelf/Slots/ScanOpt/
64
-
65
- ### fgaray
66
- Hey @maliberty ,
67
-
68
- I am trying to write a function to perform the scan replace and I need to find a scan equivalent cell in the DB.
69
-
70
- I am iterating over the libs in the odb::dbDatabase like this:
71
-
72
- ```
73
- bool IsLibScanCell(odb::dbLib* lib) {
74
- for (odb::dbMaster* master: lib->getMasters()) {
75
- for (odb::dbMTerm* mterm: master->getMTerms()) {
76
- std::cout << mterm->getSigType().getString() << std::endl;
77
- }
78
- }
79
- return false;
80
- }
81
- ```
82
-
83
- But looks like the enum of getSigType does not have enough information to see if a pin is scan_enable or scan_in. Is this something that I need to implement or is there another way to access this information inside the lib? I see that in sta/ there are some parsing to detect scan signals, but I do not see them being used in any place outside sta/.
84
-
85
-
86
-
87
- ### fgaray
88
- For scan replace, I have a working demo where I collect the scan lib cells and then I try to iterate over the design's blocks applying "sta::dbNetwork" replaceCell, but I found that this and some methods in sta:: check before performing the replace if both cells are "equivalent" by counting the number of ports in each cell (See inDbInstSwapMasterBefore in dbSta/src/dbSta.cc ).
89
-
90
- This is not really true when you try to replace a cell with an scan equivalent because the scan one will have additional ports like scan in and scan enable.
91
-
92
- I was wondering if there is a function that I am missing that could help me with this kind of replace.
93
-
94
- If not, should I provide a new method for dbSta and odb that checks for functional equivalent ports instead of index/name based matching (See swapMaster in odb/src/db/dbInst.cpp) or roll one myself just for DFT? Would you prefer this method to be available in odb or maybe just complete change the way we do the replaceCell?
95
-
96
- Thanks!
97
-
98
- ### fgaray
99
- Hi @maliberty,
100
-
101
- Do you know how to test if two ports/pins are functional equivalent?, I am trying to use sta::LibertyPort's function() and sta::FuncExpr::equiv() but looks like sta::LibertyPort's function() returns most of the time nullptr. I see that in odb/src/db/dbInst.cpp:1201, we test that mterms are equivalent by checking that the name of the ports is the same (line 1236).
102
-
103
- I can work with that but I think my scan replace function will be more robust if I can check if the ports are functional equivalent.
104
-
105
-
106
- ### fgaray
107
- Implemented basic DFT in #2900, #3079 and #3356
108
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3331.md DELETED
@@ -1,37 +0,0 @@
1
- # Synthesis support
2
-
3
- Subcategory: Synthesis support
4
-
5
- ## Conversation
6
-
7
- ### Blebowski
8
- Hello, Is there any idea whether eventually OpenRoad would like to support also synthesis as part of the tool (instead of using Yosys externally) ?
9
-
10
- ### maliberty
11
- There is no plan for that. At most we would integrate yosys. What motivates your question?
12
-
13
- ### Blebowski
14
- Hi, @maliberty , several points:
15
- - Yosys does not support full System Verilog (AFAIK, only with proprietary parser). Nor it supports VHDL (only via GHDL yosys plugin, and getting this to work is tricky)
16
- - Synthesis is mentioned several times in: https://github.com/The-OpenROAD-Project/OpenROAD/issues/1759
17
- 1. Hard to get netlist that passes the timing
18
- 2. Possible QoR gains
19
-
20
- - Generally, placement / routing aware synthesis reaches better results in lower geometries
21
- - Industry high-end goes this direction
22
- - Easier user interface (single database, simpler to pass information from synthesis to PnR)
23
-
24
-
25
- ### Blebowski
26
- However,
27
- I need to note that I am not actively using nor developing OpenRoad. As for the usage, in company where I work, we use Synopsys for Synthesis and PnR tools due to stability, documentation, good performance and vast features available. We are looking forward for the day when we will be able to use open source.
28
- As for the development, I would really like to join, but I already have little time for the open-source work due to my daily job.
29
-
30
- ### maliberty
31
- SV support is unrelated to OR and is being addressed in https://github.com/chipsalliance/Surelog. The pain is mostly related to missing ASIC oriented features in yosys/abc (eg operator mapping).
32
-
33
- In general its a large effort and would need dedicated resources to work on it. There isn't enough bandwidth to make it a priority. If you want to start addressing some of these concerns I'd be happy to discuss further.
34
-
35
- ### maliberty
36
- OR has already integrated abc which handles the technology dependent operations. Its less clear that there is a big advantage to the technology independent operations in yosys itself. There is some work slowly progressing on remapping in OR using abc.
37
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3359.md DELETED
@@ -1,53 +0,0 @@
1
- # Ideas for multithreaded rendering
2
-
3
- Tool: Graphical User Interface
4
-
5
- Subcategory: Performance enhancement
6
-
7
- ## Conversation
8
-
9
- ### rovinski
10
- I saw the conversation in #3330, #3342, #3348, etc. and started thinking about how the rendering time might be improved aside from not drawing objects that aren't useful.
11
-
12
- It seems that the rendering, for the most part, is embarrassingly parallel. In `LayoutViewer::drawInstanceOutlines()` for example (the most expensive layer measured in #3348), a function is applied to every instance in a vector to determine how / if it gets drawn, and then it makes a draw call.
13
-
14
- https://github.com/The-OpenROAD-Project/OpenROAD/blob/6920ac530b6959f742e75b7f55062b1642730b14/src/gui/src/layoutViewer.cpp#L2041-L2091
15
-
16
- This is essentially a [map reduce](https://en.wikipedia.org/wiki/MapReduce) paradigm which should be extremely parallel, except that Qt painter objects aren't thread safe. There was discussion in #3279 that the database is not "thread safe", but I think that was in the context of read/write operations and not read-only operations.
17
-
18
- Beyond that, each layer is independently but sequentially rendered:
19
-
20
- https://github.com/The-OpenROAD-Project/OpenROAD/blob/6920ac530b6959f742e75b7f55062b1642730b14/src/gui/src/layoutViewer.cpp#L2555-L2599
21
-
22
- There may or may not be ordering required on draw calls to these layers, I'm not sure.
23
-
24
- I looked at a few [Qt guides](https://doc.qt.io/qt-5/qtcore-threads-mandelbrot-example.html) and they all recommend separating computation from rendering in separate threads for better performance (and interactivity).
25
-
26
- I think some of these things were mentioned by @QuantamHD in #1930 but the PR itself seems to be unrelated (and abandoned).
27
-
28
- I have some interest in looking into this, but I wonder what the best approach would be for this? Qt has [QThread](https://doc.qt.io/qt-5/qthread.html) and [QThreadPool](https://doc.qt.io/qt-5/qthreadpool.html) for this exact purpose. There's also [QtConcurrent](https://doc.qt.io/qt-5/qtconcurrent.html) which may have some uses.
29
-
30
- In an ideal world, it seems like there should be a thread pool with `getThreadCount()-1` threads and the main thread. The main thread has a render queue which atomically pops the queue for draw calls and does nothing else until all worker threads yield. The worker threads can parallelize at the granularity of layer or instance/shape depending on what's optimal. But this could take a lot of refactoring, possibly be overkill, or possibly be difficult to tune the amount of threading correctly.
31
-
32
- An intermediate step could be to simply parallelize some of the heavy-weight tasks that block the GUI thread, such as using `std::transform(std::execution::par)` or `std::for_each(std::execution::par)` and separating the draw calls from the filtering. But also these functions don't seem to be well supported on old versions of GCC/Clang (it requires Intel `libtbb` and may error or just use a serial version if it isn't linked, but I think it also think it doesn't work on Apple silicon). Maybe using OpenMP would work?
33
-
34
- Any thoughts or opinions?
35
- @gadfort @maliberty @oharboe @QuantamHD
36
-
37
- ### oharboe
38
- After the recent performance improvement, the biggest performance problem currently for us, is that all intermediate zoom levels are rendered https://github.com/The-OpenROAD-Project/OpenROAD/issues/3278
39
-
40
- So if that could be fixed FIRST... :-)
41
-
42
- Afterwards, the suggestions that you propose sound wicked!
43
-
44
- It is true that Qt nor the database is thread safe, then on the other hand, this means that the GUI is single threaded, so I suppose that you are right that you can use as many or few threads as you like as long as Qt nor the database functions are non-the-wiser. Rendering should be especially amenable to this since it is a read only operation...
45
-
46
- Also, rendering doesn't have to be perfect. I have been thinking that a draft could be rendered while zooming and then when the user stops giving input, more detail can be rendered.
47
-
48
-
49
- ### maliberty
50
- @rovinski I've considered all these ideas but lack the time to implement them all. Right now I'm looking at moving the rendering to a separate thread and making it interruptible. The main challenge is figuring out how to trigger the right events at the right time to make Qt happy. It is very particular that you can only paint in paintEvent.
51
-
52
- Assuming I can land this then it is more straightforward to split that worker thread into multiple threads. KLayout works in a similar fashion where they render per layer in each thread and composite in the main thread.
53
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3531.md DELETED
@@ -1,17 +0,0 @@
1
- # UPF support
2
-
3
- Tool: Unified Power Format (UPF)
4
-
5
- Subcategory: UPF support
6
-
7
- ## Conversation
8
-
9
- ### kareefardi
10
- Is https://github.com/The-OpenROAD-Project/OpenROAD/tree/master/src/upf a work in progress (it doesn't have documentation). If it is, is it planned to have it supported?
11
-
12
- ### maliberty
13
- UPF support is a work in progress. The reader in upf/ works but the rest of the tool needs to support it.
14
-
15
- ### maliberty
16
- I think the last remaining piece was pdn inserting the power switches. @gadfort would you update on this?
17
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3619.md DELETED
@@ -1,110 +0,0 @@
1
- # DFT: Storing scan chain information in odb
2
-
3
- Tool: Design for Test
4
-
5
- Subcategory: Scan chain representation
6
-
7
- ## Conversation
8
-
9
- ### fgaray
10
- Hello everyone,
11
-
12
- To support drawing scan chains in OpenROAD in the gui and CTL ([Core Test Language](https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=966626)) parser/generator we need to make the scan info available in odb. Here I describe a DB schema to represent CTLs in odb.
13
-
14
- ## Objective
15
-
16
- * Make scan chains available to the gui to be able to highlight scan structures and display information related to scan.
17
- * Provide a common source of true for other commands to query and find scan structures on the design.
18
- * Provide support for a future parser for CTLs files.
19
-
20
- ## Out of scope
21
-
22
- * Parsing and writing a CTL file at this time.
23
- * Handling test patterns.
24
-
25
- ## CTLs
26
-
27
- CTLs are files based on the [STIL standard](https://grouper.ieee.org/groups/1450/dot1/p1450.1-D14.pdf) that allow us to share information between DFT tools for optimizations and interoperability between different vendors. They describe the internal scan structures and they are useful too to share information between internal commands inside OpenROAD, like scan insertion and the GUI.
28
-
29
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/1619948/b513ec9c-4575-45c4-bac0-0401ab164da2)
30
-
31
- Source: CTL for Test Information of Digital ICs, Rohit Kapur
32
-
33
- **The schema described here is not a 1:1 map of a CTL file but should be enough to create one or to store the information coming from a CTL.**
34
-
35
- ## dbScanChain
36
-
37
- * name: string
38
- * length: uint
39
- * cells: list\<dbInst\>
40
- * scanIn: dbScanPin
41
- * scanOut: dbScanPin
42
- * scanClock: dbScanPin
43
- * scanEnable: dbScanPin
44
- * testMode: string
45
- * partitions: list\<dbScanPartition\>
46
- * scanInsts: list\<dbScanInst\>
47
-
48
- ## dbScanPin
49
-
50
- Wrapper to handle either dbBTerm or dbITerm:
51
-
52
- * bterm: dbBTerm
53
- * iterm: dbITerm
54
-
55
- ## dbScanPartition
56
-
57
- * start: dbScanPin
58
- * stop: dbScanPin
59
- * name: string
60
-
61
- ## dbScanInst
62
-
63
- * scanIn: dbScanPin
64
- * scanOut: dbScanPin
65
- * bits: uint
66
- * type: enum [OneBit, ShiftRegister, BlackBox]
67
-
68
-
69
- ## Flow
70
-
71
- At the end of *insert_dft* we will be storing the scan chain information into odb using the created dbCTL object. The gui will be able to query this information and display it. dbScanChain will be attached to the dbBlock that contains the scan chains that are inside it.
72
-
73
-
74
- ## Open Questions
75
-
76
- * Where can I add a new dbObject (dbCTL) to dbBlock?
77
- * How can I represent maps (will std::map work?) string-> dbObject in the json db format?
78
- * I am creating a wrapper for dbBTerm and dbITerm for cases where the scan pins are not coming from the block, is there any way to handle this in the json db format?
79
-
80
-
81
-
82
- ### fgaray
83
- Hi @maliberty , who do you think is the right person to review this? Specially the open questions at the end.
84
-
85
- Thanks!
86
-
87
- ### maliberty
88
- Scan chain is used in both CTL and in DEF. Can dbScanChain accommodate both or do we need to consider having two classes? If the later, how would you suggest naming them ?
89
-
90
- ### maliberty
91
- Looking at DEF:
92
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/761514/6a7a39c4-e9b0-44e5-8beb-0fe450c7928d)
93
-
94
- It looks like the dbScanPartition needs a way to store MAXBITS.
95
-
96
- Each element of the chain also has a BITS field which seems hard to represent here. Perhaps we need a dbScanInst?
97
-
98
- How would you infer FLOATING or ORDERED from the design?
99
-
100
- ### fgaray
101
- Hi @maliberty ,
102
-
103
- Do you think we are ready to start the implementation?
104
-
105
- Thanks!,
106
- Felipe
107
-
108
- ### QuantamHD
109
- LGTM. Thanks for the contirbution @fgaray
110
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3788.md DELETED
@@ -1,434 +0,0 @@
1
- # Clock Tree Viewer and trying to get to the information in the tooltips
2
-
3
- Tool: Graphical User Interface
4
-
5
- Subcategory: Data visualization improvement
6
-
7
- ## Conversation
8
-
9
- ### oharboe
10
- The tooltips in the Clock Tree Viewer contain interesting information: which macro or flip flop has how much clock network latency.
11
-
12
- However, the tooltip information is hard to get to and it is not searchable.
13
-
14
- I mocked a text report that is printed when I click update...
15
-
16
- I think perhaps a better approach to this is to have a tcl utility procedure that navigates the clock tree and print out this information. Perhaps that is possible today, but I don't know how to...
17
-
18
- Change I used to mock the report below: https://github.com/Pinata-Consulting/OpenROAD/commit/17283788372d98cafc1e6bbd056f1074a1539bc6
19
-
20
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/1a874efa-2b66-41f2-b365-7990f3ed5399)
21
-
22
-
23
- ### oharboe
24
- Mocking a heatmap of arrival times. I only have location information for the Elements. Got the idea here. https://www.eng.biu.ac.il/temanad/files/2017/02/Lecture-8-CTS.pdf
25
-
26
- ```
27
- import re
28
- import matplotlib.pyplot as plt
29
- import numpy as np
30
-
31
- # Extracting data from the text
32
- text = """
33
- # Your log content goes here
34
- """
35
-
36
- lines = text.strip().split('\n')
37
- filtered_lines = [line for line in lines if "ces" in line]
38
-
39
- # Finding the maximum indices for rows and columns
40
- max_row = max_col = 0
41
- data_dict = {}
42
- for line in filtered_lines:
43
- match = re.search(r"ces_(\d+)_(\d+)/clock\s*Arrival:\s*(\d+\.\d+)", line)
44
- if match:
45
- row, col, value = map(int, map(float, match.groups()))
46
- max_row = max(max_row, row)
47
- max_col = max(max_col, col)
48
- data_dict[(row, col)] = value
49
-
50
- # Constructing the data matrix
51
- data_matrix = np.zeros((max_row + 1, max_col + 1))
52
- for (row, col), value in data_dict.items():
53
- data_matrix[row, col] = value
54
-
55
- # Plotting the heatmap
56
- plt.imshow(data_matrix, origin='upper', cmap='viridis', aspect='auto')
57
- plt.colorbar(label='Arrival (ps)')
58
- plt.title('Heatmap of Arrival Times')
59
- plt.xlabel('Column')
60
- plt.ylabel('Row')
61
- plt.show()
62
- ```
63
-
64
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/aceb47d0-d673-461d-a0f0-e0b3df96a8d6)
65
-
66
-
67
- ### oharboe
68
- Adding some 3d bling...
69
-
70
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/ad5b4839-f77e-4643-89d8-4426b4137f89)
71
-
72
-
73
- ```
74
- # Sample data
75
- text = """
76
- [INFO GUI-0088] Register _131_/CLK Arrival: 258.115 ps
77
- [INFO GUI-0088] Register _132_/CLK Arrival: 257.882 ps
78
- [INFO GUI-0088] Register _133_/CLK Arrival: 258.944 ps
79
- [INFO GUI-0088] Register _134_/CLK Arrival: 258.636 ps
80
- [INFO GUI-0088] Register _135_/CLK Arrival: 258.069 ps
81
- [INFO GUI-0088] Register _137_/CLK Arrival: 257.527 ps
82
- [INFO GUI-0088] Register _139_/CLK Arrival: 246.072 ps
83
- [INFO GUI-0088] Register _143_/CLK Arrival: 245.541 ps
84
- [INFO GUI-0088] Register _145_/CLK Arrival: 245.360 ps
85
- [INFO GUI-0088] Register ces_0_7/clock Arrival: 232.150 ps
86
- [INFO GUI-0088] Register ces_1_7/clock Arrival: 225.713 ps
87
- [INFO GUI-0088] Register _136_/CLK Arrival: 250.476 ps
88
- [INFO GUI-0088] Register _138_/CLK Arrival: 250.337 ps
89
- [INFO GUI-0088] Register _140_/CLK Arrival: 246.845 ps
90
- [INFO GUI-0088] Register _141_/CLK Arrival: 246.075 ps
91
- [INFO GUI-0088] Register _142_/CLK Arrival: 246.461 ps
92
- [INFO GUI-0088] Register _144_/CLK Arrival: 246.885 ps
93
- [INFO GUI-0088] Register _146_/CLK Arrival: 245.560 ps
94
- [INFO GUI-0088] Register ces_0_6/clock Arrival: 235.587 ps
95
- [INFO GUI-0088] Register _147_/CLK Arrival: 239.633 ps
96
- [INFO GUI-0088] Register _148_/CLK Arrival: 238.903 ps
97
- [INFO GUI-0088] Register _149_/CLK Arrival: 239.580 ps
98
- [INFO GUI-0088] Register _150_/CLK Arrival: 238.791 ps
99
- [INFO GUI-0088] Register _151_/CLK Arrival: 239.418 ps
100
- [INFO GUI-0088] Register _152_/CLK Arrival: 239.304 ps
101
- [INFO GUI-0088] Register _153_/CLK Arrival: 238.755 ps
102
- [INFO GUI-0088] Register _154_/CLK Arrival: 238.495 ps
103
- [INFO GUI-0088] Register _155_/CLK Arrival: 250.283 ps
104
- [INFO GUI-0088] Register _156_/CLK Arrival: 250.061 ps
105
- [INFO GUI-0088] Register _157_/CLK Arrival: 250.546 ps
106
- [INFO GUI-0088] Register _158_/CLK Arrival: 250.032 ps
107
- [INFO GUI-0088] Register _159_/CLK Arrival: 250.760 ps
108
- [INFO GUI-0088] Register _160_/CLK Arrival: 249.826 ps
109
- [INFO GUI-0088] Register _161_/CLK Arrival: 250.433 ps
110
- [INFO GUI-0088] Register _162_/CLK Arrival: 249.956 ps
111
- [INFO GUI-0088] Register ces_2_7/clock Arrival: 233.829 ps
112
- [INFO GUI-0088] Register ces_3_6/clock Arrival: 243.645 ps
113
- [INFO GUI-0088] Register ces_3_7/clock Arrival: 239.027 ps
114
- [INFO GUI-0088] Register ces_0_0/clock Arrival: 290.815 ps
115
- [INFO GUI-0088] Register ces_0_1/clock Arrival: 287.224 ps
116
- [INFO GUI-0088] Register ces_0_2/clock Arrival: 276.359 ps
117
- [INFO GUI-0088] Register ces_1_0/clock Arrival: 262.155 ps
118
- [INFO GUI-0088] Register ces_1_1/clock Arrival: 258.349 ps
119
- [INFO GUI-0088] Register ces_1_2/clock Arrival: 248.148 ps
120
- [INFO GUI-0088] Register ces_0_3/clock Arrival: 226.049 ps
121
- [INFO GUI-0088] Register ces_1_3/clock Arrival: 224.760 ps
122
- [INFO GUI-0088] Register ces_0_4/clock Arrival: 242.883 ps
123
- [INFO GUI-0088] Register ces_0_5/clock Arrival: 247.124 ps
124
- [INFO GUI-0088] Register ces_1_4/clock Arrival: 237.497 ps
125
- [INFO GUI-0088] Register ces_1_5/clock Arrival: 249.241 ps
126
- [INFO GUI-0088] Register ces_1_6/clock Arrival: 253.384 ps
127
- [INFO GUI-0088] Register ces_2_0/clock Arrival: 253.191 ps
128
- [INFO GUI-0088] Register ces_2_1/clock Arrival: 248.850 ps
129
- [INFO GUI-0088] Register ces_2_2/clock Arrival: 236.498 ps
130
- [INFO GUI-0088] Register ces_3_0/clock Arrival: 261.042 ps
131
- [INFO GUI-0088] Register ces_3_1/clock Arrival: 257.075 ps
132
- [INFO GUI-0088] Register ces_3_2/clock Arrival: 244.640 ps
133
- [INFO GUI-0088] Register ces_2_3/clock Arrival: 225.582 ps
134
- [INFO GUI-0088] Register ces_2_4/clock Arrival: 239.450 ps
135
- [INFO GUI-0088] Register ces_2_5/clock Arrival: 251.290 ps
136
- [INFO GUI-0088] Register ces_2_6/clock Arrival: 255.456 ps
137
- [INFO GUI-0088] Register ces_3_3/clock Arrival: 227.135 ps
138
- [INFO GUI-0088] Register ces_3_4/clock Arrival: 243.496 ps
139
- [INFO GUI-0088] Register ces_3_5/clock Arrival: 248.000 ps
140
- [INFO GUI-0088] Register _163_/CLK Arrival: 253.020 ps
141
- [INFO GUI-0088] Register _164_/CLK Arrival: 253.796 ps
142
- [INFO GUI-0088] Register _165_/CLK Arrival: 254.691 ps
143
- [INFO GUI-0088] Register _166_/CLK Arrival: 254.356 ps
144
- [INFO GUI-0088] Register _167_/CLK Arrival: 254.200 ps
145
- [INFO GUI-0088] Register _168_/CLK Arrival: 253.308 ps
146
- [INFO GUI-0088] Register _169_/CLK Arrival: 252.967 ps
147
- [INFO GUI-0088] Register _170_/CLK Arrival: 253.566 ps
148
- [INFO GUI-0088] Register ces_4_6/clock Arrival: 236.291 ps
149
- [INFO GUI-0088] Register ces_4_7/clock Arrival: 232.063 ps
150
- [INFO GUI-0088] Register _171_/CLK Arrival: 231.291 ps
151
- [INFO GUI-0088] Register _172_/CLK Arrival: 231.543 ps
152
- [INFO GUI-0088] Register _173_/CLK Arrival: 231.174 ps
153
- [INFO GUI-0088] Register _174_/CLK Arrival: 230.426 ps
154
- [INFO GUI-0088] Register _175_/CLK Arrival: 231.306 ps
155
- [INFO GUI-0088] Register _176_/CLK Arrival: 231.412 ps
156
- [INFO GUI-0088] Register _177_/CLK Arrival: 231.484 ps
157
- [INFO GUI-0088] Register _178_/CLK Arrival: 231.182 ps
158
- [INFO GUI-0088] Register _179_/CLK Arrival: 244.957 ps
159
- [INFO GUI-0088] Register _180_/CLK Arrival: 242.174 ps
160
- [INFO GUI-0088] Register _181_/CLK Arrival: 244.902 ps
161
- [INFO GUI-0088] Register _182_/CLK Arrival: 242.804 ps
162
- [INFO GUI-0088] Register _183_/CLK Arrival: 244.575 ps
163
- [INFO GUI-0088] Register _184_/CLK Arrival: 243.857 ps
164
- [INFO GUI-0088] Register _185_/CLK Arrival: 242.949 ps
165
- [INFO GUI-0088] Register _186_/CLK Arrival: 244.352 ps
166
- [INFO GUI-0088] Register _187_/CLK Arrival: 253.289 ps
167
- [INFO GUI-0088] Register _189_/CLK Arrival: 253.454 ps
168
- [INFO GUI-0088] Register _193_/CLK Arrival: 253.757 ps
169
- [INFO GUI-0088] Register _194_/CLK Arrival: 253.377 ps
170
- [INFO GUI-0088] Register _188_/CLK Arrival: 249.940 ps
171
- [INFO GUI-0088] Register _190_/CLK Arrival: 250.214 ps
172
- [INFO GUI-0088] Register _191_/CLK Arrival: 250.193 ps
173
- [INFO GUI-0088] Register _192_/CLK Arrival: 250.440 ps
174
- [INFO GUI-0088] Register ces_6_6/clock Arrival: 237.794 ps
175
- [INFO GUI-0088] Register ces_6_7/clock Arrival: 237.428 ps
176
- [INFO GUI-0088] Register ces_7_6/clock Arrival: 248.387 ps
177
- [INFO GUI-0088] Register ces_7_7/clock Arrival: 243.904 ps
178
- [INFO GUI-0088] Register ces_4_0/clock Arrival: 269.870 ps
179
- [INFO GUI-0088] Register ces_4_1/clock Arrival: 265.959 ps
180
- [INFO GUI-0088] Register ces_4_2/clock Arrival: 254.121 ps
181
- [INFO GUI-0088] Register ces_4_3/clock Arrival: 227.621 ps
182
- [INFO GUI-0088] Register ces_5_0/clock Arrival: 260.539 ps
183
- [INFO GUI-0088] Register ces_5_1/clock Arrival: 256.515 ps
184
- [INFO GUI-0088] Register ces_5_2/clock Arrival: 245.174 ps
185
- [INFO GUI-0088] Register ces_5_3/clock Arrival: 225.957 ps
186
- [INFO GUI-0088] Register ces_5_4/clock Arrival: 233.418 ps
187
- [INFO GUI-0088] Register ces_5_5/clock Arrival: 250.130 ps
188
- [INFO GUI-0088] Register ces_5_6/clock Arrival: 260.323 ps
189
- [INFO GUI-0088] Register ces_5_7/clock Arrival: 263.816 ps
190
- [INFO GUI-0088] Register ces_4_4/clock Arrival: 219.267 ps
191
- [INFO GUI-0088] Register ces_4_5/clock Arrival: 223.542 ps
192
- [INFO GUI-0088] Register ces_7_0/clock Arrival: 287.112 ps
193
- [INFO GUI-0088] Register ces_7_1/clock Arrival: 283.560 ps
194
- [INFO GUI-0088] Register ces_7_2/clock Arrival: 273.081 ps
195
- [INFO GUI-0088] Register ces_7_3/clock Arrival: 272.904 ps
196
- [INFO GUI-0088] Register ces_6_0/clock Arrival: 250.403 ps
197
- [INFO GUI-0088] Register ces_6_1/clock Arrival: 246.424 ps
198
- [INFO GUI-0088] Register ces_6_2/clock Arrival: 235.488 ps
199
- [INFO GUI-0088] Register ces_6_3/clock Arrival: 230.871 ps
200
- [INFO GUI-0088] Register ces_6_4/clock Arrival: 234.553 ps
201
- [INFO GUI-0088] Register ces_6_5/clock Arrival: 238.808 ps
202
- [INFO GUI-0088] Register ces_7_4/clock Arrival: 240.051 ps
203
- [INFO GUI-0088] Register ces_7_5/clock Arrival: 244.454 ps
204
- """
205
-
206
- import re
207
- import matplotlib.pyplot as plt
208
- import numpy as np
209
-
210
- lines = text.strip().split('\n')
211
- filtered_lines = [line for line in lines if "ces" in line]
212
-
213
- # Finding the maximum indices for rows and columns
214
- max_row = max_col = 0
215
- data_dict = {}
216
- for line in filtered_lines:
217
- match = re.search(r"ces_(\d+)_(\d+)/clock\s*Arrival:\s*(\d+\.\d+)", line)
218
- if match:
219
- row, col, value = map(int, map(float, match.groups()))
220
- max_row = max(max_row, row)
221
- max_col = max(max_col, col)
222
- data_dict[(row, col)] = value
223
-
224
- # Constructing the data matrix
225
- data_matrix = np.zeros((max_row + 1, max_col + 1))
226
- for (row, col), value in data_dict.items():
227
- data_matrix[row, col] = value
228
-
229
- # Plotting the heatmap
230
- plt.imshow(data_matrix, origin='upper', cmap='viridis', aspect='auto')
231
- plt.colorbar(label='Arrival (ps)')
232
- plt.title('Heatmap of Arrival Times')
233
- plt.xlabel('Column')
234
- plt.ylabel('Row')
235
- plt.show()
236
- ```
237
-
238
-
239
-
240
- ### oharboe
241
- The reason I was interested in the skew heatmap as a function of location, is that for the mock-array, it is more important to have low skew at the edges, where the input/output pins are, than in the center... If I understand the presentation correctly...
242
-
243
- From the heatmap, I can see that I get the "inverse" of what I am looking for: I get high latency/skew at the edges and lowest clock network latency in the middle of the array.
244
-
245
-
246
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/2a02ed4e-16d1-40cf-87cb-b356f6942acb)
247
-
248
-
249
- ### oharboe
250
- Export to yaml everywhere? Then ask ChatGPT to write a program to graph whatever you want :-)
251
-
252
- ### oharboe
253
- One more round of tinkering, extracting leaf latency (macros & flip flops) for mock-array:
254
-
255
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/f9324f3c-9b7d-4186-b202-73ddd958e478)
256
-
257
- ```
258
- import numpy as np
259
- import matplotlib.pyplot as plt
260
- from scipy.interpolate import griddata
261
- from mpl_toolkits.mplot3d import Axes3D
262
-
263
- # Extract data from the table
264
- data = """
265
- [INFO GUI-0090] 353016 31894 _131_/CLK 2.5811525e-10
266
- [INFO GUI-0090] 353610 32704 _132_/CLK 2.578819e-10
267
- [INFO GUI-0090] 353448 29734 _133_/CLK 2.589441e-10
268
- [INFO GUI-0090] 353232 30004 _134_/CLK 2.5863567e-10
269
- [INFO GUI-0090] 353232 32164 _135_/CLK 2.5806945e-10
270
- [INFO GUI-0090] 353232 33514 _137_/CLK 2.5752672e-10
271
- [INFO GUI-0090] 352962 73474 _139_/CLK 2.4607233e-10
272
- [INFO GUI-0090] 353070 75094 _143_/CLK 2.455407e-10
273
- [INFO GUI-0090] 353232 74554 _145_/CLK 2.4536023e-10
274
- [INFO GUI-0090] 325488 49596 ces_0_7/clock 2.3215012e-10
275
- [INFO GUI-0090] 325488 97116 ces_1_7/clock 2.2571323e-10
276
- [INFO GUI-0090] 353070 30274 _136_/CLK 2.504758e-10
277
- [INFO GUI-0090] 353232 32974 _138_/CLK 2.5033728e-10
278
- [INFO GUI-0090] 353988 74824 _140_/CLK 2.468455e-10
279
- [INFO GUI-0090] 353070 74284 _141_/CLK 2.4607524e-10
280
- [INFO GUI-0090] 353124 73744 _142_/CLK 2.4646146e-10
281
- [INFO GUI-0090] 353718 74014 _144_/CLK 2.4688515e-10
282
- [INFO GUI-0090] 353070 75364 _146_/CLK 2.455602e-10
283
- [INFO GUI-0090] 282288 49596 ces_0_6/clock 2.3558674e-10
284
- [INFO GUI-0090] 354312 120724 _147_/CLK 2.396328e-10
285
- [INFO GUI-0090] 353070 120184 _148_/CLK 2.389033e-10
286
- [INFO GUI-0090] 353070 120724 _149_/CLK 2.3958044e-10
287
- [INFO GUI-0090] 353070 120454 _150_/CLK 2.3879088e-10
288
- [INFO GUI-0090] 352908 120994 _151_/CLK 2.3941804e-10
289
- [INFO GUI-0090] 354150 120994 _152_/CLK 2.3930402e-10
290
- [INFO GUI-0090] 353070 121264 _153_/CLK 2.387552e-10
291
- [INFO GUI-0090] 353070 121534 _154_/CLK 2.3849495e-10
292
- [INFO GUI-0090] 354312 167704 _155_/CLK 2.5028343e-10
293
- [INFO GUI-0090] 353070 167434 _156_/CLK 2.5006122e-10
294
- [INFO GUI-0090] 354312 167974 _157_/CLK 2.5054644e-10
295
- [INFO GUI-0090] 353070 167704 _158_/CLK 2.5003247e-10
296
- [INFO GUI-0090] 353502 168244 _159_/CLK 2.5076013e-10
297
- [INFO GUI-0090] 353070 167974 _160_/CLK 2.4982605e-10
298
- [INFO GUI-0090] 353070 168514 _161_/CLK 2.504333e-10
299
- [INFO GUI-0090] 353070 168784 _162_/CLK 2.4995647e-10
300
- [INFO GUI-0090] 325488 144636 ces_2_7/clock 2.3382915e-10
301
- [INFO GUI-0090] 282288 192156 ces_3_6/clock 2.4364535e-10
302
- [INFO GUI-0090] 325488 192156 ces_3_7/clock 2.3902733e-10
303
- [INFO GUI-0090] 23088 49596 ces_0_0/clock 2.908152e-10
304
- [INFO GUI-0090] 66288 49596 ces_0_1/clock 2.8722372e-10
305
- [INFO GUI-0090] 109488 49596 ces_0_2/clock 2.7635877e-10
306
- [INFO GUI-0090] 23088 97116 ces_1_0/clock 2.6215496e-10
307
- [INFO GUI-0090] 66288 97116 ces_1_1/clock 2.5834934e-10
308
- [INFO GUI-0090] 109488 97116 ces_1_2/clock 2.481484e-10
309
- [INFO GUI-0090] 152688 49596 ces_0_3/clock 2.2604922e-10
310
- [INFO GUI-0090] 152688 97116 ces_1_3/clock 2.2475981e-10
311
- [INFO GUI-0090] 195888 49596 ces_0_4/clock 2.4288263e-10
312
- [INFO GUI-0090] 239088 49596 ces_0_5/clock 2.4712393e-10
313
- [INFO GUI-0090] 195888 97116 ces_1_4/clock 2.3749708e-10
314
- [INFO GUI-0090] 239088 97116 ces_1_5/clock 2.492415e-10
315
- [INFO GUI-0090] 282288 97116 ces_1_6/clock 2.5338398e-10
316
- [INFO GUI-0090] 23088 144636 ces_2_0/clock 2.5319144e-10
317
- [INFO GUI-0090] 66288 144636 ces_2_1/clock 2.488505e-10
318
- [INFO GUI-0090] 109488 144636 ces_2_2/clock 2.3649802e-10
319
- [INFO GUI-0090] 23088 192156 ces_3_0/clock 2.6104152e-10
320
- [INFO GUI-0090] 66288 192156 ces_3_1/clock 2.5707544e-10
321
- [INFO GUI-0090] 109488 192156 ces_3_2/clock 2.4463978e-10
322
- [INFO GUI-0090] 152688 144636 ces_2_3/clock 2.2558208e-10
323
- [INFO GUI-0090] 195888 144636 ces_2_4/clock 2.394501e-10
324
- [INFO GUI-0090] 239088 144636 ces_2_5/clock 2.5129004e-10
325
- [INFO GUI-0090] 282288 144636 ces_2_6/clock 2.5545635e-10
326
- [INFO GUI-0090] 152688 192156 ces_3_3/clock 2.271353e-10
327
- [INFO GUI-0090] 195888 192156 ces_3_4/clock 2.4349614e-10
328
- [INFO GUI-0090] 239088 192156 ces_3_5/clock 2.4800043e-10
329
- [INFO GUI-0090] 353070 215224 _163_/CLK 2.5301994e-10
330
- [INFO GUI-0090] 354150 215764 _164_/CLK 2.5379573e-10
331
- [INFO GUI-0090] 352908 215764 _165_/CLK 2.546915e-10
332
- [INFO GUI-0090] 353016 215494 _166_/CLK 2.543563e-10
333
- [INFO GUI-0090] 352908 216034 _167_/CLK 2.5420016e-10
334
- [INFO GUI-0090] 353016 216304 _168_/CLK 2.53308e-10
335
- [INFO GUI-0090] 353070 216844 _169_/CLK 2.529667e-10
336
- [INFO GUI-0090] 353070 216574 _170_/CLK 2.5356642e-10
337
- [INFO GUI-0090] 282288 239676 ces_4_6/clock 2.362913e-10
338
- [INFO GUI-0090] 325488 239676 ces_4_7/clock 2.3206259e-10
339
- [INFO GUI-0090] 353556 262474 _171_/CLK 2.312914e-10
340
- [INFO GUI-0090] 353070 263014 _172_/CLK 2.3154262e-10
341
- [INFO GUI-0090] 353070 263284 _173_/CLK 2.3117369e-10
342
- [INFO GUI-0090] 353070 262744 _174_/CLK 2.3042566e-10
343
- [INFO GUI-0090] 353070 263824 _175_/CLK 2.3130639e-10
344
- [INFO GUI-0090] 353070 263554 _176_/CLK 2.3141243e-10
345
- [INFO GUI-0090] 353124 264094 _177_/CLK 2.3148382e-10
346
- [INFO GUI-0090] 354312 263554 _178_/CLK 2.3118188e-10
347
- [INFO GUI-0090] 353502 308644 _179_/CLK 2.4495708e-10
348
- [INFO GUI-0090] 353232 308914 _180_/CLK 2.421735e-10
349
- [INFO GUI-0090] 353394 309184 _181_/CLK 2.449023e-10
350
- [INFO GUI-0090] 352962 309454 _182_/CLK 2.428038e-10
351
- [INFO GUI-0090] 354204 309454 _183_/CLK 2.4457464e-10
352
- [INFO GUI-0090] 353664 309994 _184_/CLK 2.4385688e-10
353
- [INFO GUI-0090] 353070 309724 _185_/CLK 2.4294902e-10
354
- [INFO GUI-0090] 353664 310264 _186_/CLK 2.443524e-10
355
- [INFO GUI-0090] 353934 351844 _187_/CLK 2.53289e-10
356
- [INFO GUI-0090] 353016 353734 _189_/CLK 2.5345356e-10
357
- [INFO GUI-0090] 353016 354274 _193_/CLK 2.5375743e-10
358
- [INFO GUI-0090] 353556 352114 _194_/CLK 2.533773e-10
359
- [INFO GUI-0090] 353880 349684 _188_/CLK 2.4994018e-10
360
- [INFO GUI-0090] 353286 351574 _190_/CLK 2.502142e-10
361
- [INFO GUI-0090] 354042 349414 _191_/CLK 2.501929e-10
362
- [INFO GUI-0090] 353016 354004 _192_/CLK 2.5043992e-10
363
- [INFO GUI-0090] 282288 334716 ces_6_6/clock 2.3779426e-10
364
- [INFO GUI-0090] 325488 334716 ces_6_7/clock 2.374277e-10
365
- [INFO GUI-0090] 282288 382236 ces_7_6/clock 2.4838717e-10
366
- [INFO GUI-0090] 325488 382236 ces_7_7/clock 2.4390354e-10
367
- [INFO GUI-0090] 23088 239676 ces_4_0/clock 2.6987032e-10
368
- [INFO GUI-0090] 66288 239676 ces_4_1/clock 2.6595942e-10
369
- [INFO GUI-0090] 109488 239676 ces_4_2/clock 2.5412097e-10
370
- [INFO GUI-0090] 152688 239676 ces_4_3/clock 2.2762146e-10
371
- [INFO GUI-0090] 23088 287196 ces_5_0/clock 2.6053942e-10
372
- [INFO GUI-0090] 66288 287196 ces_5_1/clock 2.5651453e-10
373
- [INFO GUI-0090] 109488 287196 ces_5_2/clock 2.451736e-10
374
- [INFO GUI-0090] 152688 287196 ces_5_3/clock 2.2595725e-10
375
- [INFO GUI-0090] 195888 287196 ces_5_4/clock 2.334176e-10
376
- [INFO GUI-0090] 239088 287196 ces_5_5/clock 2.501299e-10
377
- [INFO GUI-0090] 282288 287196 ces_5_6/clock 2.6032337e-10
378
- [INFO GUI-0090] 325488 287196 ces_5_7/clock 2.6381633e-10
379
- [INFO GUI-0090] 195888 239676 ces_4_4/clock 2.1926745e-10
380
- [INFO GUI-0090] 239088 239676 ces_4_5/clock 2.2354223e-10
381
- [INFO GUI-0090] 23088 382236 ces_7_0/clock 2.871124e-10
382
- [INFO GUI-0090] 66288 382236 ces_7_1/clock 2.8355954e-10
383
- [INFO GUI-0090] 109488 382236 ces_7_2/clock 2.7308103e-10
384
- [INFO GUI-0090] 152688 382236 ces_7_3/clock 2.729042e-10
385
- [INFO GUI-0090] 23088 334716 ces_6_0/clock 2.504028e-10
386
- [INFO GUI-0090] 66288 334716 ces_6_1/clock 2.464243e-10
387
- [INFO GUI-0090] 109488 334716 ces_6_2/clock 2.3548827e-10
388
- [INFO GUI-0090] 152688 334716 ces_6_3/clock 2.3087125e-10
389
- [INFO GUI-0090] 195888 334716 ces_6_4/clock 2.3455274e-10
390
- [INFO GUI-0090] 239088 334716 ces_6_5/clock 2.388076e-10
391
- [INFO GUI-0090] 195888 382236 ces_7_4/clock 2.4005137e-10
392
- [INFO GUI-0090] 239088 382236 ces_7_5/clock 2.4445432e-10
393
- """
394
- rows = data.strip().split("\n")
395
- x = []
396
- y = []
397
- z = []
398
- for row in rows:
399
- _, _, x_val, y_val, _, z_val = row.split()
400
- x.append(float(x_val))
401
- y.append(float(y_val))
402
- z.append(float(z_val)/1E-12)
403
-
404
- # Prepare grid for interpolation
405
- xi = np.linspace(min(x), max(x), 100)
406
- yi = np.linspace(min(y), max(y), 100)
407
- xi, yi = np.meshgrid(xi, yi)
408
-
409
- # Interpolate the z values
410
- zi = griddata((x, y), z, (xi, yi), method='linear')
411
-
412
- # Plot the 3D heatmap
413
- fig = plt.figure()
414
- ax = fig.add_subplot(111, projection='3d')
415
- surf = ax.plot_surface(xi, yi, zi, rstride=1, cstride=1, cmap=plt.cm.jet,
416
- linewidth=0, antialiased=False)
417
-
418
- fig.colorbar(surf)
419
-
420
- plt.show()
421
- ```
422
-
423
-
424
- ### oharboe
425
- asap7/sha3 using python snippet above...
426
-
427
- Animation: https://drive.google.com/file/d/1k64ew7QkB9AeRaFs9ZOX09EqzeMZupMG/view
428
-
429
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/01ef0594-b8dd-4188-b35d-1978d9ec762e)
430
-
431
-
432
- ### maliberty
433
- @precisionmoon any thoughts here as I know you want more visualization as well.
434
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3812.md DELETED
@@ -1,34 +0,0 @@
1
- # Feature to OpenROAD that lets us drag views with the mouse?
2
-
3
- Tool: Graphical User Interface
4
-
5
- Subcategory: User interface enhancement
6
-
7
- ## Conversation
8
-
9
- ### oharboe
10
- It would make moving around the design a lot easier.
11
-
12
- Is there a way to do this already?
13
-
14
- The arrow keys is a bit clunky compared to dragging the view with the mouse.
15
-
16
- ### stefanottili
17
- OpenRoad will hopefully take hint's from KLayout's layout view controls instead of inventing yet another scheme.
18
-
19
- zoom in/out at the cursor position (mouse wheel zoom in/out)
20
- make the cursor key "shift width" adjustable (40% seem like a better value then the defaults)
21
- have "gestures" for fit, zoom in, zoom out (move mouse right-up, down, left-up), so no "mouse drag viewpoint".
22
-
23
- There doesn't seem to be a consensus for zoom in/out, fit, level down/up keys.
24
- I override the klayout defaults and use z/Z, f, l/L, curious what other people use.
25
-
26
- ### stefanottili
27
- I stand corrected, there is dragging the view in Klayout, on a Mac is "right mouse + shift"
28
-
29
- ### maliberty
30
- There is no method currently. It could be an enhancement request as could gestures.
31
-
32
- ### oharboe
33
- Added feature request https://github.com/The-OpenROAD-Project/OpenROAD/issues/3840
34
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3856.md DELETED
@@ -1,32 +0,0 @@
1
- # Release scheme
2
-
3
- Subcategory: Release management
4
-
5
- ## Conversation
6
-
7
- ### Blebowski
8
- Hello,
9
-
10
- I see there is a lot of development going on. I am wondering whether there are any opinions/plans about introducing a release scheme (e.g. once a month or quarterly, or after major feature changes / enhancements) ? Maintaining something like `CHANGES.md` or `NEWS.md` would be also very handy. Once user does not follow the development almost day-to-day, it is easy to get lost.
11
-
12
-
13
-
14
- ### QuantamHD
15
- I would be in favor of a news/changes file, but I'm a little bit more hesitant on slow releases.
16
-
17
- Precision Innovations provides a daily release https://github.com/Precision-Innovations/OpenROAD/releases which I feel like is a good cadence given the fast pace of developement.
18
-
19
- I think slower stable releases make sense when you have an entity that's willing to cherry pick bug fixes into older releases, which we don't have yet. I feel like if we embraced a slower release process at this point in time the net result would be people stuck with bugs for longer.
20
-
21
-
22
-
23
- ### mithro
24
- A monthly newsletter might be a good way to let people see the improvements.
25
-
26
- ### dralabeing
27
- Daily releases are good to get the best , updated and stable versions especially with prebuilt binaries from the Precision releases. I do think it would be good to add some sort of updates that are not yet in the docs or clear in the commit histories. For e.g mpl2 works well but this is not yet reflected in the default flow or docs but can be flagged in an Updates.md for users to try out --somewhat like a beta version. Similar any new features that are ready for use (partitioning, pin swapping, gate cloning etc.) can be mentioned along with a special release tag .
28
- Monthly update frequency should be good as in special cases when an update is done sooner on a major feature.
29
-
30
- ### rovinski
31
- I really like the [KLayout release notes](https://www.klayout.de/development.html) which have a list of PRs in that release (with hyperlinks) and a one-line description, as well as classifying as a bugfix or enhancement. It doesn't sound like we're moving from daily releases, but it could form a semi-automatically generated monthly newsletter or similar.
32
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/3939.md DELETED
@@ -1,33 +0,0 @@
1
- # Possible summary of changes format
2
-
3
- Subcategory: Release process enhancement
4
-
5
- ## Conversation
6
-
7
- ### vvbandeira
8
- While discussion #3856 has leaned to whether there should be a release and how frequent it should be, this discussion thread is to help guide the format and the process we would use to do such releases. We are considering using an automated way to generate "release notes" or at least a summary of recent changes. One proposed way to achieve this is to use PR titles and "Merge commit" messages. We would define and enforce a style of PR title, e.g., `[<TYPE>] <TOOL>: <SUMMARY>` where `[FEAT] drt: added shiny router cmd`. While `<TOOL>` is well defined (i.e., any module under `src/third-party` and `<SUMMARY>` will depend on the PR, what would be the most relevant values `<TYPE>` to include and their definition? As a starter point:
9
-
10
- - `[FEAT]` or `[NEW]`: for when we add new features, either in the form of Tcl/Python API changes -- e.g., new commands or flags.
11
- - `[DOCS]` or `[DOC]`: for documentation only changes.
12
- - `[BUGF]` or `[FIX]`: when fixing bugs or issues.
13
- - `[UPGR]` or `[ENH]`: upgrade/enhance -- not very sure about this, but it was what I thought with 4/3 letters to keep similar to above, mostly a self-imposed restriction, open to feedback.
14
-
15
-
16
- cc/ @maliberty @dralabeing @luarss @rovinski @QuantamHD
17
-
18
- ### maliberty
19
- I don't think of it as "summary of recent changes" as we still don't plan to do releases. <TOOL> is also optional as some changes may span multiple tools though that should be less common.
20
-
21
- ### rovinski
22
- I like the spirit, but I can foresee a few issues:
23
-
24
- 1. A PR might span multiple tools, in which case it's hard to accommodate the format
25
- 2. A PR might implement a feature only partially, in which case it's not ready for full use yet
26
- 3. A PR might span multiple types (e.g. both feature and enhancement).
27
- 4. If we want this scheme to stick, we will likely need a bot to enforce it or at least remind PR authors
28
-
29
- As such,
30
-
31
- * I think that enhancement and feature should probably just be one tag. I think PRs end up being both too often to distinguish
32
- * We need some sort of mechanism or identifying a feature as not ready for general use yet.
33
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/4420.md DELETED
@@ -1,65 +0,0 @@
1
- # Report total number of paths vs. number of failing path
2
-
3
- Tool: OpenSTA
4
-
5
- Subcategory: Reporting enhancement
6
-
7
- ## Conversation
8
-
9
- ### lustefan97
10
- I was wondering if there was a way to quickly summarize the ratio between the number of paths and number of failing paths (with regard to timing) using OpenSTA. I know it can report the slack in each path in a detailled manner, but I would like to have a report in the vein of : "Here is the number of paths that the design has that needs to be tested for timing vs. the number of path that failed the test"
11
-
12
- ### maliberty
13
- I don't see an easy way as the number of paths is exponential. How would you use the result?
14
-
15
- @tspyrou any thoughts?
16
-
17
- ### oharboe
18
- @lustefan97 Could this be a use-case for timing histogram(supported in OpenROAD under Charts).
19
-
20
- ### maliberty
21
- Note that to get the histograms you have to have qt charts installed. During cmake you should get a message like "Charts widget is enabled" and not "Charts widget is not enabled"
22
-
23
- ### maliberty
24
- > It is just a shame that it is so out if the way, I forget that we do... Can it be moved to the tabs where clocktree, hierarchy, etc. is?
25
-
26
- It is a docking window the same as all the others and can be position wherever you like.
27
-
28
- ### maliberty
29
- > Anyway: I think it would be great if this feature could be made more available out of the box and not hidden away behind "cmake magic" and curious defaults.
30
-
31
- It is on by default, but qtcharts is not available for centos7 and so can't be assumed.
32
-
33
- ### oharboe
34
- A less anaemic slack histogram example:
35
-
36
- ```
37
- make DESIGN_CONFIG=designs/asap7/aes/config.mk cts
38
- ```
39
-
40
- ```
41
- make DESIGN_CONFIG=designs/asap7/aes/config.mk gui_cts
42
- ```
43
-
44
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/1caeb16b-836f-4ae5-b5a6-6b235791ab90)
45
-
46
-
47
- ### lustefan97
48
- @maliberty @oharboe Ok so I updated my OpenROAD to the latest version while having QT chart installed and enabled. After loading a database in that I knew had reported a slack, I tried to make a timing chart but resulted in the GUI crashing.
49
-
50
- I got a warning telling me that timing data was not stored in the database and must be loaded separately if needed, are they not stored in the .odb ? Is the .odb file only "geometric" (would not be surprising since a .lef and .def files are needed), even if .lib files and RC rules have been used to produce the final result ?
51
-
52
- ### maliberty
53
- @AcKoucher FYI
54
-
55
- ### oharboe
56
- @AcKoucher Just one more example of where the horizontal labels are gone completely:
57
-
58
- ![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/7fa91044-d0df-4ffe-84b2-b5a264d959db)
59
-
60
-
61
- ### lustefan97
62
- @oharboe @maliberty I have a quick but important question, right now I run my TCL scripts on a device that has no GUI, how would I need to proceed if I wanted to design a chip on it using OpenROAD but then move the results and generate the histogram on another computer (capable of opening a GUI).
63
-
64
- I ran a test script in which I saved the design in a .odb file and the timing results from OpenSTA in a .sdf file, but when I loaded the .odb back in another GUI session (as well as the .sdf file using `read_sdf`), multiple warnings were raised during loading and no histogram was created...
65
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/4422.md DELETED
@@ -1,22 +0,0 @@
1
- # Creating an FEOL Class in OpenROAD
2
-
3
- Subcategory: FEOL integration
4
-
5
- ## Conversation
6
-
7
- ### chetanyagoyal
8
- This is in regard to [https://github.com/chipsalliance/ideas/blob/main/gsoc-2023-ideas.md#create-feol-classes-in-openroad-and-a-gds-writer](url), a GSoC project proposed in 2023. I would like to take this up for 2024's run of GSoC, and was **wondering what the project entails**. For reference, I have been contributing to OpenFASoC for a large part of the year and have good knowledge of their codebase, and some knowledge of OpenROAD-flow-scripts
9
-
10
- ### maliberty
11
- @msaligane was the mentor in 2023 so it would be good to get his input. I expect the hardest part would be dealing with hierarchy (aref/sref) as OR's odb is rather flat in its lef/def orientation. New layers shouldn't be too hard. The rest of the system will have to be more aware of them and the GUI will need some updating.
12
-
13
- I would be good to have a specific use case beyond just read/write gds. There should be some value added by doing it in OR as opposed to klayout or gdsfactory.
14
-
15
- ### stefanottili
16
- Having openroad write gds/oas would be useful, because it reduces the file size compared to odb.
17
-
18
- But I would recommend not to duplicate the ability to read and collect all the gds/oas pieces coming from various other sources. This functionality already exists in strm2oas.
19
-
20
- ### msaligane
21
- We have been looking into this in the previous GSOC. Hopefully, this project gets approved for Chetanya to finish the project.
22
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/4819.md DELETED
@@ -1,19 +0,0 @@
1
- # gui highlight_inst
2
-
3
- Tool: Graphical User Interface
4
-
5
- Subcategory: Enhance GUI functionality
6
-
7
- ## Conversation
8
-
9
- ### gyu511
10
- hello.
11
- In the gui, highlight_inst only supports up to 8 groups, so I would like to expand it to support more groups.
12
- So I added to HighlightGroupDialog in OpenROAD/src/gui/src/highlightGroupDialog.cpp and added color in OpenROAD/src/gui/include/gui/gui.h and finally button in OpenROAD/src/gui/ui/highlightGroupDlg.ui I also added it, but it doesn't work. What should I do?
13
-
14
- ### maliberty
15
- Can you send a pointer to your branch with the changes? How does it not work?
16
-
17
- ### maliberty
18
- Once this works I hope you would be willing to make a PR so we can incorporate it in OR.
19
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/5041.md DELETED
@@ -1,23 +0,0 @@
1
- # Write macro placement position to a tcl file
2
-
3
- Tool: Hierarchical Macro Placement
4
-
5
- Subcategory: Macro placement output
6
-
7
- ## Conversation
8
-
9
- ### vijayank88
10
- Similar to `write_pin_placement` option, is there a option to write macro placement position while using automatic macro placement.
11
- So that macro position will be re-used for some specific requirement with manual macro placement again.
12
-
13
- ### maliberty
14
- This is more of an issue than a discussion
15
-
16
- ### maliberty
17
- https://github.com/The-OpenROAD-Project/OpenROAD/issues/5043
18
-
19
- ### AcKoucher
20
- @vijayank88 FYI, there's the `write_macro_placement` flag in mpl2. However it's currently deprecated due to a new OR command for general macro placement writing that was implemented in #4557.
21
-
22
- I've opened #5046 to update the flag so that mpl2 will write the entire macro placement instead of only those macros that it placed.
23
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
gh_discussions/Feature Request/5420.md DELETED
@@ -1,24 +0,0 @@
1
- # dft command names
2
-
3
- Tool: Design for Test
4
-
5
- Subcategory: Command naming
6
-
7
- ## Conversation
8
-
9
- ### ThomasGeroudet
10
- Hello @maliberty, @fgaray, I've just seen dft commands, but not tried it yet.
11
- The name of the commands seemed familiar, and I was wondering if it could be better to rename them to avoid potential legal issues such as this one:
12
- https://news.synopsys.com/2016-12-21-Synopsys-Obtains-Injunction-in-Copyright-Suit-Against-ATopTech
13
- In case of renaming, maybe all dft commands could start with dft_* so that it's easy to find them ?
14
-
15
- ### fgaray
16
- Yeah, changing the names SGTM to avoid this issue.
17
-
18
- Do we know what other tools use to name this commands? (Cadence for example). I tried looking online but could not find Cadence's related docs.
19
-
20
- dft_insert, dft_scan_replace, dft_preview SGTM
21
-
22
- ### maliberty
23
- Thanks, we want to avoid any such conflicts. I am not aware of the proprietary tool names but am fine with a rename to avoid such.
24
-