aziesemer / astran Goto Github PK
View Code? Open in Web Editor NEWASTRAN - Automatic Synthesis of Transistor Networks
Home Page: http://aziesemer.github.io/astran
License: BSD 2-Clause "Simplified" License
ASTRAN - Automatic Synthesis of Transistor Networks
Home Page: http://aziesemer.github.io/astran
License: BSD 2-Clause "Simplified" License
Descobrir como remover as dependências de uma interface gráfica quando usando a flag --shell
. Usar parametro deveria deixar de necessitar o uso da GUI. O problema foi visto quando executamos o astran remotamente por ssh sem usarmos a flag -X
para habilitar a GUI.
Poly routing in the top and bottom of the cell
Original issue reported on code.google.com by [email protected]
on 25 Oct 2013 at 12:22
OFFSET different from half pitch
Original issue reported on code.google.com by [email protected]
on 24 Oct 2013 at 12:37
Check if they belong to the same net before insert poly spacing rule.
Original issue reported on code.google.com by [email protected]
on 25 Oct 2013 at 11:34
Oi Adriel!
Consegui abrir as células no Cadence, mas ao fazer o DRC dá problemas com:
=> o tamanho mínimo de NPLUS e PPLUS que deve ser 0.6 (são os PPLUS e
NPLUS que ficam na parte inferior do GND e superior do VDD, no final da
célula);
=> o enclosure da VIA1 que deve ser 0.15 (deve ter uma metal 2 junto com a
VIA1);
=> o espaço mínimo entre NDIFF (difusão N) e o NTUB deve ser de 2.6;
=> e tem em algumas células que não respeita o espaço entre metal 1 que
deve ser de 0.45.
Obrigada!
Original issue reported on code.google.com by [email protected]
on 29 Jul 2009 at 1:40
Attachments:
Insert bodycentre search, some kind of reordering, steiner point technique...
Original issue reported on code.google.com by [email protected]
on 31 Jul 2009 at 1:11
The viewer doesn't open when trying to view large layouts in Mac Os (Windows ?)
Original issue reported on code.google.com by [email protected]
on 16 Jul 2009 at 5:44
Google recently released a manufacturable 130nm PDK which can be found at https://github.com/google/skywater-pdk and is also offering free tape outs for open source designs at https://efabless.com/open_shuttle_program
It would be interesting to see how you astran generated cells compare to the existing cells that are provided.
There is likely to be a 90nm PDK in the future too.
The README mentioned it was open source but I can find any reference to what LICENSE the source is released under?
Violating in contacts
Original issue reported on code.google.com by [email protected]
on 26 Oct 2013 at 9:35
Attachments:
- Insert Save Rules to file .rul inside File/Save Menu. Create respective
command.
- Edit layer_labels in rules.h inside Rules Window
Original issue reported on code.google.com by [email protected]
on 15 Jul 2009 at 7:25
The tool is violating this rule in some cases
Original issue reported on code.google.com by [email protected]
on 25 Oct 2013 at 12:21
Is being violated in small W transistors.
Original issue reported on code.google.com by [email protected]
on 26 Oct 2013 at 9:36
Attachments:
Create different perfis for intracell placement & routing costs and generate
all them simultaneously. At the end, the best one (with smaller compaction
cost, is chosen as winner.
Original issue reported on code.google.com by [email protected]
on 13 Aug 2014 at 8:40
Do not put each folded transistors in paralel when there are 2 or more
transistors in serie.
Original issue reported on code.google.com by [email protected]
on 5 Dec 2013 at 7:22
Attachments:
Is there any documentation or guide for creating rule file for Astran other than the image provided with the tool?
It's being violated in some situations (to the right)
Original issue reported on code.google.com by [email protected]
on 26 Oct 2013 at 9:33
Attachments:
The ILP model get some variables without number: e.g. xa, yb,...
Original issue reported on code.google.com by [email protected]
on 25 Oct 2013 at 12:14
I installed Astran for all users in our lab as a regular CAD tool. In this way, the users can use astran anywhere they are and not only from the instalation folder.
The problem with this is related to where Astran look for it own files and configurations. Take a look:
This problem is easly fixed creating a environment variable like ASTRAN_PATH, which Astran can use to search for the files. This can also be used to fix the issue #35.
That solution can also be used to improve the initial astran.cfg file like this:
set viewer "${ASTRAN_PATH}/Tools_MacIntel/wxSightGL.app/Contents/MacOS/wxSightGL"
set rotdl "${ASTRAN_PATH}/Tools_MacIntel/rotdl"
set placer "${ASTRAN_PATH}/Tools_MacIntel/mobralplace"
set lpsolve "${GUROBI_PATH}/bin/gurobi_cl"
set log history.log
set verbose_mode 1
With this solution, it is possible to change the astran lpsolver easly changin the GUROBI_PATH variable.
Please, consider to improve the astran to make it more flexible and professional with this feature. If you need help, I can try to help you with this feature.
PPLUS and NPLUS
Original issue reported on code.google.com by [email protected]
on 24 Oct 2013 at 12:37
Will reduce delay, specially for poly connections.
Original issue reported on code.google.com by [email protected]
on 13 Aug 2014 at 8:38
My last astran version (3.9.8) shows:
ASTRAN - IC Physical Design Tool v3.9.4 - Mon Oct 5 17:00:58 2015
I think you should consider to create a way sinc the github version number and the tool version number. This number is really useful to report bugs and ask for features.
Arquivo de log. Adriel, o arquivo history.log
salva apenas os comandos executados né? Seria interessante colocar nele o resultado de cada comando pra facilitar o debug. O que você acha?
Existe alguma forma de setar o nome dos devices? Nossos scripts trocam todos os nomes dos dispositivos pra PMOS e NMOS, poderia ser algo assim, se não tiver. O que tu acha?
Ah, o str_Name
poderia ser uma lista assim, poderia aceitar inicialmente todas as variações possiveis incluindo as variações para Vth's diferentes. E no futuro multiplos Vth's poderiam ser adicionados no layout.
SET PDEVICE <str_Name>
SET NDEVICE <str_Name>
Add a new item in the Circuit window to allow the user to change the design
name.
Original issue reported on code.google.com by [email protected]
on 17 Jul 2009 at 4:05
Olá!
Na versão para Windows está dando problema na hora de fazer o placement ou
às vezes quando manda ele compactar o layout. O erro está em anexo.
Normalmente acontece com as células de tamanho maior, se tiver uma com
tamanho 1 e outra com tamanho 3, dá erro em todas de tamanho 3.
Por exemplo:
DFECP3, DFEP3, DFP3, DFS3, DFSC3, DFSCP3, DFSE3, DFSEC3, DFSEP3 ...
IMAJ31, IMUX21 e demais
INV4 a INV15
JK3, JKC3, JKCP3 ...
MAK32
MUX só deu certo com tamanho 1
NAND24, 28, 32, 34, 42, 43 ...
NOR22, 23, 24 ...
Original issue reported on code.google.com by [email protected]
on 22 Sep 2009 at 3:40
Attachments:
Lembrete pra mim: Arrumar o makefile do wxSightGL para ele funcionar em qualquer lugar
# Dependencies: Ubuntu 14.xx/16.xx
sudo apt install libglew1.13 libglew-dev
sudo apt install freeglut3 freeglut3-dev
sudo apt install libwxgtk2.8-0 libwxgtk2.8-dev
# Compilando:
make
# Dependencies: CentOS 5.10
sudo yum install glew-devel.x86_64 glew.x86_64
sudo yum install freeglut.x86_64 freeglut-devel.x86_64
# Compilando:
make CCFLAGS="`pkg-config --cflags --libs gtk+-2.0 glib-2.0` `wx-config --cppflags`"
Hi @aziesemer, it is possible to fix this issue?
While generating cell layout I'm getting this error at command compact layout:
Error, Unable to open input file 'TimeLimit=3600'
Usage of ../Tools_Linux/lp_solve version 5.5.2.0:
../Tools_Linux/lp_solve [options] [[<]input_file]
List of options:
-h prints this message
-lp read from LP file (default)
-mps read from MPS file, default in fixed format
-mps_free use free format
-mps_ibm interprete integers accoring to ibm format
-mps_negobjconst negate objective constant
-fmps read from MPS file in free format
-rpar filename read parameters from filename.
Any workaround or solution?
Thanks in advance
The contents of Preferences and Cellgen windows are bigger than their windows
size.
Original issue reported on code.google.com by [email protected]
on 18 Jul 2009 at 2:58
Is it possible to still use lpsolve with astran? It seems that Gurobi is required by Astran, but for our application, an OpenSource solver is required. Would it be possible to add lpsolve support in again?
Also please document the system requirements (e.g. Gurobi) in the README.
Hi, guys.
The link to what I assume was an image of a layout generated by Astran in the README.md is broken...
https://gme-ufrgs.googlecode.com/files/screenshot2.png
Could you fix that?
It would be nice to have an image of a layout generated by Astran right there when someone accesses the GitHub page.
O Astran retorna 1 quando termina. Isso tá ok Adriel? Aqui no linux estou esperando um 0.
http://www.eda.ncsu.edu/wiki/FreePDK15:Contents
Original issue reported on code.google.com by [email protected]
on 11 Jun 2014 at 1:34
Hi Adriel @aziesemer ,
Thanks a lot for your excellent work! Recent Gurobi stops providing the download link of Gurobi 5.0 and now people can only find Gurobi 8.0-9.5. I am resolving the problem and a PR will be submitted.
What steps will reproduce the problem?
1. Load Netlist
2. Save Netlist
The netlist file (*.lib) loads cell's information like instances tree and
nets. However, when you save the netlist to a .lib file, this kind of
information is lost. It is not included to the file due to the data structure.
What to do:
Change the data structure to include these information in the netlist file.
Original issue reported on code.google.com by [email protected]
on 31 Aug 2009 at 9:56
Modify our spice parser to suport these netlists.
Original issue reported on code.google.com by [email protected]
on 30 Jul 2009 at 6:05
Violating in long vertical/horizontal lines
Original issue reported on code.google.com by [email protected]
on 26 Oct 2013 at 9:34
Attachments:
New standardcell templates
Original issue reported on code.google.com by [email protected]
on 24 Oct 2013 at 12:38
Create this button in the ICPD toolbar.
Original issue reported on code.google.com by [email protected]
on 16 Jul 2009 at 7:21
Where I can find a help with all the Astran commands?
É possivel aidicionar uma forma de ignorar comentários no arquivo de teconlogia? Isso facilita o processo de portar/modifica/adaptar esse arquivo. Normalmente ele ignora os comandos que nao entende, mas se achar o caractere de comentário, (por exemplo #
) ele pode simplesmente evitar dizer essa informação na tela.
Create command "SAVE HISTORY <filename>" to save all previous commands to a file
Original issue reported on code.google.com by [email protected]
on 15 Jul 2009 at 6:35
Save configuration, library, rules, routing and placement inside a folder so
that it can be opened
again and restore its state.
Original issue reported on code.google.com by [email protected]
on 31 Jul 2009 at 1:13
Adriel, tu já conhece o Read The Docs? É uma plataforma pra fazer documentações. O astran podia ter uma documentação lá. Toda documentação é gerada por arquivos texto como os README.md
do github. Ele tem integraçào com o github tb e faz tuido automático. O bacana é que cada página é um arquivo texto com alguma tags simples. No fim, ele gera uma documentação reobusta, organizada e com um sistema de busca que facilita muito. Aqui tem um exemplo do da documentação do sublime text http://docs.sublimetext.info/en/latest/index.html. Acho interessante tu pensar nesse tipo de coisa. Que tu acha?
Show Layouts as hierarchy instead of flat in Layout list window. This will
speed up the resfresh
function.
Original issue reported on code.google.com by [email protected]
on 16 Jul 2009 at 5:59
Save routing result to an output file (rotdl format and Cadence in the future)
so that an external
router can be executed and the result can be opened again and compacted.
2 new entries to the Route menu will be needed: clear and compact.
Original issue reported on code.google.com by [email protected]
on 16 Jul 2009 at 6:04
This is a DRC violation that is very easy to fix by hand but I haven't figured
out how to do it efficiently using ILP. I'm aware of this issue and will try to
fix in the future.
Original issue reported on code.google.com by [email protected]
on 3 Nov 2013 at 1:47
Attachments:
Instances in the TOP hierarchy (outside a subckt) is not saved.
Original issue reported on code.google.com by [email protected]
on 16 Jul 2009 at 7:36
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.