From 01/10/2019 to 02/08/2019
- I have a case in which I'm able to _Verilate_ @<subblock>@ with little trouble (< 2 min., < 1 Mbyte).
Then, I try to...
- 01:21 AM Usage: RE: "File truncated" error from ar when attempting to create V<my_block>__ALL.a ?!
- Aha! I just tried building up the archive file by hand and noticed that it starts failing (giving the "File truncated...
- 01:01 AM Usage: RE: "File truncated" error from ar when attempting to create V<my_block>__ALL.a ?!
- In that case likely one of the objects got truncated by gcc, perhaps you missed an error?
You should be able to e....
- 12:56 AM Usage: RE: "File truncated" error from ar when attempting to create V<my_block>__ALL.a ?!
- Hmm, I just tried rebuilding the archive again (It's been a couple hours since the compilation finished.) and got the...
- 12:49 AM Usage: RE: "File truncated" error from ar when attempting to create V<my_block>__ALL.a ?!
- You're probably on a network disk (NFS) and a compile has exited but hasn't finished flushing the .o to disk when the...
- I'm getting the following error at the end of an otherwise successful translation/compilation:...
- 02:36 PM Usage: RE: evaluating verilator
- It means that the construct itself that it is warning about might be something that will simulate differently. The W...
- 01:34 PM Usage: RE: evaluating verilator
- Hi Wilson,
Regarding this excerpt from your comment above:
> Verilator will generally give a UNOPTFLAT warning ...
- 11:26 PM Usage: RE: How to pack arrays properly in my C++ test bench?
- Just for safety added some asserts if the sizes aren't as expected, probably should have had those earlier...
- 11:24 PM Usage: RE: How to pack arrays properly in my C++ test bench?
- Ah, okay, I see that that type does use 32 bits:...
- 11:15 PM Usage: RE: How to pack arrays properly in my C++ test bench?
- What OS? The real question is is the vluint32_t 32 bits or 64? That's what the definition uses, not a long.
- 06:55 PM Usage: RE: How to pack arrays properly in my C++ test bench?
That's not what I'm seeing:...
- 12:38 AM Usage: RE: How to pack arrays properly in my C++ test bench?
- A long is 32 bits, not 64, or there's something wrong with the compiler's setup in verilatedos.h.
- 12:35 AM Usage: RE: How to pack arrays properly in my C++ test bench?
- Hi Wilson,
Thanks for your reply!
Sorry, I don't quite understand what you mean.<br>
It's clear I have twice a...
- 04:02 PM Usage: RE: FST dumping 100x slower than VCD
- Check you compiled all code including the gtkwave/*.c files with -O3. Assuming still bad, please compile and link wi...
- I gave FST dumping a shot today... but the simulation speed was quite dramatically impacted -- well over 100x slowdow...
- 11:17 PM Usage: RE: How to pack arrays properly in my C++ test bench?
- Yes, it rounds to nearest 8/16/32/64 or n*32 word for speed, pack that with zeros then index the array by 0..18.
- I'm confused about how to properly pack arrays in my C++ test bench, because the port types in my resultant @V<foo>.h...
- 12:25 AM Usage: RE: Multiple comma-separated declaration in one for-loop
- I applied code similar to your patch to trunk as it seemed a good general improvement. I changed it slightly to matc...
- 11:46 PM Usage: RE: Is VM_USER_DIR defined incorrectly in Verilator-generated makefiles?
- Maybe, and probably should have been. But, I did some research on changing this and it broke some usages.
- 07:17 PM Usage: RE: Compile errors with Mingw32 on fstapi.c
- Fix for space concatenation submitted upstream.
- 02:57 PM Usage: RE: Incremental verilate on large designs?
- Not currently, sorry.
It's not really incremental, but one method is to manually piece your design into SystemC mo...
- I'm wondering if there are any strategies or features to reduce build time when small changes are made to large desig...
- 10:15 AM Usage: RE: Multiple comma-separated declaration in one for-loop
- Hello, I find supporting this is difficult since I have little experience about compilers.
I try to work on this but...
- Is it possible that @VM_USER_DIR@ is being defined/used incorrectly in _Verilator_-generated makefiles?<br>
- 10:55 PM Development: RE: interface modport expression
- I would like to 2nd this motion.
Modports Expressions, let alone Modports, aren't particularly common. But the pro...
- 12:24 AM Usage: RE: Several questions, re: the `--inline-mult` option.
- 1. An internal ast node E.g. add.
2. Right, will fix.
3. It has been used on larger designs however gcc has bee...
- 12:03 AM Usage: RE: Can I call out to a foreign language module, via VPI, from a Verilog submodule?
- Not easily, I think you are looking for something like this:
- 11:47 PM Usage: RE: What's a good value to give the --output-split-cfuncs flag?
- Maybe start with 50000, see how long it takes, and adjust to get e.g. 5 minute gcc per c++ file.
- Any advice on what value to give in the @--output-split-cfuncs@ flag?
- In thinking about how to work around this limitation whereby we're unable to use pre-Verilated *.o files to provide V...
- I have some questions, re: the @--inline-mult@ option, that I hoped someone might know the answers to:
# What is a...
- 02:05 AM Usage: RE: Best practice for compiling top-level test bench and linking final executable, when NO...
- --exe shouldn't be slowing anything down, it just adds a make rule, you can put your .cpp's on the command line and u...
- My design is so big (2 hour _Verilator_ translation/compilation time) that I'm not using the @--exe@ option.<br>
- 10:32 PM Usage: RE: error: 'VL_CPU_RELAX' was not declared in this scope.
- D'oh! Sorry, just me being stupid.<br>
I forgot to refresh an Emacs buffer.<br>
Thanks very much for your time, and...
- 06:46 PM Usage: RE: error: 'VL_CPU_RELAX' was not declared in this scope.
- It does seem inconsistent. If you delete your vrltr_out and rebuild is it OK?
- 06:09 PM Usage: RE: error: 'VL_CPU_RELAX' was not declared in this scope.
- Thanks, Wilson!
Does the @-DVL_THREADED@ flag get put in the @V<blkname>_classes.mk@ file?
I'm looking in that fi...
- 02:25 AM Usage: RE: error: 'VL_CPU_RELAX' was not declared in this scope.
- I think you aren't passing in the flags that verilator puts into the generated makefile, specifically -DVL_THREADED
- I just started trying to use multi-threading and am getting this error, when attempting to compile my top-level C++ t...
Also available in: Atom