Readon

it would be better to avoid chars that already have meanings.'|' is a logic operator represent or.

how about '*

> I agree with you. Nice.^^

If it is possible may be it's better to continue with #9086 to support tensor parallel. Then treat numa node as a standalone computation unit and speed up by column...

I think levelBridge is not enough for this case. The Node natively is act in one stage, so the valid and ready would be managed only by one cycle case....

I have tried one at #1459 . it seems that your code can simplify it.

So, the assumption is that sh.exe is installed in the environment. It would be best to test this on various cases. Windows with MSYS2 setting, MSYS2, Linux, or even MacOS.

Unfortunately, I could not make it work on MSYS2 environment. It reports ``` Can't exec "$MSYS2_ROOT$/share/verilator//verilator_bin": No such file or directory at /mingw64/bin/verilator line 188.er / Test / testOnly 90s...

> Unfortunately, I could not make it work on MSYS2 environment. It reports > > ``` > Can't exec "$MSYS2_ROOT$/share/verilator//verilator_bin": No such file or directory at /mingw64/bin/verilator line 188.er /...

In some cases, it would not stop too. That's why I always recommand do not use the version 5.