装Quartus Lite 20.1, 在Windows WSL

26 Jun 2023

电脑坏了真烦。

在Windows WSL下没法运行32-bit程序，所以quartus最少也得20.1版本。

装完运行quartus，先设置路径。

~/.bashrc

export PATH=$PATH:~/intelFPGA_lite/20.1/quartus/bin
export PATH=$PATH:~/intelFPGA_lite/20.1/modelsim_ase/bin

有几个lib没有。

sudo apt-get install -y libsm6 libxext6 libxrender-dev

Then the following errors showed up.

warning: setlocale: LC_CTYPE: cannot change locale (en_US.UTF-8): No such file or directory
realloc(): invalid pointer
Aborted (core dumped)

To solve the first warning, run:

sudo dpkg-reconfigure locales

It brings a text based GUI. Choose “en_US.UTF-8”.

Now the realloc() problem. At first I thought this would solve it.

Greetings!

I spent a couple days chewing on this, and eventually found https://github.com/chriz2600/quartus-lite. The interesting thing about their implementation is that they’re using tcmalloc (http://goog-perftools.sourceforge.net/doc/tcmalloc.html) instead of the stock malloc, which is unusual enough that I had to try it. Sure enough, installing libtcmalloc-minimal4 and then LD_PRELOADing it did the trick:

export LD_PRELOAD=/usr/lib/libtcmalloc_minimal.so.4 ${QUARTUS_PATH}/nios2eds/nios2_command_shell.sh ./build.sh -r ${_revision} -s ${_size}

I’ve also found it beneficial(?) to delete ${QUARTUS_PATH}/quartus/linux64/libboost_system.so, libccl_curl_drl.so, and libstdc++.so.6, leaving it to use the stock Ubuntu versions instead (which are installed by the build-essential package, I believe). This is mostly because of suggestions for solving another error:

Inconsistency detected by ld.so: dl-close.c: 811: _dl_close: Assertion `map->l_init_called’ failed!

Hopefully this is useful information! Everything is working great with Quartus 17.0 for me, so far.

So I tried.

$ sudo apt install libtcmalloc-minimal4
$ export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libtcmalloc_minimal.so.4

Turns out, this is a quartus issue with WSL on Windows.

https://community.intel.com/t5/Intel-FPGA-Software-Installation/Running-Quartus-Prime-Standard-on-WSL-crashes-in-libudev-so/m-p/1189032

I don’t know what happened in my case, but I did the same thing.

$ sudo apt install libudev-dev

$ export LD_PRELOAD=/lib/x86_64-linux-gnu/libudev.so     <--- this is not necessary

Solved.

The newest version of modelsim is 20.1.1 and it still is 32-bit code.

Can’t run 32-bit programs on WSL Windows. Tried many things, didn’t work.

https://vhdlwhiz.com/modelsim-quartus-prime-lite-ubuntu-20-04/

It seems to be a pain in the ass.

It went smooth on a real machine with ubuntu22.04.

  $ ./ModelSimSetup-20.1.1.720-linux.run 
  $ sudo dpkg --add-architecture i386
  $ sudo apt-get update
  $ sudo apt-get install libc6:i386 libncurses5:i386 libstdc++6:i386
  $ sudo apt-get install libXext:i386
  $ sudo apt-get install libxtst6:i386
  $ sudo apt-get install libxft-dev:i386

晕，后来发现，windowsd的wsl是2，但是里面装的ubuntu是wsl1的。

wsl --list --verbose

  NAME            STATE           VERSION
* Ubuntu-22.04    Running         1

后来试了wsl --set-version Ubuntu-22.04 2，想把ubuntu升级到wsl2，一直不成功。

最后发现，windows里没装hyper-v。

装上以后就好了。

现在wsl里ubuntu也能运行32bit程序了。

u@home:~$

Archive

About

RSS

装Quartus Lite 20.1, 在Windows WSL