[WIP][CI] Moved Sanitized Build Test from Odin to Parmys

[AlexandreSinger]

The sanitized build test was testing the Odin build and not the parmys build (which is the default). Moved the test to Parmys.

Resolves #3191

[AlexandreSinger]

It looks like there is some issues with the Sanitized build and parmys. I may have to investigate.

I do not think the sanitized build should even be running Parmys since we should really be debugging our own code. I think we should make a special test suite for the sanitized build which will run only VPR (i.e. will only use pre-generated blif files). To discuss.

[vaughnbetz]

Right now there is an error if we try to compile parmys with the sanitizers; it's a .so error so I think it is likely that parmys doesn't have the right flags passed to it to build the sanitized version. @loglav03 : if you have time to take a look that would be great, and adding @AmirhosseinPoolad as he tends to be very good at sorting build issues like this. Goal is to fix the error and convert the basic sanitized run to be on parmys instead of odin II.
@haydar-c : until this is fixed, you'd need to add the flat-recon very small test to the odin basic tests to get sanitized runs on them.

[AmirhosseinPoolad]

I investigated this a little bit. For some reason parmys is being compiled with '-g -fsanitize=address -fsanitize=leak -fuse-ld=gold -static-libasan' (Look at build/build.ninja) when by all means it shouldn't (we're adding the parmys subdirectory before setting the various compile options for VPR). I'm not quite sure why but my suspicion is that all the PUBLIC stuff in various cmakefiles has something to do with it (parmys depends on a bunch of vtr internal libraries which do get compiled with asan).

I'm also not entirely sure why it's erroring out. I think we're compiling yosys without asan (which is good and intended) but yosys is trying to dynamically load parmys.so which has some asan function declarations without any definitions. The loader then complains about this and crashes the program.

I don't really have any idea how to fix any of this. We might need a cmake wizard.