ljedrz

@wanyvic while this suggestion is being considered, have you tried the workaround described [here](https://github.com/AleoHQ/snarkOS/issues/1574#issuecomment-1015430278)?

@howardwu do we still need to keep this branch open? The same applies to https://github.com/AleoHQ/snarkOS/pull/1139.

Just to be sure, you're using `nvtop` to check the GPU utilization, correct?

> This seems to be the same I experienced and reported on #1555 In this case the node is `Mining`, though.

This shouldn't happen with the `--release` flag since https://github.com/AleoHQ/snarkOS/pull/541. I just checked it locally, and testing with that flag didn't cause an OOM (or any memory use growth whatsoever).

Due to the nature of what the aforementioned PR was targeting, it shouldn't be OS-specific; I was trying to reproduce this by running tests with input values and I can...

Ok, so outside of `cargo test` I can indeed reproduce this and I've found the culprit; it's actually very similar to the `TestConstraintSystem` that got interning and cleanups in the...

@0rphon I have [a branch](https://github.com/ljedrz/leo/tree/fix_circuit_synthesizer_ooms) with a prototype for a fix; I no longer get OOMs with your example when I use it, can you try it on your machine?

The fix I prepared (which assumes that intermediate `ConstraintSystem` objects for the `CircuitSynthesizer` can be cleaned up once they go out of their scope) is targeting the compilation phase; for...

Confirmed: a similar fix is needed in `snarkvm`; I'm now testing a prototype, will report soon.