I only just made the connection of the WebGL 2.0 tests being flaky and the use of the Core Profile. I have a feeling that when Chrome uses the Core Profile in the GPU process, either more GPU process hangs and associated watchdog timeouts are seen, or there are some random crashes in the GPU process.

I think we need to instrument the test harness ASAP to try to symbolize these GPU process crashes. I'll work on this. In the meantime if more flakiness is seen on mac_chromium_rel_ng in the WebGL 1.0 conformance tests we may need to revert the switch. We may also need to mark the entire WebGL 1.0 conformance suite flaky.

I've been triaging the WebGL 1.0 and 2.0 conformance flakes in  Issue 596622  and  Issue 619264 . At this point, I don't think there is persistent flakiness associated with the switch to the Core Profile. This particular test hasn't flaked in a while. Closing this as WontFix (not reproducible).