Remove volk_16u_byteswap_a_orc

[argilo]

This protokernel was enabled in #719, but it appears to be broken on M1, so let's remove it for now.

Fixes #745.

[jdemel]

@argilo could the address in the homebrew build error be the culprit:
stop reason = EXC_BAD_ACCESS (code=2, address=0x100044000) ?
It's a wild guess but it might be an issue with the int16 uint16 type thing that you fixed in another commit.
In any case, the leading 1 in the address made me suspicious.

[jdemel]

On a different note: Github seems to offer paid M1 CI on large/xlarge instances. Someone posted an ad in a corresponding thread and referred to FlyCI. This might be an option for us. At least, it might improve our process.

[argilo]

It's a wild guess but it might be an issue with the int16 uint16 type thing that you fixed in another commit.

I expect #740 is the change you had in mind. It also touched volk_16u_byteswap_a_orc in this release, so it makes sense to treat it with suspicion.

[argilo]

Without a stack trace or debug symbols it's hard to know for sure what's going on. I hope tha @chenrui333 might be able to help diagnose this.

[argilo]

I converted this to a draft until we have confirmation that this protokernel is the cause of #745.

[argilo]

I used FlyCI to build VOLK on M1 with ORC support, and didn't observe any failures after trying out several configurations. I'll close this PR off for now, since there's not much reason to suspect that volk_16u_byteswap_a_orc was the cause of the test failure that Homebrew observed.

Without diagnostic information, I'm afraid there's not much more I can do to troubleshoot.