Allowing torchani to use GPU

[kexul]

Description

Added config for torchani to use GPU, related to this issue #326.
Now we can use local_options={'device':'cuda'} to assign the device:

import qcengine as qcng
import qcelemental as qcel

mol = qcel.models.Molecule.from_data("""
O  0.0  0.000  -0.129
H  0.0 -1.494  1.027
H  0.0  1.494  1.027
""")

inp = qcel.models.AtomicInput(
    molecule=mol,
    driver="energy",
    model={"method": "ani2x"}
    )

ret = qcng.compute(inp, "torchani", local_options={'device':'cuda'})
print(ret.return_result)

[loriab]

Thanks for the PR, @kexul . @dotsdl, how does this relate to how OpenMM or OpenFF in general switches cpu/gpu modes?

[codecov]

Codecov Report

Merging #327 (d7ced82) into master (0612e81) will increase coverage by 0.00%.
The diff coverage is 66.66%.

[WardLT]

Instead of having the user manually turn on GPU usage, could we use torch to detect the presence of a GPU and automatically use it?

…

On Tue, Nov 2, 2021, 11:47 AM codecov[bot] ***@***.***> wrote: Codecov <https://codecov.io/gh/MolSSI/QCEngine/pull/327?src=pr&el=h1&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=MolSSI> Report Merging #327 <https://codecov.io/gh/MolSSI/QCEngine/pull/327?src=pr&el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=MolSSI> (a4dec42 <a4dec42>) into master <https://codecov.io/gh/MolSSI/QCEngine/commit/0612e81a2ce6685866ea40692fa95649b06e2b35?el=desc&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=MolSSI> (0612e81 <0612e81>) will *decrease* coverage by 36.48%. The diff coverage is 50.00%. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#327 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABVNYDLGFNEXG4UFEAORLSLUKAW53ANCNFSM5HGOVWZQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[kexul]

Instead of having the user manually turn on GPU usage, could we use torch to detect the presence of a GPU and automatically use it?

That sounds reasonable, I've changed the code according to your suggestion.

[WardLT]

The changes look great to me. Thanks for the contribution!

Once you fix the linting problems with make format, this should be ready to merge.

@loriab , do you want to review this as well? Or, can I click the button?

[WardLT]

Introducing the whitespace here probably caused a problem with our linter.

Could you run make format and issue a commit with the changed files?

[kexul]

Done by d7ced82 😉

[loriab]

Thanks @WardLT , merge at will.