Update minimum Fortran versions

[mathomp4]

Given recent events in MAPL and GOCART, I'm proposing increasing the minimum compilers supported by ESMA_cmake to:

• Intel 18.0.5
• GNU 9.2.0
• NAG 7.0

[mathomp4]

@LiamBindle A query. Do you use any compilers older that what are proposed here.

[LiamBindle]

Yes for GCC. We frequently use GCC 8.3. What version of MAPL would this new requirement be introduced?

It's understandable that these things are necessary every once in a while, and GCC 10 has been out for some time now. As long as it truely is a necessary upgrade, it's okay on our end. I can post in our documentation that GCC 8.3 support is deprecated and will be removed in a future version of GCHP (when we upgrade to this version of MAPL).

For Intel our minimum is also 18.0.5, and we haven't ever supported NAG.

[mathomp4]

@LiamBindle Interesting. If you can run with GCC 8 then maybe that means even GFE can? I now invoke the name of @tclune to see if he thinks some of his upcoming GFE code will work with GCC 8. If it can, well, we can not up the minimum version of GCC. If not, then we can up the requirement when new GFE starts being used in MAPL.

I guess I figure "No one uses GCC 8. It's old!" 😄 Though I suppose GNU would prefer users use GCC 8.5...

[tclune]

I think that there are isolated bits of GFE that will fail under GCC8, but quite possibly (even likely) they are things that just won’t matter for most users of MAPL. E.g., not being able to use fHamcrest in pfunit. The real issue will be the next gen YAML. I can attempt to see if that works under gfortran 8 or 9, but I’m not terribly optimistic. I do have it working under 10 and 11 at the moment though. MAPL does not use this yet, but once I roll it out, that will change. * Tom From: Matthew Thompson ***@***.***> Reply-To: GEOS-ESM/ESMA_cmake ***@***.***> Date: Monday, September 20, 2021 at 10:00 AM To: GEOS-ESM/ESMA_cmake ***@***.***> Cc: "Clune, Thomas L. (GSFC-6101)" ***@***.***>, Mention ***@***.***> Subject: [EXTERNAL] Re: [GEOS-ESM/ESMA_cmake] Update minimum Fortran versions (#191) @LiamBindle<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FLiamBindle&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478476070%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=yXJEtg7dr6JCDbYMAMt0J5aCgftLOpcjICB%2Bndl5ZPo%3D&reserved=0> Interesting. If you can run with GCC 8 then maybe that means even GFE can? I now invoke the name of @tclune<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Ftclune&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478476070%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=MZSIZIip2JFkosesdRNhxdN%2F0jJ%2FqX50xZD9tDUw2Y4%3D&reserved=0> to see if he thinks some of his upcoming GFE code will work with GCC 8. If it can, well, we can not up the minimum version of GCC. If not, then we can up the requirement when new GFE starts being used in MAPL. I guess I figure "No one uses GCC 8. It's old!" 😄 Though I suppose GNU would prefer users use GCC 8.5... — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FGEOS-ESM%2FESMA_cmake%2Fpull%2F191%23issuecomment-922954899&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478486025%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=CDpXRUa0dtQIr7utYqfDk1DpY8ZzyKdco9OCHbPtYmI%3D&reserved=0>, or unsubscribe<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FABPP7YB5B5I6EZ33REKEDGDUC45AXANCNFSM47UZJI5Q&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478486025%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=PsAnvuxZ0GE948iZKMIGsgIctdz05wIbFJOWjUCP16o%3D&reserved=0>. Triage notifications on the go with GitHub Mobile for iOS<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fapps.apple.com%2Fapp%2Fapple-store%2Fid1477376905%3Fct%3Dnotification-email%26mt%3D8%26pt%3D524675&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478495984%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=DncdL7qOcpVatFwkoEFCSMuj6UURH80I51NP2Qy1bvI%3D&reserved=0> or Android<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fplay.google.com%2Fstore%2Fapps%2Fdetails%3Fid%3Dcom.github.android%26referrer%3Dutm_campaign%253Dnotification-email%2526utm_medium%253Demail%2526utm_source%253Dgithub&data=04%7C01%7Cthomas.l.clune%40nasa.gov%7C21e8404b4e764f31e5c508d97c3f0baa%7C7005d45845be48ae8140d43da96dd17b%7C0%7C0%7C637677432478495984%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=UEGHhCJTmWEcyCrQTjChxWql5pzcKMgMBywNg3rRyEQ%3D&reserved=0>.

[tclune]

Well, yafyaml is passing the unit tests with 9.4. I'll try soon with 8.3

[tclune]

Wow - gfortran 8.5 also passes the unit tests for yafyaml! I'm surprised given all the difficulties I had getting things to work with Intel and NAG.

[mathomp4]

Wow - gfortran 8.5 also passes the unit tests for yafyaml! I'm surprised given all the difficulties I had getting things to work with Intel and NAG.

8.5 is pretty new so maybe they backported all the bug fixes back to it??

[scivision]

On certain other projects, I have experienced that GCC 9.2 fails at compile time due to its bugs (later 9.x releases work). Hopefully they don't impact your project--I just use other versions like GCC 8.5 or 9.x > 9.2 etc.

[mathomp4]

On certain other projects, I have experienced that GCC 9.2 fails at compile time due to its bugs (later 9.x releases work). Hopefully they don't impact your project--I just use other versions like GCC 8.5 or 9.x > 9.2 etc.

Ooh. I'd forgotten about this PR. I should review it. Frankly, we know our code is not happy even with anything less than GCC 10+. Maaaaaybe 9.5 would work? But we have enough fancy Fortran in MAPL that we probably need to be at 10.3 or so and higher.