Discern top and bottom boundaries.

[soaringxmc]

@p-costa
I am mergingCaNS commits to CaLES.
Could you please expolain a bit what this commit is for, and when do I need this commit?
CaNS-World/CaNS@55b9a4e

[p-costa]

Hi Mao,

There will be some updates to CaNS soon, as I prepare version 3.0, then 3.1. If you'd like, I can give you a heads up in a call to talk about the differences.

About that commit, we are simply splitting the storing of the viscous momentum fluxes from the top and bottom boundaries, consistent with what we do for the scalar transport case (see scal.f90). The implications are not significant, but for your case, you do have the variable eddy viscosity, so these terms are slightly different. I hope this is clear.

[soaringxmc]

Hi Pedro:
Thank you for the explaination. I will definitely follow your future updates for CaLES updates.

[soaringxmc]

Hi, Pedro: Could we run CaNS on AMD GPUS, e.g., LUMI? Best, Mao

…

On Thu, Apr 3, 2025 at 4:44 PM Pedro Costa ***@***.***> wrote: Hi Mao, There will be some updates to CaNS soon, as I prepare version 3.0, then 3.1. If you'd like, I can give you a heads up in a call to talk about the differences. About that commit, we are simply splitting the storing of the viscous momentum fluxes from the top and bottom boundaries, consistent with what we do for the scalar transport case (see scal.f90). The implications are not significant, but for your case, you do have the variable eddy viscosity, so these terms are slightly different. I hope this is clear. — Reply to this email directly, view it on GitHub <#18 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AZM3BKN352K2KNOE7TGYWCT2XVCLXAVCNFSM6AAAAAB2LOJFGCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONZWGAZTCMJXG4> . You are receiving this because you authored the thread.Message ID: ***@***.***> [image: p-costa]*p-costa* left a comment (CaNS-World/CaLES#18) <#18 (comment)> Hi Mao, There will be some updates to CaNS soon, as I prepare version 3.0, then 3.1. If you'd like, I can give you a heads up in a call to talk about the differences. About that commit, we are simply splitting the storing of the viscous momentum fluxes from the top and bottom boundaries, consistent with what we do for the scalar transport case (see scal.f90). The implications are not significant, but for your case, you do have the variable eddy viscosity, so these terms are slightly different. I hope this is clear. — Reply to this email directly, view it on GitHub <#18 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AZM3BKN352K2KNOE7TGYWCT2XVCLXAVCNFSM6AAAAAB2LOJFGCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONZWGAZTCMJXG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[p-costa]

Hi Mao @soaringxmc,

We can run CaNS on LUMI, in our prototyped version 3.1, but it is not public yet. Do you want to wait for the release, or, if you are in a hurry, we can share with you the private repo.

[soaringxmc]

Hi @p-costa
I am asking for post-doc Kiran Satheeshchandran at KTH, who should be using FLUTAS. I'll let him know this.

[p-costa]

Ok. Kiran has also reached out to me before about other matters.

As a side note, I suggest not using Issues on repositories to discuss these matters, as we are off-topic!