-
Notifications
You must be signed in to change notification settings - Fork 51
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Adds an option to override the real precision to r4 #68
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
While this looks good for Intel on c5, what about the other templates?
@bensonr Should all templates be updated? |
@uramirez8707 - I think we should make this available in the same manner for all compiler templates contained within the mkmf system. The SHiELD model and build system also support the NVidia HPC Toolkit and GCC compilers. As the SHiELD build system is phased out in favor of the updated fre-cli, we need to ensure any templates used by this new system have the proper support. |
templates/ncrc5-intel-classic.mk
Outdated
ifdef USE_R4 | ||
REAL_PRECISION := -r4 | ||
CPPDEFS += -DOVERLOAD_R4 | ||
else | ||
REAL_PRECISION := -r8 | ||
endif | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bensonr
should we be using -real-size 32
and real-size 64
instead?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@uramirez8707 - Yes, I think we should be using the updated -real-size
for both OneAPI and classic ifort.
This can be used in an xmls by doing
For the component that you want to use r4 for.
This option will allow us to compile different components at different precisions (i.e compile fv3 in r4 and everything else in r8)