[Nek5000-users] Increase nek5000 problem size
nek5000-users at lists.mcs.anl.gov
nek5000-users at lists.mcs.anl.gov
Tue Feb 21 08:34:40 CST 2017
Note that the checkpoint file size will scale as the number of elements in your
problem (i.e., the number of elements in the .rea file: grep -i nel eddy_uv.rea to see this).
The checkpoint file will not scale as lelt in the SIZE file.
The requirement for lelt is that:
P*lelt > nel
where P is the number of processors used in the simulation.
You also need the global number of elements, lelgt > nel.
________________________________
From: nek5000-users-bounces at lists.mcs.anl.gov [nek5000-users-bounces at lists.mcs.anl.gov] on behalf of nek5000-users at lists.mcs.anl.gov [nek5000-users at lists.mcs.anl.gov]
Sent: Monday, February 20, 2017 11:46 PM
To: nek5000-users at lists.mcs.anl.gov
Subject: [Nek5000-users] Increase nek5000 problem size
Hi,
I'm new to use nek5000, I have a problem about checkpoint file size. I did test for eddy and set like 1x1=12, 1xd=18, lelt=3000, checkpoint file size is 0.7 MB, it seems very small, If I continue increase memory size, I cannot compile nek5000 correctly, I will get an error like bellow:
obj/bdry.o: In function `trstax_':
bdry.f:(.text+0x10f6): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
bdry.f:(.text+0x14f4): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
bdry.f:(.text+0x15c7): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
bdry.f:(.text+0x1793): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
obj/bdry.o: In function `nekasgn_':
bdry.f:(.text+0x20d7): relocation truncated to fit: R_X86_64_32S against symbol `screv_' defined in COMMON section in obj/postpro.o
bdry.f:(.text+0x20e8): relocation truncated to fit: R_X86_64_32S against symbol `screv_' defined in COMMON section in obj/postpro.o
obj/coef.o: In function `area3_':
coef.f:(.text+0x417): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
coef.f:(.text+0x46f): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
coef.f:(.text+0x477): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
coef.f:(.text+0x682): relocation truncated to fit: R_X86_64_32S against symbol `ctmp0_' defined in COMMON section in obj/navier5.o
coef.f:(.text+0x6ef): additional relocation overflows omitted from the output
I already set
# enable BIG MEMORY suuport
BIGMEM="true"
So, if there any other way to to increase problem size? Thanks.
Best,
Kai
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/nek5000-users/attachments/20170221/24ded092/attachment.html>
More information about the Nek5000-users
mailing list