makenek error

271 views
Skip to first unread message

Sai Pranay

unread,
Jul 18, 2024, 7:35:36 AM7/18/24
to Nek5000
Hello

I am new to NEK5000 and I am trying to get a 2D mesh smoother implemented as shown in the file below.

If my understanding is correct, running NEK5000 would give a 2D smoothed mesh. I wish to preserve the resolution near the wall. 
Mesh is designed in gmsh and converted to re2 format and then to rea format and generating the corresponding map file.

When I am trying to execute makenek I am getting an error at the end as shown in the build.log. Let me know the source of the error.

Thank you
Pranay
SIZE
final.usr
final2.usr
build.log

YuHsiang Lan

unread,
Jul 18, 2024, 12:36:44 PM7/18/24
to Nek5000
Hi,

You need to add "#include "experimental/meshsmoother.f" at the beginning of the usr file.

Hope this helps,
Yu-Hsiang
--

Sai Pranay

unread,
Jul 19, 2024, 10:16:23 AM7/19/24
to Nek5000
Thank you for the message. I manged to exeute makenek using the .usr file and when I set the lelt around 5K. But the actual mesh demands lelt around 7K. When I tried to increase lelt by reducing mpi or increasing lelg makenek fails with the error below. Tried updating the FFLAGS/CFLAGS to large and still no use.

Error:
Warning: Type mismatch between actual argument at (1) and actual argument at (2) (INTEGER(8)/INTEGER(4)).
building libnek5000.a ... ar: `u' modifier ignored since `D' is the default (see `U')
done
mpif77  -O2 -cpp -fdefault-real-8 -fdefault-double-8 -std=legacy  -DMPI -DUNDERSCORE -DGLOBAL_LONG_LONG -DTIMER -I/home/sai/Desktop/NEK5000/Nek5000/run/final -I/home/sai/Desktop/NEK5000/Nek5000/core -I./ -I /home/sai/Desktop/NEK5000/Nek5000/core/experimental -o nek5000 /home/sai/Desktop/NEK5000/Nek5000/core/drive.f /home/sai/Desktop/NEK5000/Nek5000/run/final/final.o libnek5000.a -L/home/sai/Desktop/NEK5000/Nek5000/3rd_party/blasLapack -lblasLapack -L/home/sai/Desktop/NEK5000/Nek5000/3rd_party/gslib/lib -lgs -Wl,--allow-multiple-definition
/home/sai/Desktop/NEK5000/Nek5000/run/final/final.o: in function `fixcurs_':
final.f:(.text+0x31e0): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in libnek5000.a(navier5.o)
final.f:(.text+0x31fb): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in libnek5000.a(navier5.o)
final.f:(.text+0x3216): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in libnek5000.a(navier5.o)
final.f:(.text+0x325c): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in libnek5000.a(navier5.o)
final.f:(.text+0x3287): relocation truncated to fit: R_X86_64_PC32 against symbol `ctmp0_' defined in COMMON section in libnek5000.a(navier5.o)
libnek5000.a(drive2.o): in function `plan3_vol_':
drive2.f:(.text+0x5ce): relocation truncated to fit: R_X86_64_PC32 against symbol `scrvh_' defined in COMMON section in libnek5000.a(convect.o)
libnek5000.a(drive2.o): in function `plan4_vol_':
drive2.f:(.text+0x877): relocation truncated to fit: R_X86_64_PC32 against symbol `scrvh_' defined in COMMON section in libnek5000.a(convect.o)
drive2.f:(.text+0x916): relocation truncated to fit: R_X86_64_PC32 against symbol `scrvh_' defined in COMMON section in libnek5000.a(convect.o)
libnek5000.a(drive2.o): in function `plan2_vol_':
drive2.f:(.text+0x5f88): relocation truncated to fit: R_X86_64_PC32 against symbol `scrvh_' defined in COMMON section in libnek5000.a(convect.o)
libnek5000.a(plan5.o): in function `midstep_':
plan5.f:(.text+0x5a): relocation truncated to fit: R_X86_64_PC32 against symbol `scrvh_' defined in COMMON section in libnek5000.a(convect.o)
plan5.f:(.text+0x9d): additional relocation overflows omitted from the output
collect2: error: ld returned 1 exit status
make: *** [makefile:133: nek5000] Error 1

Is it beacuse of my system limitations, this is not happening?

Thank you
Pranay
final.usr

YuHsiang Lan

unread,
Jul 19, 2024, 11:39:45 AM7/19/24
to Nek5000
HI,

> mpif77  -O2 -cpp -fdefault-real-8 -fdefault-double-8 -std=legacy  -DMPI -DUNDERSCORE -DGLOBAL_LONG_LONG -DTIMER -I/home/sai/Desktop/NEK5000/Nek5000/run/final -I/home/sai/Desktop/NEK5000/Nek5000/core -I./ -I /home/sai/Desktop/NEK5000/Nek5000/core/experimental -o nek5000 /home/sai/Desktop/NEK5000/Nek5000/core/drive.f /home/sai/Desktop/NEK5000/Nek5000/run/final/final.o libnek5000.a -L/home/sai/Desktop/NEK5000/Nek5000/3rd_party/blasLapack -lblasLapack -L/home/sai/Desktop/NEK5000/Nek5000/3rd_party/gslib/lib -lgs -Wl,--allow-multiple-definition
/home/sai/Desktop/NEK5000/Nek5000/run/final/final.o

I don't find -mcmodel=large in your compile command.
Can you post your makenek? and SIZE?

> Is it beacuse of my system limitations, this is not happening?
What's the size of your RAM?

Thanks,
Yu-Hsiang
--

Sai Pranay

unread,
Jul 19, 2024, 11:57:26 AM7/19/24
to Nek5000
Here you go :

Architecture:            x86_64
  CPU op-mode(s):        32-bit, 64-bit
  Address sizes:         39 bits physical, 48 bits virtual
  Byte Order:            Little Endian
CPU(s):                  12
  On-line CPU(s) list:   0-11
Vendor ID:               GenuineIntel
  Model name:            Intel(R) Xeon(R) W-1250P CPU @ 4.10GHz
    CPU family:          6
    Model:               165
    Thread(s) per core:  2
    Core(s) per socket:  6
    Socket(s):           1
    Stepping:            5
    CPU max MHz:         4800.0000
    CPU min MHz:         800.0000
    BogoMIPS:            8199.79

Thanks
Pranay
makenek
SIZE

Sai Pranay

unread,
Jul 19, 2024, 12:02:22 PM7/19/24
to Nek5000
RAM availability: 64 GB
        64008 M total memory
         6974 M used memory
         9781 M active memory
        29160 M inactive memory
        21557 M free memory

Thanks
Pranay

Fischer, Paul

unread,
Jul 19, 2024, 12:43:28 PM7/19/24
to Sai Pranay, Nek5000
Lelt is the upper bound on the number of elements per rank.  The local (per rank) memory scales as O(lelt), to leading order.

I suggest to cut lelt by a factor of two and then use twice as many MPI ranks.

Or, a factor of 4 and use 4X the number of ranks, etc.

hth,
Paul


From: nek...@googlegroups.com <nek...@googlegroups.com> on behalf of Sai Pranay <padakan...@gmail.com>
Sent: Friday, July 19, 2024 11:02 AM
To: Nek5000 <nek...@googlegroups.com>
Subject: [nek5000] Re: makenek error
 
--
You received this message because you are subscribed to the Google Groups "Nek5000" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nek5000+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nek5000/425a4bbb-f767-4f9a-a3c2-f19f16603aaen%40googlegroups.com.

YuHsiang Lan

unread,
Jul 19, 2024, 12:55:41 PM7/19/24
to Nek5000
Hi,

It looks like you didn't uncomment these lines in your makenek
You will need to recompile with a full clean (makenek clean)
  FFLAGS="-mcmodel=large"
  CFLAGS="-mcmodel=large"

Hope this helps,
Yu-Hsiang
--

Sai Pranay

unread,
Jul 20, 2024, 3:59:03 AM7/20/24
to Nek5000
Thank you this worked.  
Reply all
Reply to author
Forward
0 new messages