Hello,
I am trying to run GaMD-PPI through Amber24. My system is relatively large, which meant that I needed to update some atom limits in the $AMBERHOME/amber/pmemd24_src/src/pmemd/src/gti_controlVariable.i file according to the instructions here: http://archive.ambermd.org/202412/0029.html. Specifically, I made the following changes:
MaxNumberTIAtom = 750 -> 18750
MaxNumberTIPair = 300 -> 7500
MaxNumberREAFAtom = 500 -> 12500
Before making these changes, I was able to compile Amber24 without issue. After making these changes, I cannot compile Amber24 anymore; I run into various vague errors while doing "make install" such as:
CMake Error at pmemd_cuda_DPFP_generated_gti_NBList_kernels.cu.o.RELEASE.cmake:278 (message):
Error generating file
/lus/eagle/projects/FoundEpidem/cseitz/amber/pmemd24_src/build/src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_DPFP.dir//./pmemd_cuda_DPFP_generated_gti_NBList_kernels.cu.o
And more errors like this:
make[2]: *** [src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_SPFP.dir/build.make:847: src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_SPFP.dir/pmemd_cuda_SPFP_generated_gti_NBList_kernels.cu.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:2672: src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_DPFP.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
make[1]: *** [CMakeFiles/Makefile2:2620: src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_SPFP.dir/all] Error 2
[ 28%] Building CXX object AmberTools/src/emil/CMakeFiles/emil_mpi.dir/linearAssignment.C.o
[ 28%] Building CXX object AmberTools/src/emil/CMakeFiles/emil_mpi.dir/mcMoves.C.o
ptxas error : Entry function '_Z25kgBuildTI2TINBList_kernelv' uses too much shared data (0x36f78 bytes, 0x10000 max)
ptxas error : Entry function '_Z35kgBuildSpecial2RestNBPreList_kernelN6GTI_NB11specialTypeE' uses too much shared data (0x36f78 bytes, 0x10000 max)
CMake Error at pmemd_cuda_DPFP_mpi_generated_gti_NBList_kernels.cu.o.RELEASE.cmake:278 (message):
Error generating file
/lus/eagle/projects/FoundEpidem/cseitz/amber/pmemd24_src/build/src/pmemd/src/cuda/CMakeFiles/pmemd_cuda_DPFP_mpi.dir//./pmemd_cuda_DPFP_mpi_generated_gti_NBList_kernels.cu.o
Any ideas on how to increase the atom limits for GaMD-PPI while still being able to compile Amber24 would be much appreciated! Thanks in advance~
Best,
Christian
Christian Seitz
PhD, Dept. of Computer Science | UChicago
cgseitz.uchicago.edu<mailto:cgseitz.uchicago.edu>
[www.linkedin.com/in/christianseitz21]<
http://www.linkedin.com/in/christianseitz21>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Sep 30 2025 - 16:00:02 PDT
