离线安装CP2K2024.1

1138711019

请教一下，顺利按完了2024.2版本，也设置好了环境变量，把之前的注释掉了。为啥mpirun -np 128 cp2k.popt -i X.inp 出来计算的还是之前的版本啊？

1138711019

1138711019 发表于 2024-8-29 17:27
请教一下，顺利按完了2024.2版本，也设置好了环境变量，把之前的注释掉了。为啥mpirun -np 128 cp2k.popt - ...

解决了，犯了个低级错误

薛定谔的老猫

configure: error: in `/home/yczhu/Soft/cp2k-2024.1/tools/toolchain/build/elpa-2023.05.001/build_cpu':
configure: error: C++ compiler cannot create executables
See `config.log' for more details
make: *** No targets specified and no makefile found.  Stop.
make: *** No rule to make target `install'.  Stop.
/home/yczhu/Soft/cp2k-2024.1/tools/toolchain/scripts/tool_kit.sh: line 663: /home/yczhu/Soft/cp2k-2024.1/tools/toolchain/install/elpa-2023.05.001/install_successful: No such file or directory
请问有人遇到这个问题吗？

zmjsce

感谢楼主，我在centos 7.9上按照您的方法安装成功了~

goldNAN

薛定谔的老猫 发表于 2024-9-25 21:16
configure: error: in `/home/yczhu/Soft/cp2k-2024.1/tools/toolchain/build/elpa-2023.05.001/build_cpu' ...

C++没编译好

zjxitcc

薛定谔的老猫 发表于 2024-9-25 21:16
configure: error: in `/home/yczhu/Soft/cp2k-2024.1/tools/toolchain/build/elpa-2023.05.001/build_cpu' ...

如果你使用GNU系列编译器（如gcc-9.2）编译CP2K，编译前应运行以下6行命令，检查是否存在相应的编译器，以及版本是否一致
which gcc
which g++
which gfortran
gcc --version
g++ --version
gfortran --version
哪一个有问题 就不用编译CP2K了，先把编译器问题解决。（虽然也可以让CP2K工具链现场安装一个gcc，但你没交待细节，我默认你机器上已经有某种gcc了）

薛定谔的老猫

zjxitcc 发表于 2024-9-29 10:51
如果你使用GNU系列编译器（如gcc-9.2）编译CP2K，编译前应运行以下6行命令，检查是否存在相应的编译器， ...

好的，就是使用CP2K工具连接的gcc，把gcc的包下载到了buIld文件里面，先确认下编译器的问题，感谢！

zjxitcc

薛定谔的老猫 发表于 2024-9-29 16:11
好的，就是使用CP2K工具连接的gcc，把gcc的包下载到了buIld文件里面，先确认下编译器的问题，感谢！

如果你是让CP2K现编译一个gcc，那出现这种错误就是gcc没编译成功，要往前找报错。

ziv

想问一下，安装完成后，有一些文件可以删除吗。在超算上，文件数量有限制

Santz

abin 发表于 2024-3-26 14:09
这玩意不是我编译的.

无法提供协助.

问下 CentOS7.9，3.10.0-1160.el7.x86_64 内核能否正常支持NFSoRDMA。我最近在MLNX_OFED_LINUX-23.10-2.1.3.1 驱动下测试，该驱动手册里明确支持该内核的， ib 速度测试也正常。但是在实际配置中发现，NFS server 端 rpcinfo 中并未见 20049 端口。很奇怪，客户端 rdma 挂载可以正常挂载，但是实际测试效果不佳。我在 NVIDIA 官方看到很多说法是NFSoRDMA is not supported in the Mellanox OFED version 3.4 and higher，但是 23.10驱动安装时明确添加--add-kernel-support --with-nfsrdma，从其 help 来看又是支持NFSoRDMA的。

还有一点就是我发现 rdma 挂载的满载一段时间 slurm 计算节点有可能会挂掉，显示的是内核级别的报错，我怀疑3.10.0-1160.el7.x86_64 内核的NFSoRDMA存在问题。故想咨询下老师 CentOS7.9，3.10.0-1160.el7.x86_64 内核的NFSoRDMA配置情况/方式。

abin

Santz 发表于 2025-3-31 10:06
问下 CentOS7.9，3.10.0-1160.el7.x86_64 内核能否正常支持NFSoRDMA。我最近在MLNX_OFED_LINUX-23.10-2.1 ...

使用支持CentOS 7.9的版本.

完毕.


补充，如果使用太新的卡，老系统未必能用。

zhj45318403

贴主您好，我在按照您的帖子步骤正常安装无报错后尝试提交服务器进行计算，在日志.err文件中发现大量
[cnode01:1181191] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181194] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181190] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181182] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181185] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181165] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181201] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181186] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181168] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
[cnode01:1181196] mca_base_component_repository_open: unable to open mca_btl_openib: libosmcomp.so.3: cannot open shared object file: No such file or directory (ignored)
请问是什么问题呢？谢谢您

goldNAN

不是软件问题，可能 MPI不兼容某些高端设备，比如高速IB卡。  
参考 http://bbs.keinsci.com/thread-31660-1-1.html

五月雨

请问楼主，我按照您的方法在Centos 7.6安装，进行到Getting proc arch info using OpenBLAS tools时总是出现这个问题：

1. etarch_2nd.c:22:35: error: ‘SGEMM_DEFAULT_UNROLL_M’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_UNROLL_M’?
   
2.    22 |     printf("SGEMM_UNROLL_M=%d\n", SGEMM_DEFAULT_UNROLL_M);
   
3.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
   
4.       |                                   SBGEMM_DEFAULT_UNROLL_M
   
5. getarch_2nd.c:22:35: note: each undeclared identifier is reported only once for each function it appears in
   
6. getarch_2nd.c:23:35: error: ‘SGEMM_DEFAULT_UNROLL_N’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_UNROLL_N’?
   
7.    23 |     printf("SGEMM_UNROLL_N=%d\n", SGEMM_DEFAULT_UNROLL_N);
   
8.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
   
9.       |                                   SBGEMM_DEFAULT_UNROLL_N
   
10. getarch_2nd.c:24:35: error: ‘DGEMM_DEFAULT_UNROLL_M’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_M’?
    
11.    24 |     printf("DGEMM_UNROLL_M=%d\n", DGEMM_DEFAULT_UNROLL_M);
    
12.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
13.       |                                   XGEMM_DEFAULT_UNROLL_M
    
14. getarch_2nd.c:25:35: error: ‘DGEMM_DEFAULT_UNROLL_N’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_N’?
    
15.    25 |     printf("DGEMM_UNROLL_N=%d\n", DGEMM_DEFAULT_UNROLL_N);
    
16.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
17.       |                                   XGEMM_DEFAULT_UNROLL_N
    
18. getarch_2nd.c:29:35: error: ‘CGEMM_DEFAULT_UNROLL_M’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_M’?
    
19.    29 |     printf("CGEMM_UNROLL_M=%d\n", CGEMM_DEFAULT_UNROLL_M);
    
20.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
21.       |                                   XGEMM_DEFAULT_UNROLL_M
    
22. getarch_2nd.c:30:35: error: ‘CGEMM_DEFAULT_UNROLL_N’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_N’?
    
23.    30 |     printf("CGEMM_UNROLL_N=%d\n", CGEMM_DEFAULT_UNROLL_N);
    
24.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
25.       |                                   XGEMM_DEFAULT_UNROLL_N
    
26. getarch_2nd.c:31:35: error: ‘ZGEMM_DEFAULT_UNROLL_M’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_M’?
    
27.    31 |     printf("ZGEMM_UNROLL_M=%d\n", ZGEMM_DEFAULT_UNROLL_M);
    
28.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
29.       |                                   XGEMM_DEFAULT_UNROLL_M
    
30. getarch_2nd.c:32:35: error: ‘ZGEMM_DEFAULT_UNROLL_N’ undeclared (first use in this function); did you mean ‘XGEMM_DEFAULT_UNROLL_N’?
    
31.    32 |     printf("ZGEMM_UNROLL_N=%d\n", ZGEMM_DEFAULT_UNROLL_N);
    
32.       |                                   ^~~~~~~~~~~~~~~~~~~~~~
    
33.       |                                   XGEMM_DEFAULT_UNROLL_N
    
34. getarch_2nd.c:79:50: error: ‘SGEMM_DEFAULT_Q’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_Q’?
    
35.    79 |     printf("#define SLOCAL_BUFFER_SIZE\t%ld\n", (SGEMM_DEFAULT_Q * SGEMM_DEFAULT_UNROLL_N * 4 * 1 *  sizeof(float)));
    
36.       |                                                  ^~~~~~~~~~~~~~~
    
37.       |                                                  SBGEMM_DEFAULT_Q
    
38. getarch_2nd.c:80:50: error: ‘DGEMM_DEFAULT_Q’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_Q’?
    
39.    80 |     printf("#define DLOCAL_BUFFER_SIZE\t%ld\n", (DGEMM_DEFAULT_Q * DGEMM_DEFAULT_UNROLL_N * 2 * 1 *  sizeof(double)));
    
40.       |                                                  ^~~~~~~~~~~~~~~
    
41.       |                                                  SBGEMM_DEFAULT_Q
    
42. getarch_2nd.c:81:50: error: ‘CGEMM_DEFAULT_Q’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_Q’?
    
43.    81 |     printf("#define CLOCAL_BUFFER_SIZE\t%ld\n", (CGEMM_DEFAULT_Q * CGEMM_DEFAULT_UNROLL_N * 4 * 2 *  sizeof(float)));
    
44.       |                                                  ^~~~~~~~~~~~~~~
    
45.       |                                                  SBGEMM_DEFAULT_Q
    
46. getarch_2nd.c:82:50: error: ‘ZGEMM_DEFAULT_Q’ undeclared (first use in this function); did you mean ‘SBGEMM_DEFAULT_Q’?
    
47.    82 |     printf("#define ZLOCAL_BUFFER_SIZE\t%ld\n", (ZGEMM_DEFAULT_Q * ZGEMM_DEFAULT_UNROLL_N * 2 * 2 *  sizeof(double)));
    
48.       |                                                  ^~~~~~~~~~~~~~~
    
49.       |                                                  SBGEMM_DEFAULT_Q
    
50. make: *** [getarch_2nd] Error 1
    
51. OpenBLAS detected LIBCORE = unknown
    
52. OpenBLAS detected ARCH    = x86_64

复制代码

因此，到Installing OpenBLAS时直接停止：

1. Installing OpenBLAS library for target UNKNOWN
   
2. getarch.c:1698:2: error: #error "The TARGET specified on the command line or in Makefile.rule is not supported. Please choose a target from TargetList.txt"
   
3. 1698 | #error "The TARGET specified on the command line or in Makefile.rule is not supported. Please choose a target from TargetList.txt"
   
4.       |  ^~~~~
   
5. make: *** [getarch] Error 1
   
6. Makefile.system:1531: Makefile.: No such file or directory
   
7. make: *** No rule to make target `Makefile.'.  Stop.
   
8. getarch.c:1698:2: error: #error "The TARGET specified on the command line or in Makefile.rule is not supported. Please choose a target from TargetList.txt"
   
9. 1698 | #error "The TARGET specified on the command line or in Makefile.rule is not supported. Please choose a target from TargetList.txt"
   
10.       |  ^~~~~
    
11. make: *** [getarch] Error 1
    
12. Makefile.system:1531: Makefile.: No such file or directory
    
13. make: *** No rule to make target `Makefile.'.  Stop.

复制代码

请问楼主，这种问题应该如何解决