參考鏈接:linux下安裝blas和lapack
BLAS 和 LAPACK 這兩個數學庫是很多 Linux 科學計算軟件需要調用的,所以經常會用到。
LAPACK,其名為Linear Algebra PACKage的縮寫,是一以Fortran編程語言編寫,用於數值計算的函式集。LAPACK提供了豐富的工具函式,可用於諸如解多元線性方程式、線性系統方程組的最小平方解、計算特征向量、用於計算矩陣QR分解的Householder轉換、以及奇異值分解等問題。
LAPACK的源碼可以從http://www.netlib.org/lapack/處下載,BLAS也包含在其中。
BLAS,全稱Basic Linear AlgebraSubprograms,即基礎線性代數子程序庫,里面擁有大量已經編寫好的關於線性代數運算的程序。
BLAS的源碼可以從 http://www.netlib.org/blas/ 下載,但實際上LAPACK中已經包含了BLAS。
0. 寫在前面的:
之前采用gfortran來編譯生成了lapack的庫文件,但是在后續采用pgf90命令(pgf90 -llapack)來編譯其它文件時,產生了以下的類似錯誤:
[She@she-centos7 TEC_She]$ make pgf90 -g -fast -c m_bern.f90 pgf90 -g -fast -c d_inpkey.f90 pgf90 -g -fast -c p_menaux.f90 ... pgf90 -g -fast -o main_igsTec -L/usr/local/lib m_bern.o d_inpkey.o ... -llapack -lblas /usr/local/lib/liblapack.a(dormlq.o):在函數‘dormlq_’中: dormlq.f:(.text+0x32b):對‘_gfortran_concat_string’未定義的引用 dormlq.f:(.text+0x887):對‘_gfortran_concat_string’未定義的引用 /usr/local/lib/liblapack.a(dormqr.o):在函數‘dormqr_’中: dormqr.f:(.text+0x2f8):對‘_gfortran_concat_string’未定義的引用 dormqr.f:(.text+0x81c):對‘_gfortran_concat_string’未定義的引用 /usr/local/lib/liblapack.a(ilaenv.o):在函數‘ilaenv_’中: ilaenv.f:(.text+0x58):對‘_gfortran_compare_string’未定義的引用 ilaenv.f:(.text+0x287):對‘_gfortran_compare_string’未定義的引用 ilaenv.f:(.text+0x2b4):對‘_gfortran_compare_string’未定義的引用 ilaenv.f:(.text+0x2d5):對‘_gfortran_compare_string’未定義的引用 ilaenv.f:(.text+0x2f4):對‘_gfortran_compare_string’未定義的引用 /usr/local/lib/liblapack.a(ilaenv.o):ilaenv.f:(.text+0x313): more undefined references to `_gfortran_compare_string' follow /usr/local/lib/liblapack.a(xerbla.o):在函數‘xerbla_’中: xerbla.f:(.text+0x49):對‘_gfortran_st_write’未定義的引用 xerbla.f:(.text+0x54):對‘_gfortran_string_len_trim’未定義的引用 xerbla.f:(.text+0x66):對‘_gfortran_transfer_character_write’未定義的引用 xerbla.f:(.text+0x76):對‘_gfortran_transfer_integer_write’未定義的引用 xerbla.f:(.text+0x7e):對‘_gfortran_st_write_done’未定義的引用 xerbla.f:(.text+0x87):對‘_gfortran_stop_string’未定義的引用 /usr/local/lib/liblapack.a(iparmq.o):在函數‘iparmq_’中: iparmq.f:(.text+0x150):對‘_gfortran_compare_string’未定義的引用 iparmq.f:(.text+0x16f):對‘_gfortran_compare_string’未定義的引用 iparmq.f:(.text+0x18f):對‘_gfortran_compare_string’未定義的引用 iparmq.f:(.text+0x273):對‘_gfortran_compare_string’未定義的引用 iparmq.f:(.text+0x28e):對‘_gfortran_compare_string’未定義的引用 /usr/local/lib/liblapack.a(iparam2stage.o):iparam2stage.F:(.text+0x263): more undefined references to `_gfortran_compare_string' follow make: *** [all] 錯誤 2
這是由於gfortran和pgf90編譯命令不同導致的,因而本文以PGI編譯器來執行lapack源代碼的編譯,即,Fortran程序采用pgf90命令,C程序采用pgcc命令來編譯,相應的具體過程及參數記錄如下。
1. 確保機器上安裝了PGI gfortran編譯器。如果沒有安裝的話,手動安裝:
sudo yum install gfortran
PGI編譯器需要去官網下載,具體安裝過程參見我的另一篇博客《CentOS 7上安裝PGI 2017編譯器》。
2. 下載blas, cblas, lapack 源代碼, 這些源碼都可以在http://www.netlib.org 上找到,下載並解壓。
我下載的版本是lapack-3.7.1.tgz,解壓之后會有一個文件夾,lapack-3.7.1,它含有BLAS,CBLAS,LAPACKE等文件夾,其中BLAS是BLAS的源碼,CBLAS是BLAS的C語言接口。
3. 這里就是具體的編譯步驟
(0) 復制lapack目錄下的make.in.example文件,並修改其中的內容
首先進入lapack-3.7.1文件夾,然后根據平台的特點,將該目錄下對應的 make.inc.example 文件另存為 make.inc。
cd .. cp make.inc.example make.inc vi make.inc #################################################################### # LAPACK make include file. # # LAPACK, Version 3.7.1 # # June 2017 # #################################################################### SHELL = /bin/sh # CC is the C compiler, normally invoked with options CFLAGS. # CC = pgcc # gcc CFLAGS = -O3 # Modify the FORTRAN and OPTS definitions to refer to the compiler # and desired compiler options for your machine. NOOPT refers to # the compiler options desired when NO OPTIMIZATION is selected. # # Note: During a regular execution, LAPACK might create NaN and Inf # and handle these quantities appropriately. As a consequence, one # should not compile LAPACK with flags such as -ffpe-trap=overflow. # FORTRAN = pgf90 # gfortran OPTS = -O2 -Mrecursive # -frecursive DRVOPTS = $(OPTS) NOOPT = -O0 -Mrecursive # -frecursive # Define LOADER and LOADOPTS to refer to the loader and desired # load options for your machine. # LOADER = pgf90 # gfortran LOADOPTS = # The archiver and the flag(s) to use when building an archive # (library). If your system has no ranlib, set RANLIB = echo. # ARCH = ar ARCHFLAGS = cr RANLIB = ranlib # Timer for the SECOND and DSECND routines # # Default: SECOND and DSECND will use a call to the # EXTERNAL FUNCTION ETIME #TIMER = EXT_ETIME # For RS6K: SECOND and DSECND will use a call to the # EXTERNAL FUNCTION ETIME_ #TIMER = EXT_ETIME_ # For gfortran compiler: SECOND and DSECND will use a call to the # INTERNAL FUNCTION ETIME #TIMER = INT_ETIME # If your Fortran compiler does not provide etime (like Nag Fortran # Compiler, etc...) SECOND and DSECND will use a call to the # INTERNAL FUNCTION CPU_TIME TIMER = INT_CPU_TIME # If none of these work, you can use the NONE value. # In that case, SECOND and DSECND will always return 0. #TIMER = NONE # Uncomment the following line to include deprecated routines in # the LAPACK library. # #BUILD_DEPRECATED = Yes # LAPACKE has the interface to some routines from tmglib. # If LAPACKE_WITH_TMG is defined, add those routines to LAPACKE. # #LAPACKE_WITH_TMG = Yes # Location of the extended-precision BLAS (XBLAS) Fortran library # used for building and testing extended-precision routines. The # relevant routines will be compiled and XBLAS will be linked only # if USEXBLAS is defined. # #USEXBLAS = Yes #XBLASLIB = -lxblas # The location of the libraries to which you will link. (The # machine-specific, optimized BLAS library should be used whenever # possible.) # BLASLIB = ../../librefblas.a CBLASLIB = ../../libcblas.a LAPACKLIB = liblapack.a TMGLIB = libtmglib.a LAPACKELIB = liblapacke.a
(1) 編譯blas
進入 BLAS/SRC 文件夾,執行以下幾條命令
cd BLAS/SRC
# gfortran -c -O3 *.f # 編譯所有的 .f 文件,生成 .o文件,這里采用PGI編譯器的pgf90命令來編譯 pgf90 -c -O3 *.f # 編譯所有的 .f 文件,生成 .o文件,這個pgf90編譯命令與~/lapack*/make.inc保持一致 ar rv libblas.a *.o # 鏈接所有的 .o文件,生成.a 文件 sudo cp libblas.a /usr/local/lib #將庫文件復制到系統庫目錄
sudo cp libblas.a /usr/lib
(2) 編譯cblas
進入CBLAS 文件夾,首先根據你自己的計算機平台,將目錄下某個 Makefile.XXX復制為 Makefile.in , XXX表示計算機的平台,如果是linux,那么就將Makefile.LINUX 復制為Makefile.in,然后執行以下命令
cd .. && cd ../CBLAS
cp ../BLAS/SRC/libblas.a ./testing/ # 將上一步編譯成功的 libblas.a復制到 CBLAS目錄下的testing子目錄 make # 編譯所有的目錄 sudo cp ../libcblas.a /usr/local/lib #將庫文件復制到系統庫目錄下
sudo cp ../libcblas.a /usr/lib
(3) 編譯 lapack 以及 lapacke
這一步比較麻煩,首先進入lapack-3.7.1文件夾,根據平台的特點,編輯 Makefile,編譯 lapack 和 lapacke 文件,並將 lapacke 目錄下的頭文件、lapack 目錄下生成的 *.a 文件拷貝到系統目錄(/usr/local/lib, /usr/lib)下。
cd ..
vi Makefile # 修改 lapack-3.7.1/Makefile 文件,因為 lapack 依賴於 blas 庫 # 舊版本
lib: lapacklib tmglib #lib: blaslib variants lapacklib tmglib # 新版本
#lib: lapacklib tmglib lib: blaslib variants lapacklib tmglib
make # 編譯所有的lapack文件 cd LAPACKE # 進入LAPACKE 文件夾,這個文件夾包含lapack的C語言接口文件
make # 編譯lapacke
sudo cp include/*.h /usr/local/include #將lapacke的頭文件復制到系統頭文件目錄,
# 包括: lapacke.h, lapacke_config.h, lapacke_mangling.h,lapacke_mangling_with_flags.h lapacke_utils.h
cd .. # 返回到 lapack-3.7.1 目錄
sudo cp *.a /usr/local/lib # 將生成的所有庫文件復制到系統庫目錄,
# 包括:liblapack.a, liblapacke.a, librefblas.a,libtmglib.a。
sudo cp *.a /usr/lib
Ques: 事實上,編譯 lapack 時生成的 librefblas.a 文件與編譯 BLAS 時生成的 libblas.a 文件大小基本一樣,這里生成了兩次,是否可以省去第(1)-(2)步?
至此blas,cblas 和 lapack 就成功安裝到你的電腦上了。
4. lapack子程序測試
測試程序 Console.f
! 測試程序來自:http://blog.sina.com.cn/s/blog_5f350c9601014ejc.html program Console1 external dgesv integer n, lda, nrhs, ldb parameter (n=2,lda=2,nrhs=1,ldb=2) double precision A(lda,n) double precision b(ldb,nrhs) character byebye integer ipiv(n), info, i, j A(1,1)=1 A(1,2)=2 A(2,1)=3 A(2,2)=4 B(1,1)=5 B(2,1)=6 write(*,*) 'Hello World' call dgesv(n,nrhs,A,lda,ipiv,b,ldb,info) write(*,*) 'INFO =', info write(*,*) ((A(i,j),i=1,lda),j=1,n) write(*,*) ((B(i,j),i=1,ldb),j=1,nrhs) write(*,*) "END OF PROGRAM..." end program Console1
使用pgf90來編譯該程序,可以得到預期的結果,而使用 gfortran 來編譯則會報錯:
[She@she-centos7 LSQtest]$ pgf90 Console1.f -lblas -llapack # 編譯及運行正常 [She@she-centos7 LSQtest]$ ./a.out Hello World INFO = 0 3.000000000000000 0.3333333333333333 4.000000000000000 0.6666666666666667 -3.999999999999999 4.499999999999999 END OF PROGRAM... [She@she-centos7 LSQtest]$ [She@she-centos7 LSQtest]$ [She@she-centos7 LSQtest]$ [She@she-centos7 LSQtest]$ gfortran Console1.f -lblas -llapack # 編譯報錯 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgesv.o):在函數‘.C1_322’中: dgesv.f:(.data+0x18):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrf.o):在函數‘dgetrf_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf.f:206:對‘dtrsm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf.f:213:對‘dgemm_’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrf.o):在函數‘.C1_331’中: dgetrf.f:(.data+0x38):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrs.o):在函數‘dgetrs_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrs.f:191:對‘dtrsm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrs.f:191:對‘dtrsm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrs.f:202:對‘dtrsm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrs.f:202:對‘dtrsm_’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrs.o):在函數‘.C1_292’中: dgetrs.f:(.data+0x48):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dlaswp.o):(.data+0x0):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(ilaenv.o):在函數‘ilaenv_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:703:對‘pgf90_str_copy’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:261:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:274:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:687:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:353:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:479:對‘pgf90_strcmp’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(ilaenv.o):/home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./ilaenv.f:485: more undefined references to `pgf90_strcmp' follow /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(ilaenv.o):在函數‘.STATICS1’中: ilaenv.f:(.data+0x128):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(ieeeck.o):在函數‘.C1_352’中: ieeeck.f:(.data+0x10):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(xerbla.o):在函數‘xerbla_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90io_src_info03’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90io_fmtw_init’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90_lentrim’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90io_fmt_write’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90io_sc_i_fmt_write’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90io_fmtw_end’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./xerbla.f:90:對‘pgf90_stop08’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(xerbla.o):在函數‘.STATICS1’中: xerbla.f:(.data+0xa0):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(iparmq.o):在函數‘iparmq_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:265:對‘__gss_log’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:265:對‘__gss_log’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:265:對‘__mth_i_nint’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:321:對‘pgf90_str_copy’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:336:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:369:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:387:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:379:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparmq.f:379:對‘pgf90_strcmp’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(iparmq.o):在函數‘.C1_289’中: iparmq.f:(.data+0x20):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(iparam2stage.o):在函數‘iparam2stage_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:207:對‘pgf90_str_copy’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:341:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:350:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:359:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:360:對‘pgf90_strcmp’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:378:對‘pgf90_strcmp’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(iparam2stage.o):/home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./iparam2stage.F:378: more undefined references to `pgf90_strcmp' follow /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(iparam2stage.o):在函數‘.C1_327’中: iparam2stage.F:(.data+0x48):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(lsame.o):(.data+0x0):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrf2.o):在函數‘dgetrf2_’中: /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf2.f:242:對‘dtrsm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf2.f:247:對‘dgemm_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf2.f:192:對‘idamax_’未定義的引用 /home/She/Software/Fortran/Lapack/lapack-3.7.1/SRC/./dgetrf2.f:207:對‘dscal_’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dgetrf2.o):在函數‘.C1_292’中: dgetrf2.f:(.data+0x40):對‘f90_compiled’未定義的引用 /usr/lib/gcc/x86_64-redhat-linux/4.8.5/../../../liblapack.a(dlamch.o):在函數‘.C1_291’中: dlamch.f:(.data+0x68):對‘f90_compiled’未定義的引用 collect2: 錯誤:ld 返回 1
測試完畢。
5. 結論和心得
在同一台電腦上,最好對Fortran程序和C程序使用一致的編譯命令,庫文件和源代碼都遵循這樣的做法,可以避免不必要的奇怪bug。
如果一組程序中,某些文件采用了gfortran 來編譯,而一些文件采用了 pgf90 命令來編譯,則鏈接時容易產生一些難以檢查的錯誤,浪費生命!