Skip to content

Latest commit

 

History

History
213 lines (159 loc) · 7.4 KB

USAGE.md

File metadata and controls

213 lines (159 loc) · 7.4 KB

Notes on OpenBLAS usage

Usage

Program is Terminated. Because you tried to allocate too many memory regions

In OpenBLAS, we mange a pool of memory buffers and allocate the number of buffers as the following.

#define NUM_BUFFERS (MAX_CPU_NUMBER * 2)

This error indicates that the program exceeded the number of buffers.

Please build OpenBLAS with larger NUM_THREADS. For example, make NUM_THREADS=32 or make NUM_THREADS=64. In Makefile.system, we will set MAX_CPU_NUMBER=NUM_THREADS.

Despite its name, and due to the use of memory buffers in functions like SGEMM, the setting of NUM_THREADS can be relevant even for a single-threaded build of OpenBLAS, if such functions get called by multiple threads of a program that uses OpenBLAS. In some cases, the affected code may simply crash or throw a segmentation fault without displaying the above warning first.

Note that the number of threads used at runtime can be altered to differ from the value NUM_THREADS was set to at build time. At runtime, the actual number of threads can be set anywhere from 1 to the build's NUM_THREADS (note however, that this does not change the number of memory buffers that will be allocated, which is set at build time). The number of threads for a process can be set by using the mechanisms described below.

How can I use OpenBLAS in multi-threaded applications?

If your application is already multi-threaded, it will conflict with OpenBLAS multi-threading. Thus, you must set OpenBLAS to use single thread in any of the following ways:

  • export OPENBLAS_NUM_THREADS=1 in the environment variables.
  • Call openblas_set_num_threads(1) in the application on runtime.
  • Build OpenBLAS single thread version, e.g. make USE_THREAD=0

If the application is parallelized by OpenMP, please use OpenBLAS built with USE_OPENMP=1

How to choose TARGET manually at runtime when compiled with DYNAMIC_ARCH

The environment variable which control the kernel selection is OPENBLAS_CORETYPE (see driver/others/dynamic.c) e.g. export OPENBLAS_CORETYPE=Haswell and the function char* openblas_get_corename() returns the used target.

How could I disable OpenBLAS threading affinity on runtime?

You can define the OPENBLAS_MAIN_FREE or GOTOBLAS_MAIN_FREE environment variable to disable threading affinity on runtime. For example, before the running,

export OPENBLAS_MAIN_FREE=1

Alternatively, you can disable affinity feature with enabling NO_AFFINITY=1 in Makefile.rule.

Linking with the library

  • Link with shared library

gcc -o test test.c -I /your_path/OpenBLAS/include/ -L/your_path/OpenBLAS/lib -lopenblas

If the library is multithreaded, please add -lpthread. If the library contains LAPACK functions, please add -lgfortran or other Fortran libs.

  • Link with static library

gcc -o test test.c /your/path/libopenblas.a

You can download test.c from https://gist.github.com/xianyi/5780018

On Linux, if OpenBLAS was compiled with threading support (USE_THREAD=1 by default), custom programs statically linked against libopenblas.a should also link with the pthread library e.g.:

gcc -static -I/opt/OpenBLAS/include -L/opt/OpenBLAS/lib -o my_program my_program.c -lopenblas -lpthread

Failing to add the -lpthread flag will cause errors such as:

/opt/OpenBLAS/libopenblas.a(memory.o): In function `_touch_memory':
memory.c:(.text+0x15): undefined reference to `pthread_mutex_lock'
memory.c:(.text+0x41): undefined reference to `pthread_mutex_unlock'
...

Code examples

Call CBLAS interface

This example shows calling cblas_dgemm in C. https://gist.github.com/xianyi/6930656

#include <cblas.h>
#include <stdio.h>

void main()
{
  int i=0;
  double A[6] = {1.0,2.0,1.0,-3.0,4.0,-1.0};
  double B[6] = {1.0,2.0,1.0,-3.0,4.0,-1.0};
  double C[9] = {.5,.5,.5,.5,.5,.5,.5,.5,.5};
  cblas_dgemm(CblasColMajor, CblasNoTrans, CblasTrans,3,3,2,1,A, 3, B, 3,2,C,3);

  for(i=0; i<9; i++)
    printf("%lf ", C[i]);
  printf("\n");
}

gcc -o test_cblas_open test_cblas_dgemm.c -I /your_path/OpenBLAS/include/ -L/your_path/OpenBLAS/lib -lopenblas -lpthread -lgfortran

Call BLAS Fortran interface

This example shows calling dgemm Fortran interface in C. https://gist.github.com/xianyi/5780018

#include "stdio.h"
#include "stdlib.h"
#include "sys/time.h"
#include "time.h"

extern void dgemm_(char*, char*, int*, int*,int*, double*, double*, int*, double*, int*, double*, double*, int*);

int main(int argc, char* argv[])
{
  int i;
  printf("test!\n");
  if(argc<4){
    printf("Input Error\n");
    return 1;
  }

  int m = atoi(argv[1]);
  int n = atoi(argv[2]);
  int k = atoi(argv[3]);
  int sizeofa = m * k;
  int sizeofb = k * n;
  int sizeofc = m * n;
  char ta = 'N';
  char tb = 'N';
  double alpha = 1.2;
  double beta = 0.001;

  struct timeval start,finish;
  double duration;

  double* A = (double*)malloc(sizeof(double) * sizeofa);
  double* B = (double*)malloc(sizeof(double) * sizeofb);
  double* C = (double*)malloc(sizeof(double) * sizeofc);

  srand((unsigned)time(NULL));

  for (i=0; i<sizeofa; i++)
    A[i] = i%3+1;//(rand()%100)/10.0;

  for (i=0; i<sizeofb; i++)
    B[i] = i%3+1;//(rand()%100)/10.0;

  for (i=0; i<sizeofc; i++)
    C[i] = i%3+1;//(rand()%100)/10.0;
  //#if 0
  printf("m=%d,n=%d,k=%d,alpha=%lf,beta=%lf,sizeofc=%d\n",m,n,k,alpha,beta,sizeofc);
  gettimeofday(&start, NULL);
  dgemm_(&ta, &tb, &m, &n, &k, &alpha, A, &m, B, &k, &beta, C, &m);
  gettimeofday(&finish, NULL);

  duration = ((double)(finish.tv_sec-start.tv_sec)*1000000 + (double)(finish.tv_usec-start.tv_usec)) / 1000000;
  double gflops = 2.0 * m *n*k;
  gflops = gflops/duration*1.0e-6;

  FILE *fp;
  fp = fopen("timeDGEMM.txt", "a");
  fprintf(fp, "%dx%dx%d\t%lf s\t%lf MFLOPS\n", m, n, k, duration, gflops);
  fclose(fp);

  free(A);
  free(B);
  free(C);
  return 0;
}

gcc -o time_dgemm time_dgemm.c /your/path/libopenblas.a

./time_dgemm <m> <n> <k>

Troubleshooting

  • Please read Faq at first.
  • Please use gcc version 4.6 and above to compile Sandy Bridge AVX kernels on Linux/MingW/BSD.
  • Please use Clang version 3.1 and above to compile the library on Sandy Bridge microarchitecture. The Clang 3.0 will generate the wrong AVX binary code.
  • The number of CPUs/Cores should less than or equal to 256. On Linux x86_64(amd64), there is experimental support for up to 1024 CPUs/Cores and 128 numa nodes if you build the library with BIGNUMA=1.
  • OpenBLAS does not set processor affinity by default. On Linux, you can enable processor affinity by commenting the line NO_AFFINITY=1 in Makefile.rule. But this may cause the conflict with R parallel.
  • On Loongson 3A. make test would be failed because of pthread_create error. The error code is EAGAIN. However, it will be OK when you run the same testcase on shell.

BLAS reference manual

If you want to understand every BLAS function and definition, please read Intel MKL reference manual or netlib.org

Here are OpenBLAS extension functions

How to reference OpenBLAS.

You can reference our papers.

Alternatively, you can cite the OpenBLAS homepage http://www.openblas.net directly.