Magma Installation RC5 error message

[nabil]

Good Evening everyone,
I am trying to install the new magma RC5 with CUDA 4.0 installed on Fedora 11 linux based system with Tesla GPU. Also, I have installed gotoBlas.

I have the following make.inc:

#//////////////////////////////////////////////////////////////////////////////
# -- MAGMA (version 1.0) --
# Univ. of Tennessee, Knoxville
# Univ. of California, Berkeley
# Univ. of Colorado, Denver
# November 2010
#
# Contributed by: Allan Menezes (Ontario, Canada)
#//////////////////////////////////////////////////////////////////////////////

#
# GPU_TARGET specifies for which GPU you want to compile MAGMA
# 0: Tesla family
# 1: Fermi Family
#
GPU_TARGET = 0

CC = gcc
NVCC = nvcc
FORT = gfortran


ARCH = ar
ARCHFLAGS = cr
RANLIB = ranlib

OPTS = -O3 -DADD_
FOPTS = -O3 -DADD_ -x f95-cpp-input
NVOPTS = --compiler-options -fno-strict-aliasing -DUNIX -O3 -DADD_
LDOPTS = -fPIC -Xlinker -zmuldefs

LIB = -lgoto2_nehalemp-r1.13 -lpthread -lcublas -lcudart -lm -lgfortran

CUDADIR = /usr/local/cuda

LIBDIR = -L/home/bukharin/Nabil/GotoBLAS2 -L$(CUDADIR)/lib64 -L/usr/lib64
INC = -I$(CUDADIR)/include

LIBMAGMA = ../lib/libmagma.a
LIBMAGMABLAS = ../lib/libmagmablas.a

however, I got the following error message:

../lib/libmagmablas.a(zhemv_tesla.cu_o): In function `magmablas_zhemv':
tmpxft_0000564b_00000000-4_zhemv_tesla.compute_13.cudafe1.cpp:(.text+0x116d): undefined reference to `cublasZhemv'
collect2: ld returned 1 exit status
make[1]: *** [testing_zhemv] Error 1
rm testing_zgemm.o testing_zhemv.o
make[1]: Leaving directory `/home/bukharin/Nabil/cuda_magma/magma_1.0.0-rc5/magma_1.0.0-rc5/testing'
make: *** [test] Error 2

any help will be appreciated.

[nabil]

I have found the solution to my problem.
Actually I have found my stupid mistake.
I have two cuda versions 3.2 and 4.0 in my computer and I am setting both of them at the same time in my environment variable.
Thanks everyone...

[jfsebastian]

I'm trying to install Magma 1.0 RC5. I have an Ubuntu 10.10 os, CUDA 4.0 and a GTX260.

My makefile is the following (filename: make.inc)

Code: Select all

	GPU_TARGET = 0

	CC        = gcc
	NVCC      = nvcc
	FORT      = gfortran

	ARCH      = ar
	ARCHFLAGS = cr
	RANLIB    = ranlib

	OPTS      = -O3 -DADD_
	FOPTS     = -O3 -DADD_ -x f95-cpp-input 
	NVOPTS    = -DADD_ --compiler-options -fno-strict-aliasing -DUNIX -O3 
        LDOPTS    = -fPIC

	LIB       = -framework -lcublas -lcudart -lm -lstdc++.6

	CUDADIR   = /usr/local/cuda

	LIBDIR    = -L/usr/local/cuda/lib64 
	INC       = -I../include -I$(CUDADIR)/include

	LIBMAGMA     = ../lib/libmagma.a
	LIBMAGMABLAS = ../lib/libmagmablas.a

The error message is

make[1]: Entering directory `/home/work/Desktop/magma_1.0.0-rc5/testing'
gcc -O3 -DADD_ -DGPUSHMEM=130 -I../include -I/usr/local/cuda/include -I../include -I../quark/include -c testing_zgemm.cpp -o testing_zgemm.o
gcc -O3 -DADD_ -DGPUSHMEM=130 -fPIC -DGPUSHMEM=130 testing_zgemm.o -o testing_zgemm lin/liblapacktest.a -L../lib \
-lcuda -lmagma -lmagmablas -lmagma -L/usr/local/cuda/lib64 -framework -lcublas -lcudart -lm -lstdc++.6
/usr/bin/ld: cannot find -lstdc++.6
collect2: ld returned 1 exit status
make[1]: *** [testing_zgemm] Error 1
rm testing_zgemm.o
make[1]: Leaving directory `/home/work/Desktop/magma_1.0.0-rc5/testing'
make: *** [test] Error 2

I've checked, and the library libstdc++.so.6 is installed and located in /usr/lib and /usr/lib64.

Thank you for any help.