Hi Guys 

 

System spec I am trying to compile on is. 

 

Fedora 10 64 bit

Intel Q9550

8GB RAM 

Nvidia GTX 275X 633M 896MB DDR3 

 

I am having a complete nightmare trying to get the SVN code compiled.
Below is the file is the files I have:

 

Algorithm                                           device
logger                                   profconf
stxxl-1.2.1           utility

calculate_chain.cpp                        doc
Makefile                              reference
stxxl.errlog         work_consumer

calculate_chain_cuda.cu               fpga
Makefile.local                    round_function
stxxl.hpp             work_consumers.cpp

condition                                             highlevel
Makefile.local.dist           scripts
stxxl.log               work_generator

configuration                                     Implementation
misc                                       speedcomp
stxxl-svn              work_generators.cpp

configuration.hpp                            implementation.hpp
nvopencc-2.2.patch        static_configuration_host.hpp
table                      work_intermediate

cpp
implementation_methods.hpp  obj
static_configuration.hpp              test
work_management

data                                                        includes.hpp
params                                 static_sequence
tests

data_types                                          INSTALL
pretty.pl                              storage
tmto

 

I have modified the Maklefile.local with what I think is the correct
paths as follows:

 

# SDK is not needed anymore

 

HOST_CFLAGS=-malign-double

 

ifeq ($(CROSS), 32bit)

CUDA_DIR=/usr/local/cuda

#SDK_DIR=/home/sascha/NVSDK-32/C

STXXL_DIR=/usr/include/stxxl

BOOST_DIR=/usr/include/boost

BOOST_INFIX=-gcc43

M32=-m32

else

CUDA_DIR=/usr/include/cuda

#SDK_DIR=/home/sascha/NVIDIA_CUDA_SDK

STXXL_DIR=/usr/include/stxxl

#BOOST_INFIX=-gcc43

Endif

 

I got these paths from doing whereis but I am unsure if this is correct.

 

I have followed the guide to compile from
http://reflextor.com/trac/a51/wiki/CompilingTheProgram but when i then
change to /obj and do make c I get the errors

 

[r...@signius-02 obj]# make c

(echo "#!/bin/sh"; echo "exec g++-4.1  \"\...@\"") > gcc

chmod 755 gcc

PATH=.:$PATH /usr/include/cuda/bin/nvcc -c -keep -O3 -g -D_DEBUG
-DSTXXL_BOOST_THREADS -DSTXXL_BOOST_CONFIG -DSTXXL_BOOST_FILESYSTEM
-DSTXXL_BOOST_RANDOM -DSTXXL_BOOST_TIMESTAMP -I..
-I/usr/include/cuda/include -I/usr/include/stxxl/include  -o
calculate_chain_cuda.o ../calculate_chain_cuda.cu

/bin/sh: /usr/include/cuda/bin/nvcc: No such file or directory

make: *** [calculate_chain_cuda.o] Error 127

 

Which is making me think the CUDA stuff isnt installed correctly yet I
have compiled and installed PYRIT from http://code.google.com/p/pyrit/
and when I run the benchmark tests it sees the CUDA Device correctly so
the CUDA stuff must be installed correctly.

 

[r...@signius-02 obj]# /usr/bin/pyrit benchmark

Pyrit 0.2.3 (C) 2008, 2009 Lukas Lueg http://pyrit.googlecode.com

This code is distributed under the GNU General Public License v3

 

The ESSID-blobspace seems to be empty; you should create an ESSID...

 

Running benchmark for at least 60 seconds...

 

#1: 'CUDA-Device #1 'GeForce GTX 275'': 12815.2 PMKs/s (Occ. 99.6%; RTT
2.7)

#2: 'CPU-Core (SSE2)': 553.5 PMKs/s (Occ. 98.4%; RTT 2.8)

#3: 'CPU-Core (SSE2)': 544.9 PMKs/s (Occ. 98.9%; RTT 3.1)

#4: 'CPU-Core (SSE2)': 566.8 PMKs/s (Occ. 99.7%; RTT 3.0)

 

Benchmark done. 14480 PMKs/s total.

 

Please can someone tell me what dumb thing I am doing or not doing this
is driving me nuts now. 

 

Thank you in advance. I am itching to start generating some tables and
contribute. 

 

Any information you require to help me I am more than happy to give.

 

Thanks

 

 

 

 

 

 

 

 

 

 

 

 

 

 

_______________________________________________
A51 mailing list
[email protected]
http://lists.lists.reflextor.com/cgi-bin/mailman/listinfo/a51

Reply via email to