search

IntelLabs / HPAT.jl

High Performance Analytics Toolkit (HPAT) is a Julia-based framework for big data analytics on clusters.
BSD 2-Clause "Simplified" License
120 stars 16 forks source link

Error trying to execute 1D Sum example #16

Closed rafaelcarv closed 7 years ago

rafaelcarv commented 7 years ago

Hello, I am trying to execute the 1D Sum example, I updated all my environment, and i am using Julia 0.5

Did the checkout from all the intel packages (ParalellAccelerator, CompilerTools and HPAT). However, i am not able to execute the example.

I am using Ubuntu 16.04.

I am attaching a file with the error output.

out.txt

ehsantn commented 7 years ago

Seems like your HDF5 is not installed properly. You could try (Ubuntu):

sudo apt-get install -qq libmpich2-dev mpich2 hdf5-tools libhdf5-mpich2-dev

or build from source: https://support.hdfgroup.org/ftp/HDF5/current/src/unpacked/release_docs/INSTALL_parallel You might want to add the include directory to the system environment:

export C_INCLUDE_PATH=$C_INCLUDE_PATH:/usr/local/hdf5/include
export CPLUS_INCLUDE_PATH=$CPLUS_INCLUDE_PATH:/usr/local/hdf5/include
Wajihulhassan commented 7 years ago

@rafaelcarv For Ubuntu, in addition to steps mentioned on HPAT's README, run commands from this wiki: https://github.com/IntelLabs/HPAT.jl/wiki/HPAT-Installation-on-Ubuntu-16.04

rafaelcarv commented 7 years ago

Hello,

After following these instructions (Ehsan instruction, plus the wiki instruction), the HDF5 error is not occurring anymore, however I am getting the following error:

file= /tmp/1D_large.hdf5 OptFramework failed to optimize function ##calc1Dsum#271 in optimization pass HPAT.runDistributedPass with error UndefVarError(:computeDependenciesAST) ERROR: LoadError: error compiling ##calc1Dsum#271: unsupported or misplaced expression "alloc" in function ##calc1Dsum#271 in calc1Dsum(::String) at /home/rafael/.julia/v0.5/CompilerTools/src/OptFramework.jl:577 in main() at /home/rafael/.julia/v0.5/HPAT/examples/1D_sum.jl:62 in include_from_node1(::String) at ./loading.jl:488 in process_options(::Base.JLOptions) at ./client.jl:262 in _start() at ./client.jl:318 while loading /home/rafael/.julia/v0.5/HPAT/examples/1D_sum.jl, in expression starting on line 75

Wajihulhassan commented 7 years ago

@rafaelcarv . I fired up a new VM of Ubuntu 16.04. I built and installed release-0.5 Julia from the github and followed the instructions on README and wiki. And successfully ran HPAT tests. I did not get your error.

wajih@wajih:~/.julia/v0.5/HPAT/examples (master)*$ julia --depwarn=no 1D_sum.jl 
WARNING: Method definition (::Type{ParallelAccelerator.ParallelIR.PIRParForAst})(Any, Any, Any, Any, Any, Any, Any, Any, Any, Any, Any) in module ParallelIR at /home/wajih/.julia/v0.5/ParallelAccelerator/src/parallel-ir.jl:359 overwritten at /home/wajih/.julia/v0.5/ParallelAccelerator/src/parallel-ir.jl:363.
file= /home/wajih/.julia/v0.5/HPAT/src/../input_data/1D_large.hdf5
Distributed-memory MPI mode.
OpenMP is not used.
Main will be generated in file main1.cc
Data for main will be in file main1.data
Script to compile is in main1.sh
SELFPRIMED 35.689845293
Main will be generated in file main2.cc
Data for main will be in file main2.data
Script to compile is in main2.sh
result = 499762.93314082833
SELFTIMED 0.003996779
wajih@wajih:~/.julia/v0.5/HPAT/examples (master)*$

Just make sure all the packages HPAT, ParallelAccelerator, and CompilerTools are checkout to master branch. Then generate dataset with the script in the HPAT repo generate_data/generate_1D_array.jl and run the example. Try to follow the steps again and let me know if you see the error again.

rafaelcarv commented 7 years ago

Thanks,

I just reinstalled everything in a new VM and it worked.