*****************************************************************************
**			TAU Performance System(R)                          **
**			http://tau.uoregon.edu                             **
*****************************************************************************
**    Copyright 1997-2025				   	   	   **
**    Department of Computer and Information Science, University of Oregon **
**    Advanced Computing Laboratory, Los Alamos National Laboratory        **
**    Research Center Juelich, ZAM Germany			           **
*****************************************************************************
/*******************************************************************
 *                                                                 *
 *        Tuning and Analysis Utilities Installation Procedure     *
 *                           Version 2.35.1                        *
 *                                                                 *
 *******************************************************************
 *    For installation help, see INSTALL.                          *
 *    For release notes, see README.                               *
 *    For JAVA instructions, see README.JAVA                       *
 *    For licensing information, see LICENSE.                      *
 *    For a tutorial on using TAU, open html/index.html in your    *
 *        web browser.                                             *
 *    For more information, including updates and new releases,    *
 *        see http://tau.uoregon.edu                               *
 *    For help, reporting bugs, and making suggestions, please     *
 *        send e-mail to tau-bugs@cs.uoregon.edu                   *
 *******************************************************************/
/* NOTE: PLEASE REFER TO tools/src/contrib/LICENSE* files for open *
 * source licenses of other packages that TAU uses internally.     *
 *******************************************************************/


General Installation Procedure: 
-------------------------------
Microsoft Windows users should refer to instructions in README.WINDOWS.txt. 



Quickstart Guide:
=================
Typically, you will need third party libraries to use with TAU. We have created
an external third party library package that contains libunwind, binutils, 
OMPT, Weka, etc. You may download it from:
http://tau.uoregon.edu/ext.tgz
You may also download the Program Database Toolkit (PDT) package from:
http://tau.uoregon.edu/pdt.tgz

PDT is installed using ./configure ; make ; make install 

Please untar the ext.tgz in this top level TAU directory:
tar zxf ext.tgz

If you are using TAU on a Mac OS X system, please download and install
http://tau.uoregon.edu/java.dmg  to better support ParaProf's 3D visualizer. 
If you have a newer Java, you may need to set: 
export PATH=/Library/Java/JavaVirtualMachines/jdk-11.0.3.jdk/Contents/Home/bin:$PATH
in your ~/.zshrc file after installing java. 

After uncompressing and untarring tau, the user needs to configure, compile and
install the package. This is done by invoking:

% ./configure 

We recommend using BFD, libunwind, dwarf, and ompt download options that extract 
the files from the external third party library directory. 

% ./configure -bfd=download -dwarf=download -unwind=download -iowrapper -pdt=<dir> 
% make install

OpenMP Support in TAU
---------------------
TAU supports OMPT v5.0 with Clang 8.x compilers and Intel 19.0.4+ and TR7 with Intel 19.0.1 compilers.
To use this support, simply configure TAU with -ompt. 
NOTE: If these compilers are not used, -ompt automatically downloads the appropriate TR6 or TR4 downloads depending on the compiler. GNU does not currently support OMPT but it will in the future. 

To use MPI compilers and OpenMP, please specify the appropriate compilers as:
% ./configure -c++=mpicxx -cc=mpicc -fortran=mpif90 -mpi -ompt -iowrapper -bfd=download -dwarf=download -otf=download -unwind=download -pdt=<path_to_pdt_dir>
% make install 

To use this new OMPT capability using tau_exec to preload the libTAU.so library from <taudir>/x86_64/lib/shared-icpc-ompt-v5-... directory, please use:

% export TAU_OMPT_SUPPORT_LEVEL=full
% export OMP_NUM_THREADS=4
% mpirun -np 4 tau_exec -T mpi,openmp,ompt,v5,pdt -ompt ./a.out
% pprof -a | more
% paraprof

By default the OMPT support level is basic (which generates events on thread 0, but not on threads 1.. N-1. With full
it generates events on all 0..N-1 threads. 

To use other packages such as PAPI [http://icl.cs.utk.edu/papi] to access 
hardware performance counters or Score-P [http://www.score-p.org] to generate
OTF2 traces that may be viewed in Vampir [http://www.vampir.eu], please use:

% ./configure -papi=<dir> -scorep=<dir> ...
options. Each configuration of TAU builds a unique stub makefile (configuration options file) that stores key configuration parameters (paths to MPI, compilers used, etc). It typically looks like this:
/usr/tau-<version>/<arch>/lib/Makefile.tau-<tags>
We typically use TAU by choosing a specific configuration by setting the 
TAU_MAKEFILE environment variable to point to the desired configuration. 

e.g.,
% export TAU_MAKEFILE=/path/to/tau-2.x/craycnl/lib/Makefile.tau-cray-papi-mpi-pdt
and then after putting TAU in the path:
% export PATH=/path/to/tau-2.x/craycnl/bin:$PATH
we substitute the compiler used with the TAU compiler scripts:

% make CC=tau_cc.sh CXX=tau_cxx.sh F90=tau_f90.sh 
and run the job:
% aprun -n 256 ./a.out

and analyze the performance data:
% paraprof 
(or pprof for text instead of the GUI).
There are some key runtime environment variables that you may set:
% export TAU_METRICS=TIME,PAPI_FP_INS,PAPI_L1_DCM,ENERGY

(ENERGY is only available on newer Cray systems at this time in user space). 
% export TAU_PROFILE_FORMAT="merged"
generates a single tauprofile.xml file instead of profile.<rank>.0.<thread> files.
% paraprof tauprofile.xml & 
allows you to browse the data. For large runs (> 100,000 cores, we recommend using this option). 

Other variables that can be useful are:
% export TAU_SAMPLING=1
which generates a profile using event-based sampling.
For CUDA profiling on GPUs, please use -cuda=<dir> to configure TAU and use tau_exec to launch the job:

% ./configure -cuda=<dir>  ... 

% mpirun -np 256 tau_exec -T papi,cupti,pdt -cupti ./a.out

TAU has been tested with CUDA 10.1, 10.2, 11.0, 12.x, and 13.x and earlier versions. 
No special -cuda options are needed when configuring TAU for OpenACC profiling 
% ./configure -arch=craycnl -mpi -bfd=download -pdt=<dir> -papi=<dir> ; 
% make install
% aprun -n 256 tau_exec -T pdt,mpi -openacc ./a.out

To periodically sample the memory footprint of your application, please set:
% export TAU_TRACK_MEMORY_FOOTPRINT=1

To use TAU with AMD GPUs:
./configure -rocm=<dir> -rocprofsdk  -c++=clang++ -cc=clang -bfd=download; make install
and use
tau_exec -T rocm,serial -rocm ./a.out
TAU has been tested with ROCm 5.x as well as ROCm 6.0.0 with rocprofiler and rocprofsdk 
with ROCm 7.x. 
With ROCm 6.x, the directories have moved and we need to specify the same directory for 
rocm and rocprofiler. For e.g.,:
./configure -rocm=/opt/rocm-6.0.0 -rocprofiler=/opt/rocm-6.0.0 -bfd=download -iowrapper; 
make install
tau_exec -T rocm,serial -rocm -ebs ./a.out

To use TAU with Intel GPUs:
./configure -opencl=<dir> -level_zero=<dir> -bfd=download ; make install
and use:
tau_exec -T level_zero,serial -l0 ./ze_gemm
TAU has been tested with level-zero-devel-1.0.16-i405.el8.x86_64 RPM on an
Intel PVC. It needs libdrm-devel RPM. 

Please see http://tau.uoregon.edu/tau.pptx for sample TAU slides. Please drop us
an e-mail at tau-bugs@cs.uoregon.edu if you get stuck or have a question. 
=================================== END OF QUICKSTART GUIDE ===================

Configuring TAU
----------------

1.  Configure the package for your system. We strongly urge you to see the section
"1) INSTALLING TAU" below for examples (Linux clusters, BGQ, Cray XC40)

TAU is configured by running the configure script with appropriate options that
select the profiling and tracing components that are used to build the TAU 
library.  The `configure' shell script attempts to guess correct values for 
various system-dependent variables used during compilation, and creates the 
Makefile(s) (one in each subdirectory of the source directory).

NOTE: It is highly recommended that you select the *minimal* set of options 
that satisfies the instrumentation and measurement parameters that you need. 
Multiple configurations can be created by using configure several times 
using a different set of options each time. Commonly used configurations are 
typically installed using the 'installtau' tool described below. 

NOTE: tau_setup is a Java based GUI tool for installing TAU on your system.

% ./configure -help 

TAU Configuration Utility 
***********************************************************************
Usage: configure [OPTIONS]
  where [OPTIONS] are:

Compiler Options:
-c++=<compiler>  ............................ specify the C++ compiler.
    options [mpicxx|mpiicpc|mpc_icpc|icpc|g++|*xlC*|cxx|pgCC|pgcpp|
		FCC|guidec++|aCC|c++|ecpc|
               clang++|bgclang++|g++4|icpc|scgcc|scpathCC|pathCC|orCC].
-cc=<compiler> ................................ specify the C compiler.
     options [mpicc|mpiicc|mpc_icc|icc|cc|gcc|clang|bgclang|gcc4|scgcc|
	pgcc|guidec|*xlc*|ecc|pathcc|orcc].
-fortran=<compiler> ..................... specify the Fortran compiler.
   options    [mpif90|mpiifort|gfortran|ibm|pgi|cray|intel|nagware|pathscale
	sgi|ibm64|hp|cray|pgi|absoft|fujitsu|sun|compaq|
       g95|open64|kai|nec|hitachi|intel|absoft|lahey}
-upc=<compiler> ............................. specify the UPC compiler.
     options [upc|gcc(GNU UPC) |upcc (Berkeley UPC) | cc (Cray CCE UPC)
-pdt=<dir> ........ Specify location of PDT (Program Database Toolkit).
-pdt_c++=<compiler>  ............ specify a different PDT C++ compiler.
    options [/full/path/to/compiler|g++|icpc|CC||g++|*xlC*|cxx|pgCC|pgcpp|FCC
	|guidec++|aCC|c++|ecpc|g++4|icpc|scgcc|pathCC|orCC].
-useropt='<param>' .......... arguments to compilers (defaults to -O2).

Installation Options:
-prefix=<dir> ................ Specify a target installation directory.
-exec-prefix=<arch> .......... Specify a target architecture directory.
-arch=<architecture> ................... Specify a target architecture.
       options      [mic_linux|craycnl|bgq|bgp|ibm64linux|sunx86_64
                         crayxmt|solaris2-64|mips32|sgin32|sgi64|sgio32
                                                 arm_linux|arm_android]
-bfd=<dir | download> ....... Specify a binutils directory or download.
    Note: 'download' will download and build the library automatically.
-unwind=<dir | download> ... Specify a libunwind directory or download.
    Note: 'download' will download and build the library automatically.

MPI Options:
-mpi .......................... Specify use of TAU MPI wrapper library.
-mpiinc=<dir> ............. Specify location of MPI include dir and use
                           the TAU MPI Profiling and Tracing Interface.
-mpilib=<dir> ............. Specify location of MPI library dir and use
                           the TAU MPI Profiling and Tracing Interface.
-mpilibrary=<library> ................ Specify a different MPI library.
            e.g., -mpilibrary=-lmpi#-lutil#-ldl  (# is used for a space)                                  
OpenMP Options:
-ompt .................. Use OpenMP Tools API instead of opari
(NOTE: -ompt is currently supported for GNU, Intel, MPC, and OpenUH compilers.)
-openmp ........................................... Use OpenMP threads.
-opari .................................................... Use Opari2.
-opari1 .................................................... Use Opari.
-opari_region ........ Only report performance data for OpenMP regions.
-opari_construct .. Only report performance data for OpenMP constructs.
-oparicomp=<compiler>...Specify which compiler sutie to compile Opari2.
     options [gcc|ibm|intel|pathscale|pgi|studio].

SHMEM Options:
-shmem ...................... Specify use of TAU SHMEM wrapper library.
-shmeminc=<dir> ......... Specify location of SHMEM include dir and use
                         the TAU SHMEM Profiling and Tracing Interface.
-shmemlib=<dir> ......... Specify location of SHMEM library dir and use
                           the TAU MPI Profiling and Tracing Interface.
-shmemlibrary=<library> ............ Specify a different SHMEM library.
            e.g., -shmemlibrary=-lsmac                                 

-gpi=<dir> .................. Specify use of TAU's GPI wrapper library.
GPGPU Options:
-cuda=<dir> ................ Specify location of the top level CUDA SDK
directory with include subdirectory. Enables OpenCL and CUDA profiling.

Other Options:
-iowrapper .................................... Build POSIX IO Wrapper.
-dmapp ...................................... Build Cray DMAPP Wrapper.
-pthread .................................. Use pthread thread package.
-papi=<dir> ............... Specify location of PAPI (Performance API).
-vtf=<dir> ......... Specify location of VTF3 Trace Generation Package.
-otf=<dir> ....... Specify location of Open Trace Format (OTF) Package.
-scorep=<dir>  .................... Specify location of SCOREP package.
-python .. Automatically choose python options based on Python in path.
-pythoninc=<dir> ........ Specify location of Python include directory.
-pythonlib=<dir> ............ Specify location of Python lib directory.
-tag=<unique name> ........ Specify a tag to identify the installation.
-PROFILECOMMUNICATORS ... Generate profiles with MPI communicator info.
-PROFILEPHASE .......................... Generate phase based profiles.
-BGQTIMERS .... Use fast low-overhead timers on IBM BlueGene/Q systems.
-fullhelp .............................. display the full help message.

More advanced options are available, use -fullhelp to see them.
***********************************************************************

The following  command-line options are available to configure:

-prefix=<directory>
   
   Specifies the destination directory where the header, library and binary 
   files are copied. By default, these are copied to subdirectories <arch>/bin 
   and <arch>/lib in the TAU root directory. 
   
-arch=<architecture>
   
   Specifies the architecture. If the user does not specify this option, 
   configure determines the architecture. The files are installed in the 
   <architecture>/bin and <architecture>/lib directories. 

IMPORTANT NOTE: For Cray systems, please specify -arch=craycnl.
   For IBM architectures, we use -arch=ibm64linux and -arch=bgq for IBM 
	Power 8/9 Linux and IBM BG/Q respectively. 
   To build for Intel MIC KNC systems, please use -arch=mic_linux unless 
	the MICs in the Cray XC system, in which case please use -arch=craycnl. 

-mpilibrary=<library> ................ Specify a different MPI library.
            e.g., -mpilibrary=-lmpi#-lutil#-ldl  (# is used for a space)                                  
-cuda=<dir>
		Specifies the locate of the nVidia's CUDA installation. Assumes the header are 
		in <dir>/include and libraries are in /usr/lib or another location accessible 
		by the TAU linker.

-opencl=<dir>
		Specifies the locate of an OPENCL installation. Assumes the header are in
		<dir>/CL/include and libraries are in /usr/lib or another location accessible 
		by the TAU linker.

-pthread
   
   Specifies pthread as the thread package to be used. In the default mode, no 
   thread package is used. 
   
-openmp
   Specifies OpenMP as the threads package to be used. 
   [ Ref: http://www.openmp.org ]

-opari
   Uses Opari
   
-opari_region 
   Report performance data for only OpenMP regions and not constructs. 
   By default, both regions and constructs are profiled with Opari.

-opari_construct 
   Report performance data for only OpenMP constructs and not regions.
   By default, both regions and constructs are profiled with Opari.

-pdt=<directory>
   
   Specifies the location of the installed PDT (Program Database Toolkit) root 
   directory. PDT is used to build tau_instrumentor, a C++, C and F90 
   instrumentation program that automatically inserts TAU annotations in the 
   source code. If PDT is configured with a subdirectory option (-compdir=<opt>)
   then TAU can be configured with the same option by specifying 
   -pdt=<dir> -pdtcompdir=<opt>. 

   [ Ref: http://www.cs.uoregon.edu/research/pdtoolkit ]
   
-papi=<directory>

   Specifies the location of the installed PAPI (Performance API) root 
   directory. PAPI specifies a standard application programming interface (API)    
   for accessing hardware performance counters available on most modern 
   microprocessors similar. For e.g., to measure floating point instructions 
   and level 1 data cache misses and time, please set the environment variable TAU_METRICS:
   % export TAU_METRICS=PAPI_FP_INS,PAPI_L1_DCM,TIME
   % papi_avail  (shows the list of preset metrics)
   % papi_event_chooser PRESET PAPI_FP_INS PAPI_L1_DCM 
     shows you what PAPI preset events are compatible and can be examined 
     in the same experiment with PAPI_FP_INS and PAPI_L1_DCM events. To use
     native events, please use PAPI_NATIVE_<event_name> in the TAU_METRICS. 
     Please refer to the TAU User's Guide or PAPI Documentation for other 
     event names.
   [ Ref : http://icl.cs.utk.edu/projects/papi/api/ ]

-dyninst=<directory>
   Specifies the location of the DynInst (dynamic instrumentation) package. 
   See README.DYNINST for instructions on using TAU with DynInstAPI for 
   binary runtime instrumentation (instead of manual instrumentation) or
   prior to execution by rewriting it. 
   [ Ref: http://www.dyninst.org]

-otf=<directory>
   Specifies the location of the OTF trace generation package. TAU's binary 
   traces can be converted to the Open Trace format (OTF) using tau2otf, a 
   tool that links with the OTF library. OTF traces are hierarchical (multi-stream), 
   compact, support online compression, and can be read concurrently by a parallel trace
   analysis tool such as VNG [ Ref: http://www.vampir.eu, http://www.paratools.com/otf].

-otfinc=<dir>
   Specifies the location of stand-alone OTF header files. This is intended
   for use with features orthogonal to the use of the -otf=<directory> option.
   -otf has the side-effect of making tracing default in the TAU build. This
   may not be the expected behavior for features like Event-based sampling,
   which only needs to know the header and library locations of OTF in order
   to provide conversion facilities for EBS to OTF traces.

-otflib=<dir>
   Specifies the location of stand-alone OTF library files. Please refer to
   -otfinc for further notes on intended use.

-ebs2otf
   Enables the building of utilities that support the conversion of EBS 
   traces to OTF format. This requires an appropriate version of OTF 
   (currently 1.8) to be available either through -otf or through -otfinc
   and -otflib options. Please note that this feature does not require the
   conversion utilities to be built on the same machine the EBS traces are
   generated. System administrators making pre-built versions of TAU for
   their users do not have to enable this feature if the supporting 
   software infrastructure is not available. Please read README.sampling
   for more details.

-slog2=<directory>
   Specifies the location of the SLOG2 SDK trace generation package. TAU's
   binary traces can be converted to the SLOG2 format using tau2slog2, a tool
   that uses the SLOG2 SDK. The SLOG2 format is read by the Jumpshot4 trace
   visualization software, a freely available trace visualizer from Argonne National
   Laboratories.
   [ Ref: http://www-unix.mcs.anl.gov/perfvis/download/index.htm#slog2sdk ]

-pythoninc=<dir>
   Specifies the location of the Python include directory. This is the directory
   where Python.h header file is located. This option enables python bindings to 
   be generated. The user should set the environment variable PYTHONPATH to 
   <TAUROOT>/<ARCH>/lib/bindings-<options> to use a specific version of the TAU 
   Python bindings. By importing package pytau, a user can manually instrument the source
   code and use the TAU API. On the other hand, by importing tau and 
   using tau.run('<func>'), TAU can automatically generate instrumentation. See
   examples/python directory for further information. Please see tau_python.

-pythonlib=<dir>
   
   Specifies the location of the Python lib directory. This is the directory
   where *.py and *.pyc files (and config directory) are located. This option is 
   mandatory for IBM when Python bindings are used. For other systems, this option 
   may not be specified (but -pythoninc=<dir> needs to be specified). 

-PROFILECOMMUNICATORS 
  
   This option generates MPI information partitioned by communicators. TAU 
   lists upto 8 ranks in each communicator in the listing. 

-PROFILEPARAM
 
   This option generates parameter mapped profiles. When used with the MPI 
   wrappper library (-mpi, -mpiinc=<dir>, -mpilib=<dir>) options, TAU generates
   profiles where the time spent in MPI routines is partitioned based on 
   the size of the message. It can also be used in an application to partition
   the time spent in a given routine based on a runtime parameter, such as an
   argument to the routine. See examples/param for further information. 

-PROFILEPHASE
   
   This option generates phase based profiles. It requires special instrumentation
   to mark phases in an application (I/O, computation, etc.). Phases can be 
   static or dynamic (different phases for each loop iteration, for instance).
   See examples/phase/README for further information. 

-useropt=<options-list>
   
   Specifies additional user options such as -g or -I.  For multiple options, 
   the options list should be enclosed in a single quote.

-extrashlibopts=<options-list>
   
   Specifies additional libraries and options that may be passed to the linker 
   while building TAU's shared object. e.g., for AIX -lmpi_r may be passed while
   building libTAU.so with Python and MPI. 
   
-help
   
   Lists all the available configure options and quits. 

-----------------
1) INSTALLING TAU
-----------------

i) To configure TAU for Linux clusters, you may use the MPI compiler wrapper as
the name of the compiler. If this doesn't work, please determine if your MPI 
depends upon some
other package such MPICH over GM. If so, please locate the path to the other library. To 
instrument Fortran/C/C++ code using say, Intel compilers and MPI you may install
PDT [Ref: http://www.cs.uoregon.edu/research/pdt] for automatic source instrumentation, and 
then install TAU. 

  % configure -c++=mpiicpc -cc=mpiicc -fortran=mpiifort -bfd=download -mpi -pdt=<dir> ; make install
  If this doesn't work and say if your MPICH resides in /usr/local/mpich-1.2.7 and it depends upon /opt/gm, you 
   may consider configuring TAU with:
   % configure -pdt=<dir> -c++=icpc -cc=icc -fortran=intel 
       -mpiinc=/usr/local/mpich-1.2.7/include -mpilib=/usr/local/mpich-1.2.7/lib
       -mpilibrary=-lmpich#-L/opt/gm/lib#-lgm#-lpthread#-ldl

	You may use single quotes with spaces or # for spaces in -mpilibrary.
   % make clean install

   For Infiniband, for instance, you may want to use -mpilibrary as below:
   % configure -pdt=<dir> -c++=pathCC -cc=pathcc -fortran=pathscale  
       -mpiinc=/usr/common/usg/mvapich/pathscale/mvapich-0.9.5-mlx1.0.3/include 
       -mpilib=/usr/common/usg/mvapich/pathscale/mvapich-0.9.5-mlx1.0.3/lib 
       -mpilibrary='-lmpich -L/usr/local/ibgd/driver/infinihost/lib64 -lvapi'

   To identify the dependencies of mpich, see mpif90 -v <file.f90> and identify the 
   libraries utilized to link in the application. 

iv) To configure TAU for Cray XK/XC systems or IBM BGQ

   First configure PDT using 
   % configure -XLC ; make clean install on IBM BGQ. 
   On Cray you will need to configure PDT using:
   % configure -GNU; make clean install

   Then, configure TAU using:
   % configure -arch=bgq -mpi -pdt=<dir> -pdt_c++=xlC 
	configures TAU with MPI and PDT on BG/Q. You may configure TAU using:
     Use -arch=craycnl -pdt_c++ on Cray CNL systems and 
	This creates directories
	<taudir>/[bgq,craycnl]/[bin,lib] 
        The directory <taudir>/[bgq,craycnl]/bin should be added to your PATH. 

***********************************************************************
   To install *multiple* (typical) configurations of TAU at a site, you may use the 
   script 'installtau' or 'tau_setup'. Installtau takes options similar to those described above. It 
   invokes ./configure <opts>; make clean install;  to create multiple libraries that 
   may be requested by the users at a site. 
   % installtau -help


TAU Configuration Utility 
***********************************************************************
Usage: installtau [OPTIONS]
  where [OPTIONS] are:
-arch=<arch>  
-fortran=<compiler>  
-cc=<compiler>   
-c++=<compiler>   
-useropt=<options>  
-pdt=<pdtdir>  
-pdtcompdir=<compdir>  
-pdt_c++=<C++ Compiler>  
-papi=<papidir>  
-vtf=<vtfdir>  
-slog2=<dir> (for external slog2 dir)
-slog2 (for using slog2 bundled with TAU)
-dyninst=<dyninstdir>  
-mpiinc=<mpiincdir>  
-mpilib=<mpilibdir>  
-mpilibrary=<mpilibrary>  
-perfinc=<dir> 
-perflib=<dir> 
-perflibrary=<library> 
-mpi
-tag=<unique name> 
-nocomm
-opari=<oparidir>  
-epilog=<epilogdir>  
-prefix=<dir>  
-exec-prefix=<dir>  
***********************************************************************

2. Compilation.

   Type `make clean install' to compile the package. 

   Make installs the library and its stub makefile  in <prefix>/<arch>/lib 
   subdirectory and installs utilities such as pprof and paraprof in 
   <prefix>/<arch>/bin subdirectory.

   
   Add to your .cshrc file the $(TAU_ARCH)/bin subdirectory.
   e.g.,
   # in .cshrc file
   set path=($path /usr/local/packages/tau/x86_64/bin)

   # in .bashrc file
   export PATH=/usr/local/packages/tau/x86_64/bin:$PATH

   See the examples included with this distribution in the examples/ directory.
   The README file in examples directory describes the examples. 
   
   To verify that an installation is correct, please use the tau_validate tool:

   % ./tau_validate -help
    Usage: tau_validate [-v] [--html] [--tag <tag>] [--build] [--run] <target>
    
    Options:
    
    -v           Verbose output
    --html       Output results in HTML
    --tag <tag>  Validate only the subset of TAU stub makefiles matching <tag>
    --build      Only build
    --run        Only run
    <target>     Specify an arch directory (e.g. rs6000), or the lib
                 directory (rs6000/lib), or a specific makefile.
                 Relative or absolute paths are ok.
    
    Notes:
    tau_validate will attempt to validate a TAU installation by performing
    various tests on each TAU stub Makefile.  Some degree of logic exists
    to determine if a given test applies to a given makefile, but it's not
    perfect.
    
    Example:
    
    bash : ./tau_validate --html x86_64 &> results.html
    tcsh : ./tau_validate --html --tag pgi --build craycnl >& results.html

   To upgrade from an older version of TAU, please use the tau_upgrade tool. 
   ./upgradetau  <path/to/old/tau> [extra args]
   e.g.,
   ./upgradetau /usr/local/tau-2.16 -pdt=/usr/local/pdtoolkit-5.6
   Upgrades the current configuration using older tau-2.16 configurations, but
   uses the newer PDT v5.6. 
 

3. Instrumentation.

   TAU provides compilation scripts tau_f90.sh, tau_cc.sh and tau_cxx.sh. You may
   use these scripts to automatically instrument your application if you have 
   specified the use of -pdt=<dir> while configuring TAU. PDT provides source
   code analysis for TAU to automatically insert TAU calls in a copy of the application
   source code. These scripts also link in the TAU libraries. To use this approach,
   simply set the TAU_MAKEFILE environment variable to point to the TAU stub
   makefile that is created in the <arch>/lib directory corresponding to the measurement
   option chosen. For instance, On Cray, when you configure TAU with:
   % configure -pdt=<dir> -mpi -arch=craycnl -bfd=download; make install
   % setenv TAU_MAKEFILE <taudir>/craycnl/lib/Makefile.tau-mpi-pdt
   % tau_f90.sh -c app.f90 ; tau_f90.sh app.o -o app
   % tau_cxx.sh foo.cpp -o foo
   These scripts act similar to the MPI scripts (ftn, mpif90, mpxlf90_r, etc.) 
   that internally invoke the compiler that TAU was configured with. 

   Instrumentation can be controlled by passing options to the TAU compiler. See:
   tau_compiler.sh -help for a complete listing of options and see section 8 below. 

   NEW: If you wish to use compiler-based instrumentation, simply insert
   the -optCompInst option in the TAU_OPTIONS environment variable:
   % setenv TAU_MAKEFILE <taudir>/<arch>/lib/Makefile.tau-mpi-pdt
   % setenv TAU_OPTIONS '-optCompInst -optVerbose' 
   % tau_f90.sh -c app.f90; tau_cc.sh -c foo.c; tau_f90.sh app.o -o app
   This option will use the compiler to insert calls to the TAU measurement
   API. TAU currently support GNU, IBM, Pathscale, Intel and PGI compilers.

   Under Windows, we also support instrumentation using Intel PIN. Use:
   c:\> tau_pin \Path\To\Application.exe
   to spawn the program under PIN. See PIN documentation in TAU for further
   information.

   % cd examples/mm; make
   % mpirun -np 4 ./matmult
   % pprof
   % paraprof

   To illustrate the use of TAU Fortran 90 instrumentation API, we have 
   included the NAS Parallel Benchmarks 2.3 LU and SP suites in the 
   examples/NPB2.3 directory [Ref http://www.nas.nasa.gov/NAS/NPB/ ].
   See the config/make.def makefile that shows how TAU can be used with 
   MPI  (with the TAU MPI Wrapper library) and Fortran 90. To use this, TAU
   must be configured using the -mpiinc=<dir>  and -mpilib=<dir> options. The
   default Fortran 90 compiler used is f90. This may be changed by the user in
   the makefile. LU is completely instrumented and uses the instrumented MPI
   library whereas SP has minimal instrumentation in the top level routine
   and relies on the instrumented MPI wrapper library. 
 
4. Paraprof.

   Paraprof is the GUI for TAU performance analysis. It requires Java 1.5+. An

   IMPORTANT NOTE:
   ***************
   If you see an error that looks like:
   May 18, 2005 2:27:19 PM java.util.prefs.FileSystemPreferences 
   checkLockFile0ErrorCode
   WARNING: Could not lock User prefs. Unix error code 52.

   please make sure  that you've used ssh -Y to login to your remote node. 
   If you see windows that don't look right (in size), this may be the cause
   of the problem as well. You need a trusted ssh connection for Java's Swing. 

5. Performance Database: TAUdb and PerfExplorer

   TAU's Database (TAUdb) database is designed to store and provide 
   access to TAU profile data.  A number of utility programs have been written 
   in Java to load the data into TAUdb (taudb_loadtrial) and to query the data.    With TAUdb users can perform performance analyses such as regression 
   analysis, scalability analysis across multiple trials, and so on.  
   Comparative analyses are available through the TAUdb toolkit.  
   Work is being done to provide the user with standard analysis tools, and 
   an API has been developed to access the data with standard Java classes. 
   For further information, please refer to tools/src/perfdmf/README
   file for installation and usage instructions. 

	PerfExplorer is a framework for parallel performance data mining and
	knowledge discovery. The framework architecture enables the development
	and integration of data mining operations that will be applied to
	large-scale parallel performance profiles. For further information, please 
	refer to tools/src/perfexplorer/doc/README file for installation and usage 
	instructions. 

6. Eclipse Integration: TAU JDT & CDT Plugins for Eclipse
   
   The TAU plugins for Eclipse allow TAU instrumentation and execution of Java, 
   C/C++ and Fortran programs within the Eclipse IDE.  To install the plugin
   for java copy the plugins folder in tools/src/taujava to your Eclipse main
   directory.  To install the plugin for C/C++ and Fortran the Eclipse CDT
   [http://www.eclipse.org/cdt/] or FDT [http://www.eclipse.org/ptp/] plugins
   should be installed as well.  Copy the plugins folder in tools/src/taucdt
   to your Eclipse main directory.  The respective plugins folders contain
   README files with more information.

7. TAU Commander

TAU Commander from ParaTools, Inc. is a production-grade performance engineering solution that makes The TAU Performance System users more productive. It presents a simple, intuitive, and systemized interface that guides users through performance engineering workflows and offers constructive feedback in case of error. TAU Commander also enhances the performance engineer's ability to mine actionable information from the application performance data by connecting to a suite of cloud-based data analysis, storage, visualization, and reporting services.

Download from https://github.com/ParaToolsInc/taucmdr

Please visit: http://www.taucommander.com for further information. 

If you have any questions, please contact us at tau-bugs@cs.uoregon.edu. 


