Update doc

Signed-off-by: Chen Wang <[email protected]>

docs/source/index.rst

@@ -1,80 +1,59 @@
-|build|
-
-Recorder
+Overview
 ========
 
-**A Multi-Level Library for Understanding I/O Activity in HPC
-Applications**
-
-Quickstart
-----------
-
-**1. Build**
-
-Recorder requires MPI and HDF5 to build. Please make sure they are installed
-before building Recorder.
-
-Run the following to fetch, build, and install Recorder.
-In this example, we will install Recorder under its srouce directory.
-You can change the install localtion by modifying ``$RECORDER_INSTALL_PATH`` below. 
-
-.. code:: bash
-
-   git clone https://github.com/uiuc-hpc/Recorder.git
-   cd Recorder
-   export RECORDER_INSTALL_PATH=`pwd`/install
-   git submodule update --init --recursive
-   mkdir build
-   cd build
-   cmake .. -DCMAKE_INSTALL_PREFIX=$RECORDER_INSTALL_PATH
-   make
-   make install
+We believe that multi-level I/O tracing and trace data analysis tool can
+help end users understand the behavior of their application and I/O
+subsystem, and can provide insights into the source of I/O performance
+bottlenecks.
 
+Recorder is a multi-level I/O tracing framework that can capture I/O
+function calls at multiple levels of the I/O stack, including HDF5,
+MPI-IO, and POSIX I/O. Recorder requires no modification or
+recompilation of the application and users can control what levels are
+traced.
 
-**2. Trace an application**
+Publications
+------------
 
-Make sure ``$RECORDER_INSTALL_PATH`` points to the location where you installed Recorder.
-Now let's run an application with Recorder to collect it traces.
-We will use an example MPI program as our targeting application.
-Go to the Recorder's source directory and run the following.
+Recorder has undergone significant changes since the last
+"Recorder 2.0" paper. We have incorporated a new pattern-based
+compression algorithm, along with many new features.
+We are preparing a new paper that will describe all these
+changes in detail.
 
-.. code:: bash
 
-    cd test && mpicc test_mpi.c -o test_mpi
-    mpirun -np N -env LD_PRELOAD $RECORDER_INSTALL_PATH/lib/librecorder.so ./test_mpi
+`Wang, Chen, Jinghan Sun, Marc Snir, Kathryn Mohror, and Elsa
+Gonsiorowski. “Recorder 2.0: Efficient Parallel I/O Tracing and
+Analysis.” In IEEE International Workshop on High-Performance Storage
+(HPS), 2020. <https://doi.org/10.1109/IPDPSW50202.2020.00176>`__
 
-    # On HPC systems, you may need to use srun or 
-    # other job schedulers to replace mpirun, e.g.,
-    srun -n4 -N1 --overlap --export=ALL,LD_PRELOAD=$RECORDER_INSTALL_PATH/lib/librecorder.so ./test_mpi
 
-Once completed, you will see a folder named ``recorder-yyyymmdd``. This is a parent folder
-that contains all traces folders generated on *yyyymmdd*.
-Inside it, you will find the actual trace forlder of the application you just run.
-The folder name follows the format of ``HHmmSS.ff-hostname-username-appname-pid``.
+I/O Patterns and consistency requirements
+-----------------------------------------
 
-**3. Examine the traces**
+`Traces from 17 HPC applications <https://doi.org/10.6075/J0Z899X4>`__
 
-Recorder provides several useful tools under ``$RECORDER_INSTALL_PATH$/bin`` for analysis purposes.
+The traces were collected using an old version of Recorder. The current
+version uses a different trace format. To read those traces please use
+Recorder 2.2.1 from the
+`release <https://github.com/uiuc-hpc/Recorder/releases/tag/v2.2.1>`__
+page.
 
-*recorder-summary* can be used to reports high-level statistics:
-
-.. code:: bash
-
-   $RECORDER_INSTALL_PATH/bin/recorder-summary /path/to/your_trace_folder/
-
-*recorder2text* is used to convert the Recorder-format traces to plain text files.
-
-.. code:: bash
-
-   $RECORDER_INSTALL_PATH/bin/recorder2text /path/to/your_trace_folder/
-
-This will generate text fomart traces under ``/path/to/your_trace_folder/_text``.
-You will see *N* [pid].txt files, where *N* is the number processors you run your application.
-Each of these txt files contains the traces generated by one processor indicated by the [pid].
+The following paper presents a comprehensive study on the I/O patterns
+of these applications. Additionally, it investigates the consistency
+requirements of these applications, confirming experimentally that
+POSIX consistency is rarely required by HPC applications.
 
+`Wang, Chen, Kathryn Mohror, and Marc Snir. “File System Semantics
+Requirements of HPC Applications.” Proceedings of the 30th International
+Symposium on High-Performance Parallel and Distributed Computing (HPDC).
+2021. <https://dl.acm.org/doi/abs/10.1145/3431379.3460637>`__
 
 .. toctree::
-   overview
+   :maxdepth: 1
+   :hidden
+
+   quickstart
    build
    usage
    features