updated documentation

README.md

@@ -9,11 +9,13 @@ integrates effortlessly into existing application codes.  ReLAPACK's routines
 not only outperform the reference LAPACK but also improve upon the performance
 of tuned implementations, such as OpenBLAS and MKL.
 
+
 Coverage
 --------
 For a detailed list of covered operations and an overview of operations to which
 recursion is not efficiently applicable, see [coverage.md](coverage.md).
 
+
 Installation
 ------------
 To compile with the default configuration, simply run `make` to create the
@@ -26,17 +28,20 @@ and machine specific implementations of these libraries, which are commonly
 provided by hardware vendors or available as open source (e.g.,
 [OpenBLAS](http://www.openblas.net/)).
 
+
 Configuration
 -------------
 For an overview of the configuration options see [config.md](config.md).
 
+
 Testing
 -------
 ReLAPACK's test suite compares its routines numerically with LAPACK's
 counterparts.  To set up the test located int `test/` you need to specify link
 flags for BLAS and LAPACK (version 3.5.0 or newer) in `make.inc`; then `make
 test` runs the tests.  For details, see [test/README.md](test/README.md).
 
+
 Examples
 --------
 Since ReLAPACK replaces parts of LAPACK, any LAPACK example involving the

config.md

@@ -1,13 +1,16 @@
 RELAPACK Configuration
 ======================
+
 ReLAPACK has two configuration files: `make.inc`, which is included by the
 Makefile, and `config.h` which is included in the source files.
 
+
 Build and Testing Environment
 -----------------------------
 The build environment (compilers and flags) and the test configuration (linker
 flags for BLAS and LAPACK, and the test matrix size) are specified in `make.inc`
 
+
 Routine Selection
 -----------------
 By default, ReLAPACK provides each of its LAPACK-based recursive algorithms
@@ -22,13 +25,15 @@ LAPACK routines. By setting `RELAPACK_AS_LAPACK` to 0 in `config.h`, they will
 receive the prefix `RELAPACK_`; e.g. the LU decomposition `dgetrf` would become
 `RELAPACK_dgetrf`.
 
+
 Crossover Size
 --------------
 The crossover size determines below which matrix sizes ReLAPACK's recursive
 algorithms switch to LAPACK's unblocked routines to avoid tiny BLAS Level 3
 routines.  The crossover size is set in `config.h` and can be chosen either
 globally for the entire library, by operation, or individually by routine.
 
+
 Allowing Temporary Buffers
 --------------------------
 Two of ReLAPACK's routines make use of temporary buffers, which are allocated
@@ -50,10 +55,10 @@ affected routines are:
    redundant computations.  It thereby performs about 30% less FLOPs than
    LAPACK.
 
+
 FORTRAN symbol names
 --------------------
-ReLAPACK not only uses FORTRAN routines internally (small modifications of
-original LAPACK routines to enable recursion) but is also commonly linked to
-BLAS and LAPACK with standard FORTRAN interfaces.  Since FORTRAN compilers
-commonly append an underscore to their symbol names, ReLAPACK has configuration
-switches in `config.h` to adjust the library's symbol names.
+ReLAPACK is commonly linked to
+BLAS and LAPACK with standard FORTRAN interfaces.  Since these libraries usually
+have an underscore to their symbol names, ReLAPACK has configuration switches in
+`config.h` to adjust the corresponding routine names.

coverage.md

@@ -1,5 +1,6 @@
 Coverage of ReLAPACK
 ====================
+
 This file lists all LAPACK compute routines that are covered by recursive
 algorithms in ReLAPACK, it also lists all of LAPACK's blocked algorithms which
 are not (yet) part of ReLAPACK.

examples/README.md

@@ -1,5 +1,6 @@
 ReLAPACK Examples
 =================
+
 Since ReLAPACK replaces parts of LAPACK, any LAPACK example involving the
 covered routines applies directly to ReLAPACK.