1
0
mirror of https://github.com/paboyle/Grid.git synced 2025-04-05 03:35:55 +01:00

README update

This commit is contained in:
Antonin Portelli 2016-10-31 18:21:52 +00:00
parent 66d832c733
commit 07416e4567

112
README.md
View File

@ -16,11 +16,27 @@
**Data parallel C++ mathematical object library.**
Please send all pull requests to the `develop` branch.
License: GPL v2.
Last update 2016/08/03.
Last update Nov 2016.
_Please send all pull requests to the `develop` branch._
### Bug report
_To help us tracking and solving more efficiently issues with Grid, please report problems using the issue system of GitHub rather than sending emails to Grid developers._
When you file an issue, please go though the following checklist:
1. Check that the code is pointing to the `HEAD` of `develop` or any commit in `master` which is tagged with a version number.
2. Give a description of the target platform (CPU, network, compiler).
3. Give the exact `configure` command used.
4. Attach `config.log`.
5. Attach `config.summary`.
6. Attach the output of `make V=1`.
7. Describe the issue and any previous attempt to solve it. If relevant, show how to reproduce the issue using a minimal working example.
### Description
This library provides data parallel C++ container classes with internal memory layout
@ -42,7 +58,7 @@ optimally use MPI, OpenMP and SIMD parallelism under the hood. This is a signifi
for most programmers.
The layout transformations are parametrised by the SIMD vector length. This adapts according to the architecture.
Presently SSE4 (128 bit) AVX, AVX2 (256 bit) and IMCI and AVX512 (512 bit) targets are supported (ARM NEON and BG/Q QPX on the way).
Presently SSE4 (128 bit) AVX, AVX2, QPX (256 bit), IMCI, and AVX512 (512 bit) targets are supported (ARM NEON on the way).
These are presented as `vRealF`, `vRealD`, `vComplexF`, and `vComplexD` internal vector data types. These may be useful in themselves for other programmers.
The corresponding scalar types are named `RealF`, `RealD`, `ComplexF` and `ComplexD`.
@ -50,7 +66,7 @@ The corresponding scalar types are named `RealF`, `RealD`, `ComplexF` and `Compl
MPI, OpenMP, and SIMD parallelism are present in the library.
Please see https://arxiv.org/abs/1512.03487 for more detail.
### Installation
### Quick start
First, start by cloning the repository:
``` bash
@ -71,12 +87,10 @@ mkdir build; cd build
../configure --enable-precision=double --enable-simd=AVX --enable-comms=mpi-auto --prefix=<path>
```
where `--enable-precision=` set the default precision (`single` or `double`),
`--enable-simd=` set the SIMD type (see possible values below), `--enable-
comms=` set the protocol used for communications (`none`, `mpi`, `mpi-auto` or
`shmem`), and `<path>` should be replaced by the prefix path where you want to
install Grid. The `mpi-auto` communication option set `configure` to determine
automatically how to link to MPI. Other options are available, use `configure
where `--enable-precision=` set the default precision,
`--enable-simd=` set the SIMD type, `--enable-
comms=`, and `<path>` should be replaced by the prefix path where you want to
install Grid. Other options are detailed in the next section, you can also use `configure
--help` to display them. Like with any other program using GNU autotool, the
`CXX`, `CXXFLAGS`, `LDFLAGS`, ... environment variables can be modified to
customise the build.
@ -93,24 +107,86 @@ To minimise the build time, only the tests at the root of the `tests` directory
make -C tests/<subdir> tests
```
### Build configuration options
- `--prefix=<path>`: installation prefix for Grid.
- `--with-gmp=<path>`: look for GMP in the UNIX prefix `<path>`
- `--with-mpfr=<path>`: look for MPFR in the UNIX prefix `<path>`
- `--with-fftw=<path>`: look for FFTW in the UNIX prefix `<path>`
- `--enable-lapack[=<path>]`: enable LAPACK support in Lanczos eigensolver. A UNIX prefix containing the library can be specified (optional).
- `--enable-mkl[=<path>]`: use Intel MKL for FFT (and LAPACK if enabled) routines. A UNIX prefix containing the library can be specified (optional).
- `--enable-numa`: ???
- `--enable-simd=<code>`: setup Grid for the SIMD target `<code>` (default: `GEN`). A list of possible SIMD targets is detailed in a section below.
- `--enable-precision={single|double}`: set the default precision (default: `double`).
- `--enable-precision=<comm>`: Use `<comm>` for message passing (default: `none`). A list of possible SIMD targets is detailed in a section below.
- `--enable-rng={ranlux48|mt19937}`: choose the RNG (default: `ranlux48 `).
- `--disable-timers`: disable system dependent high-resolution timers.
- `--enable-chroma`: enable Chroma regression tests.
### Possible communication interfaces
The following options can be use with the `--enable-simd=` option to target different communication interfaces:
| `<comm>` | Description |
| ------------- | -------------------------------------------- |
| `none` | no communications |
| `mpi[-auto]` | MPI communications |
| `mpi3[-auto]` | MPI communications using MPI 3 shared memory |
| `shmem ` | Cray SHMEM communications |
For `mpi` and `mpi3` the optional `-auto` suffix instructs the `configure` scripts to determine all the necessary compilation and linking flags. This is done by extracting the informations from the MPI wrapper specified in the environment variable `MPICXX` (if not specified `configure` will scan though a list of default names).
### Possible SIMD types
The following options can be use with the `--enable-simd=` option to target different SIMD instruction sets:
| String | Description |
| `<code>` | Description |
| ----------- | -------------------------------------- |
| `GEN` | generic portable vector code |
| `SSE4` | SSE 4.2 (128 bit) |
| `AVX` | AVX (256 bit) |
| `AVXFMA4` | AVX (256 bit) + FMA |
| `AVXFMA` | AVX (256 bit) + FMA |
| `AVXFMA4` | AVX (256 bit) + FMA4 |
| `AVX2` | AVX 2 (256 bit) |
| `AVX512` | AVX 512 bit |
| `AVX512MIC` | AVX 512 bit for Intel MIC architecture |
| `ICMI` | Intel ICMI instructions (512 bit) |
| `QPX` | QPX (256 bit) |
Alternatively, some CPU codenames can be directly used:
| String | Description |
| `<code>` | Description |
| ----------- | -------------------------------------- |
| `KNC` | [Intel Knights Corner](http://ark.intel.com/products/codename/57721/Knights-Corner) |
| `KNL` | [Intel Knights Landing](http://ark.intel.com/products/codename/48999/Knights-Landing) |
| `KNC` | [Intel Xeon Phi codename Knights Corner](http://ark.intel.com/products/codename/57721/Knights-Corner) |
| `KNL` | [Intel Xeon Phi codename Knights Landing](http://ark.intel.com/products/codename/48999/Knights-Landing) |
| `BGQ` | Blue Gene/Q |
#### Notes:
- We currently support AVX512 only for the Intel compiler. Support for GCC and clang will appear in future versions.
- For BG/Q only [bgclang](http://trac.alcf.anl.gov/projects/llvm-bgq) is supported. We do not presently plan to support more compilers for this platform.
- BG/Q performances are currently rather poor. This is being investigated for future versions.
### Build setup for Intel Knights Landing platform
The following configuration is recommended for the Intel Knights Landing platform:
``` bash
../configure --enable-precision=double\
--enable-simd=KNL \
--enable-comms=mpi3-auto \
--with-gmp=<path> \
--with-mpfr=<path> \
--enable-mkl \
CXX=icpc MPICXX=mpiicpc
```
where `<path>` is the UNIX prefix where GMP and MPFR are installed. If you are working on a Cray machine that does not use the `mpiicpc` wrapper, please use:
``` bash
../configure --enable-precision=double\
--enable-simd=KNL \
--enable-comms=mpi3 \
--with-gmp=<path> \
--with-mpfr=<path> \
--enable-mkl \
CXX=CC CC=cc
```