Difference between revisions of "Dependencies"
From gem5
Cirosantilli (talk | contribs) |
|||
| (21 intermediate revisions by 7 users not shown) | |||
| Line 3: | Line 3: | ||
==Operating System== | ==Operating System== | ||
| − | gem5 runs best on Linux and Mac OS X. Most developers, and our current regression system, use Linux, so this platform has the best support. A significant number of developers also use Mac OS X, so it is also well supported. | + | gem5 runs best on Linux and Mac OS X. Most developers, and our current regression system, use Linux, so this platform has the best support. A significant number of developers also use Mac OS X, so it is also well supported. For OS X, the easiest way to run gem5 is by using macports and the default system compiler. |
gem5 is designed to be portable to other Unix-like OSes. At times in the past gem5 has worked on OpenBSD and Microsoft Windows (under Cygwin), but these platforms are not regularly tested. Cygwin in particular is no longer actively supported; if you must run on a Windows host, we recommend installing Linux (e.g., Ubuntu Server) under a VM and running gem5 there. Free virtualization solutions such as VirtualBox and VMware Player work well for this usage. | gem5 is designed to be portable to other Unix-like OSes. At times in the past gem5 has worked on OpenBSD and Microsoft Windows (under Cygwin), but these platforms are not regularly tested. Cygwin in particular is no longer actively supported; if you must run on a Windows host, we recommend installing Linux (e.g., Ubuntu Server) under a VM and running gem5 there. Free virtualization solutions such as VirtualBox and VMware Player work well for this usage. | ||
| Line 11: | Line 11: | ||
gem5 is largely agnostic about the hardware it runs on. However, there are several considerations to keep in mind when running gem5: | gem5 is largely agnostic about the hardware it runs on. However, there are several considerations to keep in mind when running gem5: | ||
| − | * A 64-bit platform is strongly preferred over a 32-bit platform. Simulating a platform with a significant amount of physical memory will require the ability to address that much memory from within the gem5 process. Specifically, a 32-bit platform will typically be limited to simulating platforms with roughly 1GB of physical memory. Also, many of the ISAs simulated by gem5 are 64 bit (e.g., x86-64 and Alpha), so simulating their operation on a 32-bit machine will incur additional slowdowns. | + | * A 64-bit platform is strongly preferred over a 32-bit platform. Simulating a platform with a significant amount of physical memory will require the ability to address that much memory from within the gem5 process. Specifically, a 32-bit platform will typically be limited to simulating platforms with roughly 1GB of physical memory. Also, many of the ISAs simulated by gem5 are 64 bit (e.g., x86-64, ARM aarch64 and Alpha), so simulating their operation on a 32-bit machine will incur additional slowdowns. |
| − | * gem5's ISA support involves some very large auto-generated C++ files, which can require up to | + | * gem5's ISA support involves some very large auto-generated C++ files, which can require up to 1GB for g++ to compile. If you intend to do parallel builds (using the scons "-j" flag), you may occasionally see significant slowdowns from paging if your system has less than 1GB per core. This constraint is particularly worth noting if you're configuring a VM to run gem5 under Windows, as suggested above. |
| − | * Ideally you should choose a host with the same endianness as the ISA you will be simulating. gem5 does support cross-endian simulation, but this feature is not extensively tested. Cross-endian simulation works best in syscall emulation (SE) mode | + | * Ideally you should choose a host with the same endianness as the ISA you will be simulating. gem5 does support cross-endian simulation, but this feature is not extensively tested. Cross-endian simulation works best in syscall emulation (SE) mode. |
=External tools and required versions= | =External tools and required versions= | ||
| Line 21: | Line 21: | ||
To build gem5, you will need the following software: | To build gem5, you will need the following software: | ||
| − | * [http://gcc.gnu.org/ g++] version 4.3 or newer. | + | * [http://gcc.gnu.org/ g++] version 4.8 or newer or clang version 3.1 or newer. |
| − | * [http://www.python.org Python], version 2. | + | * [http://www.python.org Python], version 2.6 - 2.7. By July 2019 [https://gem5-review.googlesource.com/c/public/gem5/+/15982/12 Python 3.X support was almost complete]. gem5 links in the Python interpreter, so you need the Python header files and shared library (e.g., /usr/lib/libpython2.6.so) in addition to the interpreter executable. These may or may not be installed by default. For example, on Debian/Ubuntu, you need the "python-dev" package in addition to the "python" package. If you need a newer or different Python installation but can't or don't want to upgrade the default Python on your system, see our page on [[using a non-default Python installation]]. |
| − | * [http://www.scons.org SCons], version 0.98.1 or newer. SCons is a powerful replacement for make. See [http://sourceforge.net/project/showfiles.php?group_id=30337 here] to download SCons. If you don't have administrator privileges on your machine, you can use the "scons-local" package to install scons in your m5 directory, or install SCons in your home directory using the '--prefix=' option. | + | * [http://www.scons.org SCons], version 0.98.1 or newer. SCons is a powerful replacement for make. See [http://sourceforge.net/project/showfiles.php?group_id=30337 here] to download SCons. If you don't have administrator privileges on your machine, you can use the "scons-local" package to install scons in your m5 directory, or install SCons in your home directory using the '--prefix=' option. Some scripts require argparse, which is available by default in Python 2.7 and can be installed from PyPi for older versions. |
| − | |||
* [http://www.zlib.net zlib], any recent version. For Debian/Ubuntu, you will need the "zlib-dev" or "zlib1g-dev" package to get the zlib.h header file as well as the library itself. | * [http://www.zlib.net zlib], any recent version. For Debian/Ubuntu, you will need the "zlib-dev" or "zlib1g-dev" package to get the zlib.h header file as well as the library itself. | ||
* [http://www.gnu.org/software/m4/ m4], the macro processor. | * [http://www.gnu.org/software/m4/ m4], the macro processor. | ||
| + | |||
| + | The following is optional, but highly recommended: | ||
| + | * [https://code.google.com/p/protobuf/ protobuf], version 2.1 or newer for trace capture and playback support. | ||
| + | * [https://pypi.python.org/pypi/pydot pydot], Pythons interface to graphviz, is needed for generation of graphical representations of the simulated system topology. | ||
| + | |||
| + | There a few utility scripts written in Perl, but Perl is not necessary to build or run the simulator. | ||
=Included dependencies= | =Included dependencies= | ||
Some packages which might be difficult to find or which were modified for us in gem5 are included in the ext directory. | Some packages which might be difficult to find or which were modified for us in gem5 are included in the ext directory. | ||
| + | * [http://www.denx.de/wiki/U-Boot/UBootFdtInfo libfdt] -- provides support for flattened device tree "blob" files | ||
* [http://libdnet.sourceforge.net/ dnet] -- dnet provides a simplified, portable interface to several low-level networking routines. | * [http://libdnet.sourceforge.net/ dnet] -- dnet provides a simplified, portable interface to several low-level networking routines. | ||
| − | * [ | + | * [https://github.com/madler/zlib/tree/master/contrib/iostream3 iostream3] -- A C++ stream interface to the zlib library. |
* [http://www.mr511.de/software/english.html libelf] -- ELF object file access library. | * [http://www.mr511.de/software/english.html libelf] -- ELF object file access library. | ||
* [http://www.dabeaz.com/ply/ PLY] -- PLY is an implementation of lex and yacc parsing tools for Python. | * [http://www.dabeaz.com/ply/ PLY] -- PLY is an implementation of lex and yacc parsing tools for Python. | ||
* x11ksyms -- Keycodes from X11 for VNC support. | * x11ksyms -- Keycodes from X11 for VNC support. | ||
| + | * fputils -- Compiler-independent library for 80-bit floating point arithmetic | ||
Latest revision as of 17:10, 4 November 2019
Contents
Supported Platforms
Operating System
gem5 runs best on Linux and Mac OS X. Most developers, and our current regression system, use Linux, so this platform has the best support. A significant number of developers also use Mac OS X, so it is also well supported. For OS X, the easiest way to run gem5 is by using macports and the default system compiler.
gem5 is designed to be portable to other Unix-like OSes. At times in the past gem5 has worked on OpenBSD and Microsoft Windows (under Cygwin), but these platforms are not regularly tested. Cygwin in particular is no longer actively supported; if you must run on a Windows host, we recommend installing Linux (e.g., Ubuntu Server) under a VM and running gem5 there. Free virtualization solutions such as VirtualBox and VMware Player work well for this usage.
Hardware
gem5 is largely agnostic about the hardware it runs on. However, there are several considerations to keep in mind when running gem5:
- A 64-bit platform is strongly preferred over a 32-bit platform. Simulating a platform with a significant amount of physical memory will require the ability to address that much memory from within the gem5 process. Specifically, a 32-bit platform will typically be limited to simulating platforms with roughly 1GB of physical memory. Also, many of the ISAs simulated by gem5 are 64 bit (e.g., x86-64, ARM aarch64 and Alpha), so simulating their operation on a 32-bit machine will incur additional slowdowns.
- gem5's ISA support involves some very large auto-generated C++ files, which can require up to 1GB for g++ to compile. If you intend to do parallel builds (using the scons "-j" flag), you may occasionally see significant slowdowns from paging if your system has less than 1GB per core. This constraint is particularly worth noting if you're configuring a VM to run gem5 under Windows, as suggested above.
- Ideally you should choose a host with the same endianness as the ISA you will be simulating. gem5 does support cross-endian simulation, but this feature is not extensively tested. Cross-endian simulation works best in syscall emulation (SE) mode.
External tools and required versions
To build gem5, you will need the following software:
- g++ version 4.8 or newer or clang version 3.1 or newer.
- Python, version 2.6 - 2.7. By July 2019 Python 3.X support was almost complete. gem5 links in the Python interpreter, so you need the Python header files and shared library (e.g., /usr/lib/libpython2.6.so) in addition to the interpreter executable. These may or may not be installed by default. For example, on Debian/Ubuntu, you need the "python-dev" package in addition to the "python" package. If you need a newer or different Python installation but can't or don't want to upgrade the default Python on your system, see our page on using a non-default Python installation.
- SCons, version 0.98.1 or newer. SCons is a powerful replacement for make. See here to download SCons. If you don't have administrator privileges on your machine, you can use the "scons-local" package to install scons in your m5 directory, or install SCons in your home directory using the '--prefix=' option. Some scripts require argparse, which is available by default in Python 2.7 and can be installed from PyPi for older versions.
- zlib, any recent version. For Debian/Ubuntu, you will need the "zlib-dev" or "zlib1g-dev" package to get the zlib.h header file as well as the library itself.
- m4, the macro processor.
The following is optional, but highly recommended:
- protobuf, version 2.1 or newer for trace capture and playback support.
- pydot, Pythons interface to graphviz, is needed for generation of graphical representations of the simulated system topology.
There a few utility scripts written in Perl, but Perl is not necessary to build or run the simulator.
Included dependencies
Some packages which might be difficult to find or which were modified for us in gem5 are included in the ext directory.
- libfdt -- provides support for flattened device tree "blob" files
- dnet -- dnet provides a simplified, portable interface to several low-level networking routines.
- iostream3 -- A C++ stream interface to the zlib library.
- libelf -- ELF object file access library.
- PLY -- PLY is an implementation of lex and yacc parsing tools for Python.
- x11ksyms -- Keycodes from X11 for VNC support.
- fputils -- Compiler-independent library for 80-bit floating point arithmetic
