Go to file
Nathan Binkert 1cd3757db9 add several new functions that can be called from the guest
to tell the simulator to do something.
exit -> exit_old (deprecated
exit now takes an optional parameter that tells it to execute at a
specified time in the future

The next four functions have two optional parameters.  The first
specifies a delay for how long to wait to issue the instruction.
The second will tell the simulator to repeat that command
at the specified interval.

checkpoint will trigger a checkpoint
dumpstats will cause the simulator to dump stats
resetstats will cause all stats to be reset
dumpreset will dump and reset stats


all times are in nanoseconds

util/m5/Makefile:
    Clean up to make it a bit easier to muck with
util/m5/m5.c:
    Add a bunch of new commands and clean up the command parsing path
    Convert atoi to strtoul so that we can use 64bit numbers and even
    hex if we want to.  (this runs on alpha, so a long is 64bit)
util/m5/m5op.h:
    add prototypes for new m5 instructions
    use uint64_t since it's nicer

--HG--
extra : convert_revision : 664ff00f0f0dfc5263c4e873d82fd9996a4521e9
2003-11-02 01:08:59 -05:00
arch Support for Serializable non-SimObject things like events. 2003-10-29 21:45:39 -08:00
base statistics.hh: 2003-11-01 18:17:46 -05:00
cpu Changed the naming of Serializeable derived objects. Serializeable no longer has 2003-10-31 17:32:04 -05:00
dev Fix overwritting and ordering issues with serialization routines 2003-11-01 23:00:40 -05:00
docs footer.html: 2003-10-15 20:49:12 -04:00
kern/tru64 Make FullCPU schedule its TickEvent when one of its contexts becomes active. 2003-10-24 23:02:36 -07:00
sim eventq.cc, eventq.hh: 2003-10-31 22:09:18 -05:00
test Make the to_number function work better. 2003-10-31 18:27:17 -05:00
util add several new functions that can be called from the guest 2003-11-02 01:08:59 -05:00
Doxyfile Add FALRU stats, tweak doxygen configs a bit. 2003-10-17 17:51:08 -04:00
LICENSE Import changeset 2003-10-07 10:41:54 -04:00
README Add a README for the release. Ideally we would move this up a level (to the 2003-10-17 09:32:00 -07:00

This is release m5_1.0_beta1 of the M5 simulator.

This file contains brief "getting started" information and release
notes.  For more information, see http://m5.eecs.umich.edu.  If you
have questions, please send mail to m5sim-users@lists.sourceforge.net.

WHAT'S INCLUDED (AND NOT)
-------------------------

Since you're reading this file, presumably you've managed to untar the
distribution.  The archive you've unpacked has three subdirectories:
 - m5: the simulator itself
 - m5-test: regression tests and scripts to run them
 - ext: less-common external packages needed to build m5
   (currently just "ply")

Although M5 is capable of full-system simulation, the only OS it
currently supports is the proprietary Compaq/HP Tru64 version of Unix.
We are thus unable to distribute bootable disk images freely.  If you
have a Tru64 license and are interested in obtaining disk images,
contact us at m5-dev@eecs.umich.edu.

GETTING STARTED
---------------

The following steps will build and test the simulator.  The variable
"$top" refers to the top directory where you've unpacked the files,
i.e., the one containing the m5, m5-test, and ext directories.

cd $top/m5/setup
./setup ALPHA		# set up build/ALPHA directory
cd $top/m5/build/ALPHA
make m5.opt		# use "-j N" if you've got an MP system
# wait for build...
cd $top/m5-test
./do-tests.pl -B ALPHA	# test what you just built
# wait for tests to run...
# should end with "finished do-tests successfully!"

If you run into errors regarding m5/arch/alpha/decoder.cc, just
"touch" that file to update its timestamp.  This file is generated
from a compact ISA description using a program written in Python.  If
you have Python 2.2.2 or later installed on your system, you should be
able to generate it yourself, but if you don't have Python (or have an
older version), you may run in to trouble.  Since we've shipped a
working copy of decoder.cc, it's not necessary to have Python to build
M5 (unless you start modifying the ISA decription).  Unfortunately,
sometimes make gets confused and tries to do so anyway.  The "touch"
should convince make to stop trying.