summaryrefslogtreecommitdiff
path: root/sim/ppc/README
diff options
context:
space:
mode:
authorStan Shebs <shebs@apple.com>1999-04-16 01:35:26 +0000
committerStan Shebs <shebs@apple.com>1999-04-16 01:35:26 +0000
commit14cd51f7793a9ce07bc435069f57269450141363 (patch)
tree280a2da48f771d61be5b451ddbacdf9ef8e9ad13 /sim/ppc/README
downloadgdb-14cd51f7793a9ce07bc435069f57269450141363.tar.gz
Initial revision
Diffstat (limited to 'sim/ppc/README')
-rw-r--r--sim/ppc/README353
1 files changed, 353 insertions, 0 deletions
diff --git a/sim/ppc/README b/sim/ppc/README
new file mode 100644
index 00000000000..ff3fbd2b349
--- /dev/null
+++ b/sim/ppc/README
@@ -0,0 +1,353 @@
+
+
+ PSIM 1.0.1 - Model of the PowerPC Environments
+
+
+ Copyright (C) 1994-1996, Andrew Cagney <cagney@highland.com.au>.
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program; if not, write to the Free Software
+ Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+
+
+ ----------------------------------------------------------------------
+
+
+PSIM is a program written in extended ANSI-C that implements an
+instruction level simulation of the PowerPC environment. It is freely
+available in source code form under the terms of the GNU General
+Public License (version 2 or later).
+
+The PowerPC Architecture is described as having three levels of
+compliance:
+
+ UEA - User Environment Architecture
+ VEA - Virtual Environment Architecture
+ OEA - Operating Environment Architecture
+
+PSIM both implements all three levels of the PowerPC and includes (for
+each level) a corresponding simulated run-time environment.
+
+In addition, PSIM, to the execution unit level, models the performance
+of most of the current PowerPC implementations (contributed by Michael
+Meissner). This detailed performance monitoring (unlike many other
+simulators) resulting in only a relatively marginal reduction in the
+simulators performance.
+
+
+A description of how to build PSIM is contained in the file:
+
+ ftp://ftp.ci.com.au/pub/psim/INSTALL
+ or ftp://cambridge.cygnus.com/pub/psim/INSTALL
+
+while an overview of how to use PSIM is in:
+
+ ftp://ftp.ci.com.au/pub/psim/RUN
+or ftp://cambridge.cygnus.com/pub/psim/RUN
+
+This file is found in:
+
+ ftp://ftp.ci.com.au/pub/psim/README
+or ftp://cambridge.cygnus.com/pub/psim/README
+
+
+Thanks goes firstly to:
+
+ Corinthian Engineering Pty Ltd
+ Cygnus Support
+ Highland Logic Pty Ltd
+
+who provided the resources needed for making this software available
+on the Internet.
+
+More importantly I'd like to thank the following individuals who each
+contributed in their own unique way:
+
+ Allen Briggs, Bett Koch, David Edelsohn, Gordon Irlam,
+ Michael Meissner, Bob Mercier, Richard Perini, Dale Rahn,
+ Richard Stallman, Mitchele Walker
+
+
+ Andrew Cagney
+ Feb, 1995
+
+
+ ----------------------------------------------------------------------
+
+
+ What features does PSIM include?
+
+ Monitoring and modeling
+
+ PSIM includes (thanks to Michael Meissner)
+ a detailed model of most of the PowerPC
+ implementations to the functional unit level.
+
+
+ SMP
+
+ The PowerPC ISA defines SMP synchronizing instructions.
+ This simulator implements a limited, but functional,
+ subset of the PowerPC synchronization instructions
+ behaviour. Programs that restrict their synchronization
+ primitives to those that work with this functional
+ sub-set (eg P() and V()) are able to run on the SMP
+ version of PSIM.
+
+ People intending to use this system should study
+ the code implementing the lwarx instruction.
+
+ ENDIAN SUPPORT
+
+ PSIM implements the PowerPC's big and little (xor
+ endian) modes and correctly simulates code that
+ switches between these two modes.
+
+ In addition, psim can model a true little-endian
+ machine.
+
+ ISA (Instruction Set Architecture) models
+
+ PSIM includes a model of the UEA, VEA and OEA. This
+ includes the time base registers (VEA) and HTAB
+ and BATS (OEA).
+
+ In addition, a preliminary model of the 64 bit
+ PowerPC architecture is implemented.
+
+ IO Hardware
+
+ PSIM's internals are based around the concept
+ of a Device Tree. This tree intentionally
+ resembles that of the Device Tree found in
+ OpenBoot firmware. PSIM is flexible enough
+ to allow the user to fully configure this device
+ tree (and consequently the hardware model) at
+ run time.
+
+ Run-time environments:
+
+ PSIM's UEA model includes emulation for BSD
+ based UNIX system calls.
+
+ PSIM's OEA model includes emulation of either:
+
+ o OpenBoot client interface
+
+ o MOTO's BUG interface.
+
+
+ Floating point
+
+ Preliminary support for floating point is included.
+
+
+ Who would be interested in PSIM?
+
+ o the curious
+
+ Using psim, gdb, gcc and binutils the curious
+ user can construct an environment that allows
+ them to play with PowerPC Environment without
+ the need for real hardware.
+
+
+ o the analyst
+
+ PSIM includes many (contributed) monitoring
+ features which (unlike many other simulators)
+ do not come with a great penalty in performance.
+
+ Thus the performance analyst is able to use
+ this simulator to analyse the performance of
+ the system under test.
+
+ If PSIM doesn't monitor a components of interest,
+ the source code is freely available, and hence
+ there is no hinderance to changing things
+ to meet a specific analysts needs.
+
+
+ o the serious SW developer
+
+ PSIM models all three levels of the PowerPC
+ Architecture: UEA, VEA and OEA. Further,
+ the internal design is such that PSIM can
+ be extended to support additional requirements.
+
+
+ What performance analysis measurements can PSIM perform?
+
+ Below is the output from a recent analysis run
+ (contributed by Michael Meissner):
+
+ For the following program:
+
+ long
+ simple_rand ()
+ {
+ static unsigned long seed = 47114711;
+ unsigned long this = seed * 1103515245 + 12345;
+ seed = this;
+ /* cut-cut-cut - see the file RUN.psim */
+ }
+
+ Here is the current output generated with the -I switch on a P90
+ (the compiler used is the development version of GCC with a new
+ scheduler replacing the old one):
+
+ CPU #1 executed 41,994 AND instructions.
+ CPU #1 executed 519,785 AND Immediate instructions.
+ .
+ .
+ .
+ CPU #1 executed 1 System Call instruction.
+ CPU #1 executed 207,746 XOR instructions.
+
+ CPU #1 executed 23,740,856 cycles.
+ CPU #1 executed 10,242,780 stalls waiting for data.
+ CPU #1 executed 1 stall waiting for a function unit.
+ .
+ .
+ .
+ CPU #1 executed 3,136,229 branch functional unit instructions.
+ CPU #1 executed 16,949,396 instructions that were accounted for in timing info.
+ CPU #1 executed 871,920 data reads.
+ CPU #1 executed 971,926 data writes.
+ CPU #1 executed 221 icache misses.
+ CPU #1 executed 16,949,396 instructions in total.
+
+ Simulator speed was 250,731 instructions/second
+
+
+ What motivated PSIM?
+
+ As an idea, psim was first discussed seriously during mid
+ 1994. At that time its main objectives were:
+
+
+ o good performance
+
+ Many simulators loose out by only providing
+ a binary interface to the internals. This
+ interface eventually becomes a bottle neck
+ in the simulators performance.
+
+ It was intended that PSIM would avoid this
+ problem by giving the user access to the
+ full source code.
+
+ Further, by exploiting the power of modern
+ compilers it was hoped that PSIM would achieve
+ good performance with out having to compromise
+ its internal design.
+
+
+ o practical portability
+
+ Rather than try to be portable to every
+ C compiler on every platform, it was decided
+ that PSIM would restrict its self to supporting
+ ANSI compilers that included the extension
+ of a long long type.
+
+ GCC is one such compiler, consequently PSIM
+ should be portable to any machine running GCC.
+
+
+ o flexibility in its design
+
+ PSIM should allow the user to select the
+ features required and customise the build
+ accordingly. By having the source code,
+ the compiler is able to eliminate any un
+ used features of the simulator.
+
+ After all, let the compiler do the work.
+
+
+ o SMP
+
+ A model that allowed the simulation of
+ SMP platforms with out the large overhead
+ often encountered with such models.
+
+
+ PSIM achieves each of these objectives.
+
+
+ Is PSIM PowerPC Platform (PPCP) (nee CHRP) Compliant?
+
+ No.
+
+ Among other things it does not have an Apple ROM socket.
+
+
+ Could PSIM be extended so that it models a CHRP machine?
+
+ Yes.
+
+ PSIM has been designed with the CHRP spec in mind. To model
+ a CHRP desktop the following would need to be added:
+
+ o An apple ROM socket :-)
+
+ o Model of each of the desktop IO devices
+
+ o An OpenPIC device.
+
+ o RTAS (Run Time Abstraction Services).
+
+ o A fully populated device tree.
+
+
+ Is the source code available?
+
+ Yes.
+
+ The source code to PSIM is available under the terms of
+ the GNU Public Licence. This allows you to distribute
+ the source code for free but with certain conditions.
+
+ See the file:
+
+ ftp://archie.au/gnu/COPYING
+
+ For details of the terms and conditions.
+
+
+ Where do I send bugs or report problems?
+
+ There is a mailing list (subscribe through majordomo@ci.com.au) at:
+
+ powerpc-psim@ci.com.au
+
+ If I get the ftp archive updated I post a note to that mailing list.
+ In addition your welcome to send bugs or problems either to me or to
+ that e-mail list.
+
+ This list currently averages zero articles a day.
+
+
+ Does PSIM have any limitations or problems?
+
+ PSIM can't run rs6000/AIX binaries - At present PSIM can only
+ simulate static executables. Since an AIX executable is
+ never static, PSIM is unable to simulate its execution.
+
+ PSIM is still under development - consequently there are going
+ to be bugs.
+
+ See the file BUGS (included in the distribution) for any
+ other outstanding issues.
+