File: vice_2.html

package info (click to toggle)
vice 3.0.0.dfsg-2
links: PTS
area: contrib
in suites: stretch
size: 137,780 kB
ctags: 148,758
sloc: ansic: 1,389,701; asm: 91,194; sh: 35,211; cpp: 34,595; makefile: 16,586; objc: 12,688; java: 4,250; perl: 3,513; lisp: 1,795; python: 1,095; yacc: 1,056; xml: 779; lex: 551; pascal: 215; sed: 178
file content (821 lines) | stat: -rw-r--r-- 29,004 bytes
<HTML>
<HEAD>
<!-- This HTML file has been created by texi2html 1.52
     from ../../doc/vice.texi on 15 December 2016 -->

<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<link href='http://fonts.googleapis.com/css?family=Lato:400,700' rel='stylesheet' type='text/css'>
<TITLE>VICE Manual - 2  About VICE</TITLE>
<style type="text/css">
    @import 'vice.css';
</style>
</HEAD>
<BODY>
Go to the <A HREF="vice_1.html">first</A>, <A HREF="vice_1.html">previous</A>, <A HREF="vice_3.html">next</A>, <A HREF="vice_22.html">last</A> section, <A HREF="vice_toc.html">table of contents</A>.
<P><HR><P>


<H1><A NAME="SEC5" HREF="vice_toc.html#TOC5">2  About VICE</A></H1>

<P>
VICE is the one and only <EM>Versatile Commodore Emulator</EM>.  It provides
emulation of the Commodore C64, C64DTV, C128, VIC20, PET, PLUS4, SCPU64 and CBM-II computers
within a single package.  The emulators run as separate programs, but have
the same user interface, share the same settings and support the same
file formats.

</P>
<P>
<STRONG>Important notice:</STRONG> If you have no idea what a Commodore
8-bit computer is, or have questions about how these machines are used,
how the file formats work or anything else that is not strictly
related to VICE, you should read the appropriate FAQs <EM>first</EM>, as
that kind of information is not available here.  See section <A HREF="vice_19.html#SEC418">19  Contact information</A>. for
information about how to retrieve the FAQs.

</P>
<P>
All the emulators provide an accurate 6502/6510 emulator, with emulation
of all the opcodes (both documented and undocumented ones) and accurate
timing.  Unlike other emulators, VICE aims to be cycle
accurate; it tries to emulate chip timings as precisely as possible and
does so <EM>efficiently</EM>.

</P>
<P>
Please do <EM>not</EM> expect the C64DTV, C128, PET, PLUS4, SCPU64 and CBM-II emulators to
be as good as the C64 or VIC20 one, as they are still under construction.

</P>
<P>
<EM>Notice:</EM> This documentation is written for the Unix release of VICE, but is slowly being made universal.

</P>



<H2><A NAME="SEC6" HREF="vice_toc.html#TOC6">2.1  C64 emulator features</A></H2>

<P>
As of version 2.3, two C64 emulators are provided: <SAMP>`x64'</SAMP> (fast)
and <SAMP>`x64sc'</SAMP> (accurate).

</P>
<P>
The fast C64 emulator, called <SAMP>`x64'</SAMP>, features a fairly complete
emulation of the VIC-II video chip: sprites, all registers and all video
modes are fully emulated.  The emulation has been fully cycle-accurate
since version 0.13.0.

</P>
<P>
The accurate C64 emulator, called <SAMP>`x64sc'</SAMP>, features a cycle-based
and pixel-accurate VIC-II emulation. This requires a much faster machine
than the old <SAMP>`x64'</SAMP>.

</P>
<P>
A rather complete emulation of the SID sound chip is also provided.  All
the basic features are implemented as well as most of the complex ones
including synchronisation, ring modulation and filters.  There are two
emulators of the SID chip available: first is the "standard" VICE
emulator, available since VICE 0.12; the second is Dag Lem's reSID
engine. The reSID engine is a lot more accurate than the standard engine, 
but it is also a lot slower, and only suitable for faster machines.

</P>
<P>
Naturally, also both CIAs (or VIAs, in some cases) are fully emulated
and cycle accurate.

</P>



<H2><A NAME="SEC7" HREF="vice_toc.html#TOC7">2.2  C64DTV emulator features</A></H2>

<P>
The C64DTV emulator, called <SAMP>`x64dtv'</SAMP>, features emulation of C64DTV
revisions 2 and 3. The emulator is under construction, but most of the
DTV specific features are already supported (with varying accuracy).

</P>
<P>
Video cache is disabled by default as it currently doesn't work with
some of C64DTV's new video modes. The new video modes have a simple
"fake" video cache implementation that may give incorrect results
and decreased performance.

</P>



<H2><A NAME="SEC8" HREF="vice_toc.html#TOC8">2.3  C128 emulator features</A></H2>

<P>
The C128 emulator, called <SAMP>`x128'</SAMP>, features a complete emulation of
the internal MMU (<EM>Memory Management Unit</EM>), 80 column VDC screen, fast
IEC bus emulation, 2 MHz mode, Z80 emulation plus all the features of the
C64 emulation.

</P>



<H2><A NAME="SEC9" HREF="vice_toc.html#TOC9">2.4  VIC20 emulator features</A></H2>

<P>
The VIC20 emulates all the internal hardware, including the VIA chips.
The VIC-I video chip is fully emulated except NTSC interlace mode, so most
graphical effects will work correctly.

</P>
<P>
Sound support is implemented, but is still at an experimental stage.  If
you think it could be improved and know how to do so, feel
free to contact us (see section <A HREF="vice_19.html#SEC418">19  Contact information</A>).

</P>
<P>
The VIC20 emulator now allows the use of the VIC1112 IEEE488
interface. You have to enable the hardware (by menu, resource, or
commandline option) and then load the IEEE488 ROM (see for
example <CODE>http://www.funet.fi/pub/cbm/schematics/cartridges/vic20/ieee-488/325329-04.bin</CODE>, but you have to double the size to 4k for now).
The IEEE-488 code is then started by <CODE>SYS45065</CODE>.

</P>



<H2><A NAME="SEC10" HREF="vice_toc.html#TOC10">2.5  PET emulator features</A></H2>

<P>
The PET emulator emulates the 2001, 3032, 4032, 8032, 8096, 8296 and
SuperPET (MicroMainFrame 9000) models, covering the whole series.
The hardware is pretty much the same in each and that is why one single
program is enough to emulate all of them.  For more detailed information
about PET hardware please refer to the <TT>`PETdoc'</TT> file.

</P>
<P>
Both the 40 column and 80 column CRTC video chips are emulated (from the
4032 onward), but a few of the features are not implemented yet (numbers
of rasterlines per char and lines per screen).  Fortunately, they are
not very important for average applications.

</P>
<P>
Sound is available for the PET as well, but like the VIC20's it is still
under construction.

</P>
<P>
The PET 8096 is basically a PET 8032 with a 64k extension board which
allows remapping the upper 32k with RAM.  You have to write to a special
register at <CODE>$fff0</CODE> to remap the memory.  The PET 8296 is a
8096 but with a completely redesigned motherboard with 128k RAM in
total.  Of the additional 32k RAM you can use only some in blocks of 4k,
but you have to set jumpers on the motherboard for it.  VICE uses the
command line options <SAMP>`-petram9'</SAMP> and <SAMP>`-petramA'</SAMP>
instead.  Also, the video controller can handle a larger address range.
The PET 8x96 model emulations run the Commodore LOS-96 operating system
- basically an improved BASIC 4 version with up to 32k for BASIC
text and 32k for variables.  See <TT>`PETdoc'</TT> for more information.

</P>
<P>
The PET 8296-G is a 8296 with built-in 8250 low-profile dual disk drive.

</P>
<P>
The PET 8296-GD is a 8296-GD with additionally a "HiRes Emulator" (HRE).
This is a cheaper version of a "HRG" hi-res board which was based on
Thomson chips. This version instead uses no additional hardware support
apart from some memory mapping tricks. It has supporting software in the
hre-*.bin rom files.

</P>
<P>
The SuperPET also is a PET 8032 with an expansion board.  It can map 4k
at a time out of 64k into the <CODE>$9***</CODE> area.  Also it has an ACIA
6551 for RS232 communication.  The 6809 CPU that is built into the
SuperPET is now emulated, since release 2.4, including the 6702 dongle
chip.

</P>
<P>
The Super-OS/9 MMU expansion, developed by TPUG (Toronto PET Users
Group) is also emulated.

</P>
<P>
The PET computers came with three major ROM revisions, so-called BASIC
1, 2 and 4, all of which are provided.  The PET 2001 uses the version 1,
the PET 3032 uses version 2, and the others use version 4.  The 2001 ROM
is horribly broken with respect to IEEE488 (they shipped it before they
tested it with the floppy drive, so only tape worked.  Therefore the
emulator patches the ROM to fix the IEEE488 routines.

</P>
<P>
As well as other low-level fixes the 2001 patch obtains the load address
for a program file from the first two bytes of the file.  This allows
the loading of both PET2001-saved files (that have $0400 as their load
address) and other PET files (that have $0401).  The PET2001 saves from
$0400 and not from $0401 as other PETs do.

</P>
<P>
Moreover, the secondary addresses used are now <CODE>0</CODE> and <CODE>1</CODE> for
load and save, respectively, and not arbitrary unused secondary
addresses.

</P>
<P>
To select which model to run, specify it on the
command line with the <CODE>-model MODEL</CODE> option, where
<CODE>MODEL</CODE> can be one of a list of PET model numbers, all
described in see section <A HREF="vice_7.html#SEC177">7.7.1  Changing PET model settings</A>

</P>



<H2><A NAME="SEC11" HREF="vice_toc.html#TOC11">2.6  CBM-II emulator features</A></H2>

<P>
The CBM-II emulator emulates several types of CBM-II models.  Those
models are known under different names in the USA and Europe.  In the
States they have been sold as <CODE>B128</CODE> and <CODE>B256</CODE>, in Europe as
<CODE>CBM 610</CODE>, <CODE>CBM 620</CODE> (low-profile case) or <CODE>CBM 710</CODE> and
<CODE>CBM 720</CODE> (high-profile case with monitor). In addition to that
now an experimental C510 emulation is included. The C510 (also known as
P500) is the little brother of the C600/700 machines. It runs at roughly
1 MHz and, surprise, it has a VIC-II instead of the CRTC. Otherwise
the different line of computers are very similar.

</P>
<P>
These computers are prepared to take a coprocessor board with an 8088 or
Z80 CPU.  Indeed there are models <CODE>CBM 630</CODE> and <CODE>CBM 730</CODE> that
supposedly had those processors.  However these models are not emulated.

</P>
<P>
The basic difference is the amount of RAM these machines have been
supplied with.  The <CODE>B128</CODE> and the <CODE>CBM *10</CODE> models had 128k
RAM, the others 256k. This implies some banking scheme, as the 6502 can
only address 64k.  And indeed those machines use a 6509, that can
address 1 MByte of RAM.  It has 2 registers at addresses 0 and 1.  The
indirect bank register at address 1 determines the bank (0-15) where the
opcodes <CODE>LDA (zp),Y</CODE> and <CODE>STA (zp),Y</CODE> take the data from.  The
exec bank register at address 0 determines the bank where all other read
and write addresses take place.

</P>
<P>
The business line machines (C6xx/7xx) have the RAM in banks 1-2, resp.
1-4. All available banks are used for BASIC, where program code is separated
from all variables, resp. from normal variables, strings and arrays that
are distributed over other banks. The C510 instead has RAM in banks 0 and 1,
and uses bank 1 for program and all variables. Bank 0, though, can be
accessed by the VIC-II to display graphics.

</P>
<P>
Many models have been expanded to more than the built-in memory.  In fact
some machines have been expanded to the full 1M.  Bank 15 is used as
system bank, with only little RAM, and lots of expansion cartridge ROM
area, the I/O and the kernal/basic ROMs.  Some models have been modified
to map RAM into the expansion ROM area.  Those modifications can be
emulated as well.

</P>
<P>
The different settings are described in see section <A HREF="vice_7.html#SEC196">7.8.1  Changing CBM-II model</A>.

</P>


<H2><A NAME="SEC12" HREF="vice_toc.html#TOC12">2.7  SCPU64 emulator features</A></H2>

<P>
The XSCPU64 emulator is a simulation of a C64 equipped with a SuperCPU64 V2B. Features:

</P>

<UL>
<LI>20 MHz asynchronous single cycle 65816 CPU core with proper dummy and invalid cycle handling.

<LI>128 KiB static RAM, 0-16 MiB SIMM RAM, 64-512 KiB EPROM emulated and their respective timing details.

<LI>All RAM optimization configurations supported with write buffer.

<LI>I/O area access delays, write through to SRAM implemented.

<LI>Memory mappings including cartridge and boot memory map and kernal shadow.

<LI>Hardware registers and switches implemented.

<LI>Replacement SCPU64 ROM compatible with the original to avoid distribution problems

<LI>It's using the single cycle VICII core for accurate simulation

</UL>

<P>
Still to do:

</P>

<UL>
<LI>Measure and verify VICII interrupt phase shift

<LI>Measure and verify BA phase shift

<LI>SIMM RAM extra 7.5 cycle refresh delay every 10us missing.

<LI>CPU NMI support for "reset" button

</UL>

<P>
The emulation is quite accurate but not perfect. If you code something timing
intensive using this simulation please always check it on real hardware to avoid
bad surprises.

</P>
<P>
The hardware itself is asynchronous in nature, therefore caution must be taken
to not do long timing loops without synchronization in 20 MHz mode. Also don't
squeeze out the last remaining cycles without leaving a safety buffer.
Synchronization points can be created by doing I/O reads or writes and leaving
a few hundred cycles left each frame will not hurt.

</P>
<P>
Otherwise it can happen that the code is running on this version of VICE or my
SCPU64 V2+C128D perfectly but nowhere else due to manufacturing variations and
frequency drifts.

</P>


<H2><A NAME="SEC13" HREF="vice_toc.html#TOC13">2.8  The keyboard emulation</A></H2>

<P>
There are two ways of emulating the keyboard in VICE.

</P>
<P>
The default way (<EM>symbolic mapping</EM>) is to map every key
combination to the corresponding key combination on the real machine:
for example, if you press <KBD>*</KBD>, which is bound to <KBD>Shift-8</KBD> on a
U.S. keyboard, in the C64 emulator, the emulated machine will have just
the <EM>unshifted</EM> <KBD>*</KBD> key pressed (as <KBD>*</KBD> is unshifted on the
C64 keyboard).  Likewise, pressing <KBD>'</KBD> on the same U.S. keyboard
without any shift key will cause the combination <KBD>Shift-7</KBD> to be
pressed in the emulated C64.  This way, it becomes quite obvious what
keys should be typed to obtain all the symbols.

</P>
<P>
There is, however, one problem with symbolic mapping: some keys really
need to be mapped specially regardless.  The most important examples
being, in the VIC20, C64 and C128 emulators, that <KBD>CTRL</KBD> is mapped
to <KBD>Tab</KBD> and that the <KBD>Commodore</KBD> key is mapped to the left
<KBD>Control</KBD>). The <KBD>RUN/STOP</KBD> key is mapped to the <KBD>ESC</KBD> key
on the PC keyboard. The PET emulator, lacking the <KBD>Commodore</KBD> key
but having an <KBD>ESC</KBD> key, uses the left <KBD>Control</KBD> key as
<KBD>RUN/STOP</KBD> and the <KBD>ESC</KBD> key as <KBD>ESC</KBD> of course.

</P>

<P>
The second way (<EM>positional mapping</EM>) is to map every key on the
"real" keyboard to the key which has the same position on the keyboard
of the emulated machine.  This way, no <KBD>Shift</KBD> key is forced by the
program (with the exception of the function keys <KBD>F2</KBD>, <KBD>F4</KBD>,
<KBD>F6</KBD> and <KBD>F8</KBD>, which require <KBD>Shift</KBD> on the Commodore
keyboards), and the keyboard is more comfortable to use in those
programs (such as some games) that require the keys to be in the correct
positions.

</P>
<P>
<STRONG>Warning:</STRONG> unlike the real C64, VICE "presses" the <KBD>Shift</KBD>
key <EM>together</EM> with the key to shift when the <KBD>Shift</KBD> must be
forced.  In most cases this should work fine, but some keyboard routines
are quite picky and tend not to recognize the shift key because of this.
For instance, <KBD>F6</KBD> (which on the real C64 is obtained with
<KBD>Shift + F5</KBD>) could be recognized as <KBD>F5</KBD>.  In that case, use
the shift key manually (i.e., type <KBD>Shift + F5</KBD> in the example).
Yes, we know this is a bug.

</P>
<P>
The <KBD>RESTORE</KBD> key is mapped to <KBD>Page Up</KBD> (or <KBD>Prev</KBD>) by
default.

</P>



<H2><A NAME="SEC14" HREF="vice_toc.html#TOC14">2.9  The joystick emulation</A></H2>

<P>
Joysticks can be emulated both via the keyboard and via a real joystick
connected to the host machine (the latter only works on GNU/Linux
systems).

</P>
<P>
There are two keyboard layouts for joystick use, known as <EM>numpad</EM>
and <EM>custom</EM>.

</P>
<P>
The <EM>numpad</EM> layout uses the numeric keypad keys, i.e., the numbers
<KBD>1</KBD>...<KBD>9</KBD> which emulate all the directions including the
diagonal ones; <KBD>0</KBD> emulates the fire button.

</P>
<P>
The <EM>custom</EM> layout uses the keys <KBD>w</KBD>, <KBD>e</KBD>, <KBD>r</KBD>,
<KBD>s</KBD>, <KBD>d</KBD>, <KBD>f</KBD>, <KBD>x</KBD>, <KBD>c</KBD>, <KBD>v</KBD> for the directions
and <KBD>space</KBD> for the fire button instead.

</P>



<H2><A NAME="SEC15" HREF="vice_toc.html#TOC15">2.10  The disk drive emulation</A></H2>

<P>
All the emulators support up to 4 external disk drives as
devices 8, 9, 10 and 11.  Each of these devices can emulate virtual
Commodore 1541, 1541-II, 1571, 1581, 2031, 2040, 3040, 4040, 1001, 8050 and
8250 drives in one of four ways:

</P>

<UL>
<LI>

using disk images, i.e., files that contain a dump of all the blocks
contained in a real floppy disk (if you want more information about
what a disk image is, consult the
<CODE>comp.emulators.cbm</CODE> FAQ);
<LI>

accessing file system directories, thus giving you the use of files
without having to copy them to disk images; this also allows you to
read and write files in the <CODE>P00</CODE> format (again, consult the
<CODE>comp.emulators.cbm</CODE> FAQ for more info).
<LI>

accessing a real device connected to the host machine. As of VICE 1.11
it is possible to connect real drives like Commodore 1541 to the
printer port of the host using the XA1541 or XM1541 cable. Currently
this only works on Linux or Windows using the OpenCBM library. You can
get it from <A HREF="http://www.lb.shuttle.de/puffin/cbm4linux">http://www.lb.shuttle.de/puffin/cbm4linux</A> (cbm4linux,
Linux version) or from <A HREF="http://cbm4win.sf.net">http://cbm4win.sf.net</A> (cbm4win, Windows
version).

<LI>

directly using the disk drive of the host. The 3.5" disk drive of the
host can be used to read or write Commodore 1581 formatted disks.
Currently this raw drive access feature is only available for Linux
hosts.
</UL>

<P>
When using disk images there are two available types of drive
emulation.  One of them the <EM>virtual drive</EM> emulation.  It does
<EM>not</EM> really emulate the serial line, but patches the kernal ROM
(with the so-called <EM>kernal traps</EM>) so that serial line operations
can be emulated via C language routines.  This emulation is very fast,
but only allows use of standard DOS functions (and not even all of
them).  For real device or raw drive access it is required to enable
this type of emulation.

</P>
<P>
The IEEE488 drives (2031, 2040, 3040, 4040, 1001, 8050 and 8250) do
not use kernal traps. Instead the IEEE488 interface lines are
monitored and the data is passed to the drive emulation. To use them
on the C64, you need to enable the IEEE488 interface emulation. Only
if the IEEE488 emulation is enabled, those drives can be selected.

</P>
<P>
The other alternative is a <EM>true drive</EM> emulation.  The
Commodore disk drives are provided with their own CPU (a 6502 as the
VIC20 and the PETs) and their own RAM and ROM.  So, in order to more
closely emulate its features, a complete emulation of this hardware
must be provided and that is what the <EM>hardware level</EM> emulation
does. When the <EM>hardware level</EM> emulation is used, the kernal
routines remain unpatched and the serial line is fully emulated.
The problem with this emulation is that it needs a lot of processing
power, mainly because the emulator has to emulate two CPUs instead of
one.

</P>
<P>
The PETs do not use a serial IEC bus to communicate with the floppy
drive but instead use the parallel IEEE488 bus.  This does
<EM>byte by byte</EM> transfers, as opposed to the <EM>bit by bit</EM>
transfers of the C64 and VIC20, so making it feasible to emulate the
parallel line completely while emulating the drive at DOS level only.
The IEEE488 line interpreter maps the drives 8-11 (as described
above) to the IEEE488 disk units, and no kernal traps are needed.
The same emulation of the Commodore IEEE488 bus interface is
available for the C64 and the VIC20. With IEEE488 drives you can have
true 2031 emulation at unit #8, and still have filesystem access at
units #10 or #11, because monitoring the IEEE488 lines does not
interfere with the true drive emulation.

</P>
<P>
The IEEE488 disk drives 2040, 3040, 4040, 8050 and 8250 are Dual Drive
Floppy Disks.  This means that these drives handle two disks. To
accomplish the emulation, only two disks can be emulated, namely units
#8 and #10. The attached image, track display and LED display of unit
#9 and #11 are used for the second drive of the dual disk drives. On
unix the unit number display (8 or 9, 10 or 11) in the emulation
window changes to the drive number display (0 or 1).

</P>
<P>
The Commodore 2040, 3040, 4040, 1001, 8050 and 8250 disk drives are
so-called "old-style" disk drives. Their architecture includes not
one, but two processors of the 6502 type, namely a 6502 for the file
handling and communication with the PET (IP), and a 6504 (which is a
6502 with reduced address space) for the drive handling (FDC). Both
processors communicate over a shared memory area. The IP writes
commands to read/write blocks to this area and the FDC executes them.
To make the emulation feasible, the FDC processor is not emulated
cycle-exactly as a 6504, but simply by checking the commands and
executing them on the host. This provides a fast FDC emulation, but
disallows the sending the FDC processor commands to execute code.
Applications where this is necessary are believed to be rather
seldom. Only the format command uses this feature, but this is
checked for.

</P>
<P>
The dual disk drive 2040 emulates one of the very first CBM disk
drives. This drive has DOS version 1. DOS1 uses an own disk type,
that is closely related to the 1541 disk image. Only on tracks 18-24
DOS1 disks have a sector more than 1541 disks. DOS1 disk images have
the extension .d67.

</P>
<P>
The dual disk drives 3040 and 4040 use the same logical disk format
as the VC1541 and the 2031. In fact, the 4040 was the first disk with
DOS version 2. The 3040 emulated here originally was the same as
2040, only for the european 30xx PET series. As many of the original
DOS1 disk drives were upgraded (a simple ROM upgrade!) to DOS2, I use
the 3040 number for a DOS 2.0 disk drive, and 4040 for a revised DOS
2 disk drive. It is, however, not yet clear whether the disks here
are write compatible to the 1541, as rumors exist that the write gap
between  sectors is different. But read compatible they are. As VICE
emulates the FDC processor in C and not as 6504 emulation, this does
not matter in VICE.

</P>
<P>
The drives 1001, 8050 and 8250 do actually have the very same DOS
ROM. Only the code in the FDC is different, which is taken care of by
VICE. So for all three of those disk drives, only <CODE>dos1001</CODE> is
needed. The DOS version used is 2.7.

</P>


<H2><A NAME="SEC16" HREF="vice_toc.html#TOC16">2.11  Supported file formats</A></H2>

<P>
VICE supports the most popular Commodore file formats:

</P>

<UL>

<LI>

<CODE>X64</CODE> (preferred) or <CODE>D64</CODE> disk image files; Used by the 1541, 2031, 3040, 4040 drives.

<LI>

<CODE>G64</CODE> GCR-encoded 1541 disk image files

<LI>

<CODE>P64</CODE> NRZI flux pulse disk image files

<LI>

<CODE>D67</CODE> CBM2040 (DOS1) disk image format

<LI>

<CODE>D71</CODE> VC1571 disk image format

<LI>

<CODE>D81</CODE> VC1581 disk image format

<LI>

<CODE>D80</CODE> CBM8050 disk image format

<LI>

<CODE>D82</CODE> CBM8250/1001 disk image format

<LI>

<CODE>D1M</CODE> FD2000/FD4000 DD disk image format

<LI>

<CODE>D2M</CODE> FD2000/FD4000 HD disk image format

<LI>

<CODE>D4M</CODE> FD4000 ED disk image format

<LI>

<CODE>T64</CODE> tape image files (read-only)

<LI>

<CODE>P00</CODE> program files

<LI>

<CODE>CRT</CODE> C64 cartridge image files

</UL>

<P>
An utility (<CODE>c1541</CODE>, see section <A HREF="vice_13.html#SEC301">13  c1541</A>) is provided to allow transfers
and conversions between these formats.

</P>
<P>
Notice that the use of the <CODE>X64</CODE> file format is depreciated now.

</P>
<P>
<A NAME="IDX1"></A>
You can convert an <CODE>X64</CODE> file back into a <CODE>D64</CODE> file with the
UNIX <CODE>dd</CODE> command:

</P>

<PRE>
dd bs=64 skip=1 if=IMAGE.X64 of=IMAGE.D64
</PRE>

<P>
See section <A HREF="vice_16.html#SEC313">16  The emulator file formats</A>. for a technical description of the supported file
formats.

</P>


<H2><A NAME="SEC17" HREF="vice_toc.html#TOC17">2.12  Common problems</A></H2>

<P>
This section tries to describe the most common known problems with VICE,
and how to resolve them.

</P>



<H3><A NAME="SEC18" HREF="vice_toc.html#TOC18">2.12.1  Sound problems</A></H3>

<P>
VICE should compile and run without major problems on many UNIX systems,
but there are some known issues related to the sound driver.  In fact,
the sound code is the least portable part of the emulator and has not
yet been thoroughly tested on all the supported platforms.

</P>
<P>
Linux, AIX and SGI systems should play sound without any problems; if
you are running Linux please use a 2.x kernel, as VICE needs some
features that were not implemented in older versions of the Linux sound
driver.

</P>
<P>
<A NAME="IDX2"></A>
On the other hand, HP-UX and Solaris machines are known to cause
troubles.  If you think you can help debugging the code for these
systems, your help would be really appreciated.  We are having troubles
finding HP-UX and SUN consoles to work at...

</P>
<P>
<A NAME="IDX3"></A>
Some problems have been reported with the proprietary version of the
Open Sound System for Linux.  With a Crystal sound card, sound output
was significantly delayed and, apparently, the allocated buffer size was
completely wrong.  This is not a VICE bug, but rather an OSS bug.

</P>



<H3><A NAME="SEC19" HREF="vice_toc.html#TOC19">2.12.2  Shared memory problems</A></H3>

<P>
<A NAME="IDX4"></A>
<A NAME="IDX5"></A>
If you cannot start VICE because you get errors about shared memory, try
to run it with the <SAMP>`+mitshm'</SAMP> command-line option.
 This will completely disable usage of the MITSHM extensions,
that are normally used to speed up the emulation window updates.  Of
course, this will also result in a big loss in speed.

</P>
<P>
Reasons for this failure could be:

</P>

<UL>
<LI>

IPC support has been disabled at the system level; some system
administrators disable this for security reasons.  If <EM>you</EM> are the
system administrator, use a kernel that has IPC support compiled in and
enabled.
<LI>

You are attempting to run the emulator across the network (i.e., the
emulator runs on one machine, and the output is displayed on another
machine that works as an X terminal) and for some reason VICE does not
recognize this fact.  In this case, you have found a bug, so please report it
to us.
</UL>

<P>
If you want to avoid running the emulator with <SAMP>`+mitshm'</SAMP> every
time, run it once with <SAMP>`+mitshm'</SAMP> and then choose "Save settings"
from the right-button menu.

</P>


<H3><A NAME="SEC20" HREF="vice_toc.html#TOC20">2.12.3  Printer problems</A></H3>

<P>
VICE supports the emulation of a printer either on the userport or as
IEC device 4. Unfortunately the Commodore IEC routines do not
send all commands to the IEC bus. For example an <CODE>OPEN 1,4</CODE>
is not seen on the IEC bus. Also a <CODE>CLOSE 1</CODE> after that
is not seen. VICE can see from printing that there was an <CODE>OPEN</CODE>,
but it cannot see when the close was. Also a "finish print job"
cannot be seen on the userport device.
To flush the printer buffer (write to <CODE>print.dump</CODE> or to the
printer) now a menu entry can be used. Disabling and re-enabling the printer
should work as well.

</P>
<P>
The printing services have not been extensively tested but apart
from the problem mentioned above it should work fine now.

</P>


<H3><A NAME="SEC21" HREF="vice_toc.html#TOC21">2.12.4  PET keyboard problems</A></H3>

<P>
If you find that the German keyboard mapping (plus German charset)
does not print uppercase umlauts, then you are right.
The umlauts replace the [,\ and ] characters in the charset. The keys
that make these characters do not have a different entry in the
PET editor ROM tables when shifted.
Thus it is not possible to get the uppercase umlauts in the editor.
Nevertheless other programs are reported to change the keyboard
mapping table and thus allow the use of the shifted (uppercase) umlauts.

</P>
<P>
Anyway, the VICE keyboard mappings are far from being perfect and we are open
to any suggestions.

</P>
<P><HR><P>
Go to the <A HREF="vice_1.html">first</A>, <A HREF="vice_1.html">previous</A>, <A HREF="vice_3.html">next</A>, <A HREF="vice_22.html">last</A> section, <A HREF="vice_toc.html">table of contents</A>.
</BODY>
</HTML>