Download Latest Version install.zip (29.3 MB)
Home
Name Modified Size InfoDownloads / Week
vectorpascalcom 2011-06-24
OldFiles 2008-06-25
vectorpascal_1.0-1_all.deb 2026-10-29 26.9 MB
manual.pdf 2026-10-21 1.2 MB
readme 2026-10-21 5.3 kB
install.jar 2026-10-17 29.3 MB
install.zip 2026-10-17 29.3 MB
ISO_and_extensions_conformance_test_report.html 2017-12-05 103.9 kB
Totals: 8 Items   86.9 MB 19
Latest features

October 2026 	- installation via .deb files

		- integrated vector pascal IDE, launch using 
		  $vpc myfile.pas -IDE
                  IDE supports direct entry of Greek and APL characters

From July 2026 the support for gpus via opencl has been improved. The -gpu flag will now only send matrix product operations to the gpu since these show significant speedup relative to other array ops which tend to run slower than on the host.
On an Nvidia Tesla M60 you can get roughly a 8 fold speedup relative to a Skylake cpu for matrix product. Even on an arm64 you get a 5 fold acceleration of large matrix multiplies by using the Mali gpu.

From January 2026 there is support for orthogonal persistence. See
the manual for details.
Now supports AVX512 by using the -cpuX86_64v4 flag.
Additional APL style operations supported
PREF and SUFF generate prefix and suffix vectors after style of Iversons alpha and omega operators.
In fact the appropriate APL unicode are recognised by the compiler.
The APL style rotate operator on arrays has been added, written either as ROT or as the corresponding APL symbol.
The INDEX function has been introduced to do the same as the APL select construct.
For example
program select;
const n=40;m=10;
var nums:array[1..n] of integer;
    mask:array[1..n] of boolean;
    indices:pivector;
begin
	nums := random mod 1000 ;
	mask := (nums mod m)<2;
	indices:=index(mask[]);
	writeln(indices^);
	writeln(nums[indices^]);
end.

We initialise nums with numbers in 0..999, and form a mask of those numbers
whose last digits are 0 or 1.
We use  index} to form the indices of where the mask array was true\footnote{The index funtion
takes a dynamic array slice as its parameter, we form a slice comprising the whole array with
the construct {\tt mask[]}.}.
We then produce a table of the positions and numbers at these positions


Installation details

Debian systems:
If you are on a Debian derived Linux download the .deb file for Vector Pascal and 
install via

$sudo apt install ./vectorpascal_1.0-1_all.deb

After this the command vpc will be available

Non Debian systems:

Otherwise unpack  a recent version of either the install.jar file or the install.zip file.

You can do this by using the
jar -xf install.jar
command, or by usings an unzip package.

This will create a directory mmpc containing the files needed by the compiler.  

Non Debian Dependencies
------------
Java, Nasm, Gcc, atyle, sablecc, M4 macro processor if you are doing source development work.
Nvidia GPU requires nvcc in addition to above, and a GPU with compute capability 3.0 and above.
But if you install via the .deb file the dependencies required for non gpu code will
automatically be downloaded.


 

If installing from source, you should run the command
make install
to create the installation archives.


On Linux
--------
If you are just running the compiler to compile Pascal source you
need :  Java, Nasm, Gcc and Nvcc if compiling to an Nvidia GPU.

GCC is used to build and link runtime libraries to the assembler
produced by Vector Pascal. On Linux the standard GCC compiler will work.

 
If you are not installing via the .deb file then, place the directory mmpc on your path with the following command:
export PATH=$PATH:/path/to/mmpc
 

If you have altered the locations of the installed files from the default, you
will need to define the symbol mmpcdir to point at the directory containing
mmpc.jar by running the following command:
export mmpcdir=/path/to/mmpc


On Windows
----------

Warning: The native Windows system is no longe actively maintained, you are
advised to install WSL and install via the .deb file now.

When working on Windows some incompatibilities exist between different releases
of GCC and of assemblers. The NASM assembler under Windows will only work
reliably with the djgpp release of GCC and not with Cygwin. Given Cygwin's easy
installation, when working with Windows, the prefered method is to use the
Cygwin compiler, not the DJGPP compiler. If using the cygwin compiler
specify that you are using the gnu versions of the code generator
as this is required to produce gnu compatible assembler code.
Examples of using the gnu versions of the compiler would be to specify
-cpugnuP4 or -cpugnuPentium

To install, first obtain a copy of Cygwin from http://cygwin.com/setup.exe
or http://www.cygwin.com/

When installing Cygwin, add the "make" package to the default programs to
install.

Once Cygwin is installed, extract the VPC archive using the following:
unzip install.win32.zip

You will now need to build a compatible RTL by doing the following:
Go into the directory mmpc and type
make rtl
Then type
make rtl.o

The environment variable "mmpcdir" will have to be set. This can be done by
right clicking on "My Computer" and selecting "Properties". Select the
"Advanced" tab and clicking on the "Environment Variables" button. mmpcdir
should then be set to the path to the files you extracted earlier.

Additionally, you may choose to add the same directory to your path.

To execute, you can use either Cygwin's shell or a standard Dos prompt.

=================================================================
If you want to use the pretty printer facility built into the compiler, then
you will need Latex.

 

 
Source: readme, updated 2026-10-21