Reflections - New-Age-Computer or PC XXI century.

<b>Reflections</b> - New-Age-Computer or PC XXI century.

(C) Ivan Mak
Music: BY KENOTRON ISTL



           *** New-Age-Computer ***

            or the computer of the XXI century



     This article is in some degree of information in some - 
ad. Author of the article nor in any way claim to the 
uniqueness of this path of development and 
computer-protsessorostroeniya. Everything described below is a 
real plan, that may change if appears more effective solution 
to the problem. 



     Formulation of the problem.

     We want to create a new computer with
maximum efficiency at reduced cost and limited development 
costs. 

     Everyone has long known the basic computer parts. CPU, 
memory and peripherals. Throughout the world there are many 
different platforms, based on different processors. There is a 
set of adjacent platforms in the same processor as the central.

They all have differences with each other in the implementation.

     But all computers have common problems. The problems of 
speed and memory problems ... CPU speed is important, but often

it is limited by the rate of exchange with main memory.

     Computer memory is not unlimited, its performance is not 
unlimited. The question arises as as having limited memory and 
processor speed, to get maximum efficiency? One of the ways - 
paralleling processes. Increasing the number of data buses, but 
this path has its border. It is impossible to increase the 
number of tires to infinity. 8, 16, 32, 64 ... Then there is the

many complexities associated with technology. Changing 
technology - it is one of the ways of computers, but it is 
quite expensive and in fact can not be applied to solve our 
problem. 


     Thus, what we have? Imagine that we have a limited amount 
of memory. 128 kilobytes, organized as 64K x 16bit. High-speed 
memory, for example, sampling 15 nanoseconds. What computer, 
what options can be built on that memory? 

     Existing well-known 16-bit processors, or do not have 
enough speed or have the architecture, designed to significantly

more memory, which leads to a rather
inefficient use of memory. In addition, if you look at the 
cycles of the processor memory, then the naked eye can see the 
gaps time it takes for the organization of work with the bus, 
etc. etc. Dear exotic processors of course, does not count.


     The question arises as to obtain maximum performance, 
unless there is a suitable processor? 

     The answer may seem delusional. Solution -
establish its own processor. Whether it is
another five or ten years ago, on this response could be a good 
laugh, but today such is indeed possible! 


     How?

     To develop their own processor can
use of reprogrammable logic chips (PPLM). Modern PPLM reached a 
sufficient volume so that it may be built processors. This 
means that anyone with enough experience can do a similar 
designing processors. To do this, only have the appropriate 
software, PPLM, where you can conduct experiments, and, of 
course, the experimental model computer. 

     Oddly enough, but now exists a computer that can be done
similar experiments. He developed several
other purpose and is a development line of computers 
ZX-Spectrum. 

     This computer - Sprinter.



     What is Sprinter?

     The scheme is based on a large computer re-programmable 
logic chip firms ALTERA EPM10K10QC208-4. Connecting peripherals 
PPLM allows high flexibility for machine configurations. 
Programming PPLM directly at the moment

inclusion, as well as when you restart that allows you to 
drastically change the circuit in PPLM during operation. This 
greatly distinguishes computer architecture from a number of 
existing computers and therefore many of the concepts inherent 
in conventional machines, changing its meaning. In fact, the 
computer has a variable architecture in which changes are 
possible in many parts of the circuit. Programmability the 
computer's circuitry allows it to completely change that is to 
turn virtually any computer based on the Z80 processor, or CPU, 
which Z80 is compatible to your code. But this is topic for 
another article. 

     With regard to our problem, the computer contains all the 
necessary circuits. Namely: PLA, high-speed RAM, connected 
directly to the PLA as well as peripherals and all the necessary

elements for startup and testing
new processor.

     Thus, the computer Sprinter as well as possible
better suited for experiments with the creation of a processor. 
Naturally, this processor will be limited in its capabilities, 
instruction set, but it can be easily optimized for specific

task. Moreover, since PPLM can overload
diagram of the processor can be changed almost
at any moment (time reset circuit is a few tenths of a second).

     All this allows to reduce development cycle time to a few 
minutes of experimenting with new processor.



     How real is described here is designed to develop the 
processor? 

      The author of this article, conducted the first
experiments on the creation of its processor's
RISC-architecture. Experiments clearly show that the volume of 
PPLM installed in your computer Sprinter, enough to create a 
similar processor. Basis for the scheme was adopted 
mnogostekovaya processor architecture that best fits for 
FORTH-processor. 

     The obtained result - a simple FORTH-processor, which has 
4 stack, clock, and 7MHz cycle time of the execution team in 1 
or 2 cycles. In this case, the PLA is not fully used,

and the low CPU clock speed is due to
only a lack of experience in creating its author
processors.


     What are the prospects of such a processor?

     Naturally, the application developed so
as the processor inside the computer Sprinter is rather 
limited. He can afford only to solve additional tasks on the 
Sprinter-e. The value of development may be that the resulting 
processor can be used in other devices, developed at the same 
PLA. In the specialized controllers, for example. 


     But in our case, the processor design is to create a new 
computer with the new processor. What is this computer?


     Preliminary project title Sprinter-II.



     It is natural that it should be fairly
powerful, and it means that its basis must be
sufficient memory and speed. At the
However, it should be fairly cheap and
common, if it is assumed that the developed computer designed 
for mass replication. 

     These conditions, according to our estimates at this
time, satisfies the memory based on the DIMM, which allow you 
to organize the memory access speeds up to 100MHz and a data 
bus to 64 bits. 

     As the processor, in accordance with our installation uses 
PPLM firm ALTERA. Will be installed at PPLM EPF10K20, having

two times larger volume than EPF10K10, which will organize a 
32-bit processor, developed by a similar 16-bit. 

     In addition, several PPLM 10K is compatible to the 
location of the conclusions that will set in place EPF10K20 
PPLM greater volume, that will help organize even more powerful 
processor. 

     Developed a computer must have and the periphery. In the 
Sprinter-e all control peripherals through PPLM, but if

PPLM located in the processor, it makes sense to unload it. 
Therefore, the project Sprinter-II will be installed at the 
second PPLM designed purely to work with peripherals. This will 
enhance function peripherals and add new features

computer.

     Thus, the developed computer will have a fully 
programmable architecture, that will change it in accordance 
with the problem. 



     Sprinter-II project is one of the projects the firm 
Peters-Plus and is currently at the initial stage of 
development. 



     Sprinter computer maker is also a firm Peters-Plus 
(formerly Peters). 



     Information on computer Sprinter can be found at the firm 
Peters: www.atlant.ru / peters 


 E-mail: ivan_mak@mail.ru
 FIDO: 2:5030 / 529.24 aka 2:5030 / 776.24



                 *



     Upon receipt of this article, I decided to talk with John 
personally. Here's what came of it: 

VL> Sprinter II - is a brand new board, new
computer?

     Yes, absolutely.
VL> is compatible than the first?

     Compatible only partially. PLA will be others,
increased, respectively, will be more opportunities.

VL> Does not it seem that its price will be higher than the 
first model (and she had it and so ...)?


     Price really be higher. On the first
$ 140-160 estimations obtained. Let us think about how
reduce it. All depends on the price of PLA-ca. If
be able to find a cheaper supply channel, then
price decline.

VL> How much does the PLA 10 (20) (curious)?

     EPF10K10 costs 25 $, EPF10K20 - $ 50.
And Sprinter II will be installed at the two PLA.
Anyway, so it will be the first experimental sample. And we 
will continue to watch. Possible production of intermediate 
variants with a minimal set of functions. 

   From what I want to get on the Sprinter-II:
* Improving the overall speed of the processor in de tens times 
relative to Z80 at 21MHz * Video output as on TV, and on the 
VGA. * Video input from RGB to TV with the ability FRAME Vågå 
reading the image. * In the limit of direct recording from TV 
to the HDD. * Sound input / output.


   Internal architecture:
* Hardware multitasking;
* High-speed DMA channels;
* 32bit Fort processor.

VL> Where did this idea - make your CPU?

     The idea to make your processor has revolved
long. Just because the Z80 is not satisfied by
speed and capacity, and raise intellectual or Motorola simply 
uninteresting. From Intel will lean HRC, from Motorola - a thin 
Amiga. And if the processor her, then it can do everything and 
make your better than all the (hopefully better).


VL> What about the sale of the first Sprinter?

     With the sale of the first Sprinter deal worth almost the 
same place. Only recently was sold the first computer this 
year. : (It was an opportunity, gave the machine programmers. 
Now ready machines No, not counting the demonstration. Judging 
by the lack of letters to the CC999 is unlikely to be sold a 
lot of Sprinters. (According to the most pessimistic forecasts 
generally will not). 

VL> Can not this be! Wait and see ... Thanks for the 
information. All the best! 

     Thank you.