Reply by Michael Kellett●August 16, 20182018-08-16
On 11/08/2018 20:21, gnuarm.deletethisbit@gmail.com wrote:
> On Saturday, August 11, 2018 at 12:18:27 PM UTC-4, Michael Kellett wrote:
>> On 09/08/2018 23:28, othmana@gmail.com wrote:
>>> http://mymicroprocessor.blogspot.com/2018/08/fpga-simplest-processor.html
>>>
>>> Go to my blog for more information.
>>>
>> Why not make it easy for us and just give us the UK patent reference here.
>
> What is easy about reading patents?
>
> The US transmits a time code signal from Colorado which isn't always receivable on the east coast. Some 10 years ago (or so) they modified the signal to include phase modulation which should be easier to receive. The only trouble is a company (who may have worked with the US government) got a patent for a receiver of this phase modulated signal. I can't be sure of what I'm reading in the patent, so I can't try to design around it and sell a phase demodulated receiver. If they were easy to read, I would know just what had been patented and would be able to design a non-infringing receiver.
>
> Rick C.
>
Some patents are more readable than others. This is an especially chatty
and helpful one. I get the feeling that UK patent style is tending more
towards sensible descriptions but that may just be a reflection of the
ones I've happened to look at.
MK
Reply by ●August 15, 20182018-08-15
On Thursday, 16 August 2018 04:03:15 UTC+8, gnuarm.del...@gmail.com wrote:
>=20
> Ok, I looked at it a bit. Is there something special about this particul=
ar design? What were your goals in designing it? What aspects of this des=
ign excel in small CPUs?=20
>=20
> Rick C.
decoding logic is simplest. Only 16 instructions but unlimited number of op=
erators. CPU sim cannot handle this cpu. I need to pretend a few instructio=
ns correspond to some operators.
You can start with one operator. Immediate data is all used up for the lite=
ral or constant.
If you need more operators, allocate some bits in the immediate data field.
It is the simplest because the 3 opcode bits out of 4, directly control the=
mux in the data path.
The jump bit, need to share with zero detector in order to provide conditio=
nal jump.
call return instructions are provided so make the decoding somewhat complic=
ated but not by much. In many cases they are not required. For example, the=
Lagged Fibonacci Random number generator does not need a call and therefor=
e return instruction.
I have 2 reserve instructions but committing them to work will need more co=
mplicated decoding logic, which may not be justified.
E.g. in one reserve instrucion, I wanted to implement a NOP. It is implemen=
ted by decoding its 4 bit pattern, and then use it to disable register and =
memory write. Just these two functions already crashed my Quartus. Still tr=
ying to debug them. Imagine, just by adding 2 2-bit and gates, can crash Qu=
artus FPGA CAD.
Maybe, it is better not to implement the NOP instruction. We should find ot=
her alternatives, such as duplicating instructions.
Reply by ●August 15, 20182018-08-15
On Monday, August 13, 2018 at 7:10:14 AM UTC-4, oth...@gmail.com wrote:
Ok, I looked at it a bit. Is there something special about this particular design? What were your goals in designing it? What aspects of this design excel in small CPUs?
Rick C.
Reply by ●August 13, 20182018-08-13
On Sunday, 12 August 2018 00:18:27 UTC+8, Michael Kellett wrote:
> >=20
> > Go to my blog for more information.
> >=20
> Why not make it easy for us and just give us the UK patent reference here=
.
What is easy about reading patents? =20
The US transmits a time code signal from Colorado which isn't always receiv=
able on the east coast. Some 10 years ago (or so) they modified the signal=
to include phase modulation which should be easier to receive. The only t=
rouble is a company (who may have worked with the US government) got a pate=
nt for a receiver of this phase modulated signal. I can't be sure of what =
I'm reading in the patent, so I can't try to design around it and sell a ph=
ase demodulated receiver. If they were easy to read, I would know just wha=
t had been patented and would be able to design a non-infringing receiver.=
=20
Rick C.
Reply by Michael Kellett●August 11, 20182018-08-11