From: jasen on
On 2007-02-12, Ken Smith <kensmith(a)green.rahul.net> wrote:
>>> You had to load the CX register to do a LOOP.
>>
>>What's LOOP got to do with anything?
>
> You haven't been following the discussion. I used the case of a REP
> prefix used inside a loop as the example of why the 8086's instruction set
> was so poorly designed. The CX and the REP both use the CX so the CX must
> be loaded for the REP inside the loop. This means that the current CX
> contents must be saved, the CX loaded, the REP done and the CX restored.
> This is a lot of extra work.

It's less work to just use something else to hold the loop count
and use DEC and JNZ in place of LOOP (faster too since 80486)

One of the other registers or [bp+N] is often a good choice
(for apropriate values of N)

Bye.
Jasen
From: jasen on
On 2007-02-13, MassiveProng <MassiveProng(a)thebarattheendoftheuniverse.org> wrote:

>>>>I'll repeat it since you're a slow learner:
>>>>
>>>>A properly installed Linux uses all the available
>>>>partitions.
>>>
>>> And I repeat:
>>>
>>> You're a goddamned idiot.
>>
>>
>>So when, despite being a slow learner, you do learn something,
>>it is invariably wrong.
>
> You're an idiot. You statement about Linux is absolutely 100%
> WRONG!
>

It's a matter of principle not of pragmatism.

Why pollute a linux machine with some other operating system...

Bye.
Jasen
From: jasen on
On 2007-02-16, Ken Smith <kensmith(a)green.rahul.net> wrote:
> In article <be273$45d50d69$49ecf9d$20196(a)DIALUPUSA.NET>,
> nonsense(a)unsettled.com <nonsense(a)unsettled.com> wrote:
> [.....]
>>None of the has anything to do with the OS biz.
>
>
> We just had another wonderful experience with XP. Characters pumped into
> the serial port may take up to 5 seconds before a DOS application running
> under XP gets to see them. Most of them eventually come through.
>
> Tomorrow, we may try it with "dosemu" to see how well that works.

with the right serial driver it should work well, people were running serial
games in dosemu, dunno all the tricks they employed.

Bye.
Jasen
From: jmfbahciv on
In article <er440c$p85$1(a)jasen.is-a-geek.org>,
jasen <jasen(a)free.net.nz> wrote:
>On 2007-02-13, MassiveProng <MassiveProng(a)thebarattheendoftheuniverse.org>
wrote:
>
>>>>>I'll repeat it since you're a slow learner:
>>>>>
>>>>>A properly installed Linux uses all the available
>>>>>partitions.
>>>>
>>>> And I repeat:
>>>>
>>>> You're a goddamned idiot.
>>>
>>>
>>>So when, despite being a slow learner, you do learn something,
>>>it is invariably wrong.
>>
>> You're an idiot. You statement about Linux is absolutely 100%
>> WRONG!
>>
>
>It's a matter of principle not of pragmatism.
>
>Why pollute a linux machine with some other operating system...

The people who pay you use an OS that you don't.

/BAH
From: Ken Smith on
In article <er45sh$pkf$2(a)jasen.is-a-geek.org>,
jasen <jasen(a)free.net.nz> wrote:
>On 2007-02-12, Ken Smith <kensmith(a)green.rahul.net> wrote:
>>>> You had to load the CX register to do a LOOP.
>>>
>>>What's LOOP got to do with anything?
>>
>> You haven't been following the discussion. I used the case of a REP
>> prefix used inside a loop as the example of why the 8086's instruction set
>> was so poorly designed. The CX and the REP both use the CX so the CX must
>> be loaded for the REP inside the loop. This means that the current CX
>> contents must be saved, the CX loaded, the REP done and the CX restored.
>> This is a lot of extra work.
>
>It's less work to just use something else to hold the loop count
>and use DEC and JNZ in place of LOOP (faster too since 80486)

It is less work but you are still paying some speed penalty. Remember we
are talking of where this started on the 8086. It was the first member of
the family.

>
>One of the other registers or [bp+N] is often a good choice
>(for apropriate values of N)

Using a memory location adds another delay. Moving values to and from
memory required that the address take a trip trhough the ALU because of
the silly segmentation idea.


>
>Bye.
> Jasen


--
--
kensmith(a)rahul.net forging knowledge