Prev: 32 bit microcontroller with external memory bus needed
Next: Opinions wanted on career-limiting moves (<g>)
From: Jim Stewart on 14 Apr 2010 15:09
d_s_klein wrote:
> On Apr 13, 10:21 pm, whygee <y...(a)yg.yg> wrote:
>> satish.aradhya wrote:
>>> HI,
>>> I have the following queries based on the Flash Memory,
>>> 000. What is the advantage of Top Boot Flash Memory and
>>> What is the advantage of Bottom Boot Flash memory ?
>> the Top Boot is faster.
>> The bits fall from higher.
>>
>
> Wrong, wrong, wrong. (Not true.) The bits from the top-boot have to
> filter down through all of the non-boot blocks, which takes -longer-.

That's why I always invert the address bits.

> Especially if the non-boot blocks are programmed - 1 bits have a lower
> co-efficient of friction than 0 bits do.
>
> The real problem with bottom-boot is that if you do a lot of
> debugging, all the dead bugs tend to collect there and clutter things
> up.
>
> RK.
From: whygee on 14 Apr 2010 14:51
d_s_klein wrote:
> On Apr 13, 10:21 pm, whygee <y...(a)yg.yg> wrote:
>> satish.aradhya wrote:
>>> HI,
>>> I have the following queries based on the Flash Memory,
>>> 000. What is the advantage of Top Boot Flash Memory and
>>> What is the advantage of Bottom Boot Flash memory ?
>> the Top Boot is faster.
>> The bits fall from higher.
>
> Wrong, wrong, wrong. (Not true.) The bits from the top-boot have to
> filter down through all of the non-boot blocks, which takes -longer-.
>
> Especially if the non-boot blocks are programmed - 1 bits have a lower
> co-efficient of friction than 0 bits do.
>
> The real problem with bottom-boot is that if you do a lot of
> debugging, all the dead bugs tend to collect there and clutter things
> up.

damn right. I completely forgot those implementation details.
That's why I prefer to use magnetic bubble memory.

> RK.
yg

--
http://ygdes.com / http://yasep.org
From: whygee on 14 Apr 2010 14:54
Jim Stewart wrote:
>> Wrong, wrong, wrong. (Not true.) The bits from the top-boot have to
>> filter down through all of the non-boot blocks, which takes -longer-.
> That's why I always invert the address bits.
oh, yes, just flip the MSB.
the price of a single-gate inverter like SN74LVC1G04 is not so high today.
But isn't the inverter gate's delay over-compensating the faster access ?

yg
--
http://ygdes.com / http://yasep.org
From: Jim Stewart on 14 Apr 2010 17:45
whygee wrote:
> Jim Stewart wrote:
>>> Wrong, wrong, wrong. (Not true.) The bits from the top-boot have to
>>> filter down through all of the non-boot blocks, which takes -longer-.
>> That's why I always invert the address bits.
> oh, yes, just flip the MSB.
> the price of a single-gate inverter like SN74LVC1G04 is not so high today.
> But isn't the inverter gate's delay over-compensating the faster access ?

Not if you get the inversion by soldering the memory
chip upside-down.
From: whygee on 14 Apr 2010 17:26
Jim Stewart wrote:
> whygee wrote:
>> But isn't the inverter gate's delay over-compensating the faster access ?
> Not if you get the inversion by soldering the memory chip upside-down.
but then you will get all your data negated !
so this displaces the inverter gates from
the address bus to the data bus, with no speed gain.

yg
--
http://ygdes.com / http://yasep.org
First  |  Prev  | 
Pages: 1 2
Prev: 32 bit microcontroller with external memory bus needed
Next: Opinions wanted on career-limiting moves (<g>)