On Apr 13, 11:01*pm, "RealInfo" <
[email protected]> wrote:
> Hi all
>
> I am new to VHDL and I am looking for an *expertise direction.
>
> Lately I came to conclusion that I want to focus on memories modeling from
> all kinds , SRAM , ROM, FLASH etc.
>
> My question is : Is it realistic to focus on memories modeling , is therea
> market for such expertise.
>
> I ask this because there are many free online models of memories but in the
> other hand , there are many many memories and memories stacks
> that need to be modeled for new *ASIC , MICRO and FPGA based systems.
>
> Thanks in advance
>
> EC
Having come from the memory industry (Lexar/Micron), I can say that
there are a few areas of interest - which can also be valuable outside
the memory world. Memory is a commodity product, it will always be
needed - but it won't be as "glamorous" as the latest AMD/Intel chip.
The first is error detection/correction. It is amazing how few people
actually understand this topic properly when determining how many bits
it requires for complete failure.
The second is interface. Memory is all about how fast can you get
data without corruption (there are few applications, that require long
term persistence over speed - Library of Congress being one). I put
in here the decision of how large a block of memory you need to
address at a time (2Kb vs 512b makes a big difference in performance
when updating a FAT)
These two are fairly universal in most projects, IMHO. It all boils
down to what is the best way to get data? (Pick your definition of
best)
Memory tends to be driven by how people want to access it. The end
result almost never is a perfect fit for any market, but very darn
close to several high volume markets.
Of course this was based on my interests, there are other specialties
within the memory world.