fifo with ram or flops

What is the impact of using ram vs flops for a fifo? I know ram can be smaller and so uses less power. What is the downside of using ram here?

RAM is usually architected so you are either reading or writing
but not both; a FIFO is often desired to allow simultaneous or
non-aligned (so potentially simultaneous) write and read.

Flip-flops are huge compared to a well crafted SRAM cell layout.
They have a lot of extra transistors meaning they waste power
when active, and flops sized for general use have unnecessarily
large devices.

Either way, you need a read index and write index counters,
some overflow / underflow logic and so on.

A latch is probably more sensible than a true flip-flop for
this application, which would cut cell size by roughly half.
But still far from as efficient as a SRAM cell.

so if we have rams that can do read/write simultaneously, we should always use rams for fifo? Any other cons of using rams over fifo?

The rams may be slower. If you have small fifo, the sram may be bigger, than FFs. It depends...

stanford said:

so if we have rams that can do read/write simultaneously, we should always use rams for fifo? Any other cons of using rams over fifo?

Click to expand...

RAMs are typically from third parties and you might have to pay to get access to good ones. that is a downside, more on the commercial/feasibility side of things.

Let's get into context. For memory, we have a few solutions at hand to fulfill our need: SRAM cells, latches, or registers.

Basically, latches and registers are made up of multiple digital logics. Because they are CMOS structure, their gate output Vout is independent of input Vgs, The output voltage can essentially drive loads at about Vdd voltage (pros for registers).

SRAMs are essentially 2 or 4 cells "holding" the voltage. They remember the signal by a feedback mechanism, as shown by the below image.
Because they are build with only a handful of transistors, SRAMs are more compact (pros for SRAM). This is in contrast of registers that are made of multiple logic gates (vastly more transistors).
Consisting only a handful of transistor, SRAMs are less power hungry and dissipates the least amount of power (pros of SRAM).


https://upload.wikimedia.org/wikipedia/commons/5/56/4T_SRAM_Cell.png

However, notice that SRAM output bitline has to go thru a level of resistance through resistor R and mosfet M1. There is also no PMOS+NMOS (CMOS) complementary benefit for M1 connected in series to the memory cell M2. Therefore, your SRAM bitline switches more slowly (cons of SRAM), and cannot drive bigger load, as the Vbitline = Vdd - (VR + Vm2 (cons of SRAM).
If Vbitline drops to a certain threshold, your memory cell cannot function. This is why SRAM usually operates at a voltage higher than your standard digital registers (cons of SRAM).
To help reduce voltage drop at R and M1, and to improve switching speed of bitline, SRAMs are designed in a custom manner with analog mathematics in consideration (cons of SRAM) instead of being digitally guaranteed signalling. Registers on the other hand have a guaranteed digital output and can be placed and routed in an automated fashion (pros of registers).
However, when designed in a custom manner, engineers can further increase the density of the transistors (pros of SRAM). Because of the heavy customization of the transistors, most companies find it better to just "buy" the SRAM (cons of SRAM) designs from reputable companies that have done enough research on the transistor node and with the best/densest SRAM implementation. So when the companies buys the SRAMs, they will usually buy the SRAM that has a big footprint, but with the densest amount of memory cells.

On the other hand, register banks are something that is pretty standardized based on your fab, and can be digitally placed and routed anywhere in you silicon (pros of registers), compared to bought SRAMs that come in "chunks".

To recap (but not exhaustively listed):

All in all, silicon area cost and power consumption trumps all others, hence for architectures where memory banks are less referenced (low read/write op, non-speed critical), SRAMs are generally preferred.

The size and "depth of the dive" probably define the choice for a
given design.

If you're working a synthesis-based flow and there is no synthesizable
SRAM block (with, critically, simultaneous read and write at two
different addresses, "dual port", add your own counters and over /
underflow logic) then registers (latches, although many synthesis
flows and their Methodology Harpies forbid non-synchronous storage
elements) may be the only way through the design process that
does not involve hostage negotiations and eye-gouging.

A digital design group manager is unlikely to support your impulse
to go off in the bushes and design your own RAM cells, read and
write and decode circuitry, and bet on you not screwing up one
bit of it. Especially if the questions are so basic.

But this is probably more like a homework assignment that has
already passed?

At any rate, such decisions are not always decided by the
purely technical merits, but by things like schedule, D/L cost,
risk and "smooth sailing past the Scylla of tools-chain and
the Charybdis of in-house design methodologies, which no
man may question".

arvinraj said:

Let's get into context. For memory, we have a few solutions at hand to fulfill our need: SRAM cells, latches, or registers.

Basically, latches and registers are made up of multiple digital logics. Because they are CMOS structure, their gate output Vout is independent of input Vgs, The output voltage can essentially drive loads at about Vdd voltage (pros for registers).

SRAMs are essentially 2 or 4 cells "holding" the voltage. They remember the signal by a feedback mechanism, as shown by the below image.
Because they are build with only a handful of transistors, SRAMs are more compact (pros for SRAM). This is in contrast of registers that are made of multiple logic gates (vastly more transistors).
Consisting only a handful of transistor, SRAMs are less power hungry and dissipates the least amount of power (pros of SRAM).

View attachment 150516
https://upload.wikimedia.org/wikipedia/commons/5/56/4T_SRAM_Cell.png

However, notice that SRAM output bitline has to go thru a level of resistance through resistor R and mosfet M1. There is also no PMOS+NMOS (CMOS) complementary benefit for M1 connected in series to the memory cell M2. Therefore, your SRAM bitline switches more slowly (cons of SRAM), and cannot drive bigger load, as the Vbitline = Vdd - (VR + Vm2 (cons of SRAM).
If Vbitline drops to a certain threshold, your memory cell cannot function. This is why SRAM usually operates at a voltage higher than your standard digital registers (cons of SRAM).
To help reduce voltage drop at R and M1, and to improve switching speed of bitline, SRAMs are designed in a custom manner with analog mathematics in consideration (cons of SRAM) instead of being digitally guaranteed signalling. Registers on the other hand have a guaranteed digital output and can be placed and routed in an automated fashion (pros of registers).
However, when designed in a custom manner, engineers can further increase the density of the transistors (pros of SRAM). Because of the heavy customization of the transistors, most companies find it better to just "buy" the SRAM (cons of SRAM) designs from reputable companies that have done enough research on the transistor node and with the best/densest SRAM implementation. So when the companies buys the SRAMs, they will usually buy the SRAM that has a big footprint, but with the densest amount of memory cells.

On the other hand, register banks are something that is pretty standardized based on your fab, and can be digitally placed and routed anywhere in you silicon (pros of registers), compared to bought SRAMs that come in "chunks".

To recap (but not exhaustively listed):

Large	Small
poor	high
High	Low
Any voltage above Vth, can work with as low as 0.6V	Significantly higher than your nominal core voltage, typically 1.2-1.5V (+100% Vth)
Not needed	Requires level shifters for cross voltage signals due to the above point, and repeaters
Easy and automated	Custom and difficult
Cheap, can be designed internally	Expensive and requires royalty for every SRAM block used
Low, guaranteed by digital design, easy fix upon failure	High due to possibility to enter into metastable state, or write failure by picking up latchup event from radiation
Flexible	Restrictive, and comes in chunks. You need to further isolate the silicon area used with a buffer region for SRAM due to the higher voltage
Fast	Slow

[td]Registers[/td]
[td]Pros/Cons[/td]
[td]SRAMs[/td] [td]Silicon area cost[/td]
[td]Memory Density[/td]
[td]Power Consumption[/td]
[td]Operating Voltage[/td]
[td]Requirement for Supplementary Logics[/td]
[td]Design Complexity[/td]
[td]Design Cost[/td]
[td]Potential for failure/metastability[/td]
[td]Placement in silicon[/td]
[td]Responsiveness[/td]

All in all, silicon area cost and power consumption trumps all others, hence for architectures where memory banks are less referenced (low read/write op, non-speed critical), SRAMs are generally preferred.

Click to expand...

Thanks for the table. On some design, I faced it used some small srams for the fifo from the output of Serdes , still wonder why .