PDN checks in VLSI

 There are multiple checks need to be do for proper PDN signoff.

Which are ,

1. Gridcheck Analysis

2. Static IR Drop Analysis

3. Dynamic IR Drop Analysis

Setup fixing , hold fixing , noise fixing

Setup Fixing:

1. Cell swaping from Hvt  LVT

2. Upsize the bottleneck drivers(preffered buffers and inverters)

3. Split the fanout

4. Clone the critical cell

5. Clock_skewing (either launch or capture)

Hold Fixing:

1. Cell swpping to HVT in data path

2. Inserting buffer @ endpoint if setup margin is there

3. If endpoint cell area is high density route the nets in lower layers

Noise/Double Switching Fixing:

1. Applying NDR to the nets (so that space will be increase)

2. Sheilding

3. Upsizing the victim Net Driver

some useful links

http://www.asic-world.com/

http://vlsi.pro/

http://testbench.in/

http://www.vlsi-expert.com

Congestion reduction techniques

Congestion Reduction Techniques

This post will give the idea of what is congestion and what are the reasons for congestion how to resolve the congestion problem.  

Congestion  : if the number of available routing tracks is less than the number of required routing tracks

Before going to reduction techniques will first know the reasons for it.

Reasons

• High standard cell density in small areas
•  High standard cells near macro edges
•  High pin density areas (M2 M3 routing congestion) due to AOI, OAI, flops
• Rectilinear block shape
•  At last the bad floorplan or height or width of the block.

Reduction techniques:

First analyze the root cause for congestion before setting the high effort  options to reduce congestion . ( if we enable such options run time will be high and timing may be degraded)

• After seeing the congestion maps check the congested area having high standard cell density. If you see such case than apply the partial blockages with less than the density in that particular area.

          Eg: if the congestion regions is let say 80% utilization then apply partial blockage with 60% like

•  If you see the congestion near to the macro area then apply 5-10 microns of soft or hard blockage (better is 5 microns hard and 5 is soft is better). Because high critical timing path cell will be seated in the soft blockage area in the optimization phase.

• If you see the congestion M2 M3 layers apply cell padding to the high number of pin standard cells (Eg. AOI, OAI,Full adders or other 4 i/p gates   )

• Congestion is at rectileanior areas apply the partial blockages

Reasoning Interview Questions

Q) Suppose you have a 3 liter jug and a 5 liter jug. The jugs have no measurement lines on them either. How could you measure exactly 4 liter using only those jugs and as much extra water as you need?

ANS :

Step 1 : First, fill the 5 liter jug and then pour it into the 3 liter jug. The 5 liter jug now has only 2 liters left.

Step 2 : Next, empty out the 3 liter jug. Then, pour the 2 liters from the 5 liter jug to the 3 liter jug. So, now the 3 liter jug has 2 liters.

Step 3 : Fill the 5 liter jug again, and pour 1 liter into the 3 liter jug. Now, what’s left in the 5 liter jug? Well, exactly 4 liters! There’s your answer.

Q) There are 3 baskets labeled ‘Apples’, ‘Oranges’ & ‘Mixture’. One of them contains only Apples, one only Oranges and  one has mix of apples and oranges both.

These baskets are not labeled correctly. In fact, the labels on these baskets always lie. (i.e. if the label says Oranges, Then you are sure the basket either has only Apples or Mixture).

You are allowed to pick one fruit from one basket (Not allowed to see other fruits), and you have to put all the labels correctly on the basis of that information (by seeing only one fruit from any one basket).

How will you do that ?

ANS:

Pick a fruit from the basket labeled ‘Mixture’. We know from the question that this basket does not contain ‘Mixture’ for sure.

1. If this fruit is an apple, then label this Basket as ‘Apple’ (Because this basket does not contain Mixture, so if one is apple, all are apples only). Now we’ve determined that the basket labeled as ‘Mixture’ only contains Apples.
2. If we look at the basket labeled as ‘Oranges’, we know that since the label is incorrect, this basket either has only apples in it or has Mixture. Since we already know which basket contains only apples, we know that the basket labeled as ‘Oranges’ contains ‘Mixture’. So label it as ‘Mixture’. The 3rd basket will be labeled as ‘Oranges’.

You can apply the same logic if you assume you initially picked an orange from the basket labeled as ‘Mixture’

Q) You have two sticks and matchbox. Each stick takes exactly an hour to burn from one end to the other. The sticks are not identical and do not burn at a constant rate. As a result, two equal lengths of the stick would not necessarily burn in the same amount of time.  How would you measure exactly 45 minutes by burning these sticks?

ANS: This puzzle used to be asked in Wall Street interviews long time ago. It is very rare for this question to be asked now but it is a very good question to help you think a little outside the normal thought process.

The answer is really simple. Since the sticks do not burn at a constant rate, we can not use the length of the stick as any sort of measurement of time. If we light a stick, it takes 60 minutes to burn completely. What if we light the stick from both sides? It will take exactly half the original time, i.e. 30 minutes to burn completely.

0 minutes – Light stick 1 on both sides and stick 2 on one side.
30 minutes – Stick 1 will be burnt out. Light the other end of stick 2.
45 minutes – Stick 2 will be burnt out.

For more Questions below are the references

http://www.programmerinterview.com/index.php/puzzles/measure-4-liters/

http://www.mytechinterviews.com/9-minutes

http://www.ritambhara.in/3-basket-puzzle-appleorange-puzzle/