What is OCV (on-chip variation)?

Variation in transistor and wire delay from place to place within a single chip. STA applies separate factors to launch and capture paths (for example 1.1× and 0.9×) to analyze the most pessimistic combination.

Japanese name: OCV(オンチップばらつき)

OCV (on-chip variation) appears in 3 practice questions across 2 topics, including Timing analysis and CDC, IC design flow and EDA.

Field: Timing analysis and CDC

Try it: questions about OCV (on-chip variation)

For OCV derating in setup analysis, the launch clock path and data path delays are multiplied by 1.1 and the capture clock path delay by 0.9. With a launch clock path of 1.0 ns, a data path (including clock-to-output) of 1.5 ns, a capture clock path of 1.0 ns (no common portion), a period of 3.0 ns and a setup time of 0.1 ns, what is the setup slack?

  1. A1.05 ns
  2. B1.40 ns
  3. C1.25 ns
Show answer and explanation

Answer: A. 1.05 ns

OCV derating accounts for delay varying with location even within the same chip, and for setup assumes the most pessimistic combination (slow launch side, early capture clock). The arrival time is 1.1 × (1.0 + 1.5) = 2.75 ns and the required time is 3.0 + 0.9 × 1.0 − 0.1 = 3.8 ns, so the slack is 1.05 ns. 1.40 ns is the value without derating, and 1.25 ns applies the factor to the data path only. In hold analysis the factors are applied the other way around.

In STA with OCV derating, when the launch and capture clock paths share a common portion, which correctly describes CRPR (common path pessimism removal)?

  1. AIt treats the delay of the common portion as zero and recalculates the clock arrival times
  2. BIt removes the impossible extra pessimism caused by applying separate slow and fast values to the common portion
  3. CIt applies an even larger derating factor to the common portion to stay on the safe side
Show answer and explanation

Answer: B. It removes the impossible extra pessimism caused by applying separate slow and fast values to the common portion

OCV estimates the launch clock path as slow and the capture clock path as fast, but the common portion from the root to the branch point is physically the same buffers and wires, so it cannot be both slow and fast within a single edge. If the common portion has a delay of 1.0 ns and the factors are 1.1 and 0.9, the 0.2 ns difference between 1.1 ns and 0.9 ns is extra pessimism. CRPR adds this difference back to the slack, bringing the result closer to reality. It does not set the common delay to zero, and enlarging the factor is the opposite operation.

OCV (on-chip variation) in the explanations

Questions that mention OCV (on-chip variation) (3)

  1. For OCV derating in setup analysis, the launch clock path and data path delays are multiplied by 1.1 and the capture clock path delay by 0.9. With a launch clock path of 1.0 ns, a data path (including clock-to-output) of 1.5 ns, a capture clock path of 1.0 ns (no common portion), a period of 3.0 ns and a setup time of 0.1 ns, what is the setup slack?
  2. In STA with OCV derating, when the launch and capture clock paths share a common portion, which correctly describes CRPR (common path pessimism removal)?
  3. In static timing analysis, what is the technique called that multiplies delays by factors to add margin for transistor and wire delays varying by location even within the same chip?

Books on Timing analysis and CDC Ad

As an Amazon Associate I earn from qualifying purchases.

Related terms

Derating Slack STA (static timing analysis) Setup time CRPR (clock reconvergence pessimism removal) Threshold voltage Asynchronous FIFO Clock domain Clock uncertainty Critical path

Back to the glossary