A wire load model (WLM) is a table in a timing library that estimates the length, capacitance, resistance and area of a net from its fanout, before any placement or routing exists. Synthesis and pre-layout static timing analysis use these estimates to calculate interconnect delay. Wire load models are statistical, so they are only approximations; at advanced process nodes they have largely been replaced by physically aware synthesis.
Key takeaways
- A wire load model maps fanout to an estimated wire length, then multiplies that length by per-unit-length resistance, capacitance and area.
- For fanouts beyond the end of the table, the length is extrapolated with the slope value.
- Library vendors supply several models, usually chosen by the area of the design or block.
- The wire load mode (top, enclosed or segmented) decides which model applies to a net that crosses hierarchy.
- After place and route, extracted parasitics (for example a SPEF file) replace wire load estimates.
What a wire load model contains
Wire loading models contain all the information required by compile to estimate interconnect wiring delays. A typical wire load model definition contains: area, resistance, capacitance, slope and fanout. All these attributes are given per unit length wire. Slope value is used to characterize linear fanout.
To be precise about the attributes:
| Attribute | What it gives |
|---|---|
| resistance | Resistance per unit length of wire |
| capacitance | Capacitance per unit length of wire |
| area | Routing area per unit length of wire |
| fanout_length | A table of pairs: fanout count and the estimated wire length for that fanout |
| slope | Extra length per extra fanout, used to extrapolate beyond the last table entry |
A wire load model in Liberty format
Wire load models are written in the Liberty (.lib) library format. The example below uses illustrative numbers; real values come from the library vendor and depend on the library's units.
wire_load("wlm_10k") {
resistance : 0.0012 ;
capacitance : 0.00015 ;
area : 0.07 ;
slope : 45.0 ;
fanout_length(1, 22.0) ;
fanout_length(2, 40.0) ;
fanout_length(3, 56.0) ;
fanout_length(4, 70.0) ;
}
Why wire load models are used
Generally wire load models are used in ASIC design. These wire load models will contain statistical values which are used in pre-layout simulation of ASIC. Since we are extracting resistance (R) and capacitance (C) values in the back end after the place and route (P&R) phase, we need to perform pre-layout simulation before P&R.
A wire load model attempts to predict the capacitance and resistance of nets in the absence of placement and routing information. The estimated net capacitance and resistance are used for delay calculation. Technology library vendors supply statistical wire load models to support estimation of wire loads based on the number of fanout pins on a net. You can set wire load models manually or automatically.
Worked example: estimating a net from its fanout
Using the model above, with capacitance in pF per unit length and resistance in kilo-ohms per unit length (assumed units for the example):
Net with fanout 3 (inside the table)
- Length from the table: 56.0 units.
- Capacitance: 56.0 x 0.00015 = 0.0084 pF (8.4 fF).
- Resistance: 56.0 x 0.0012 = 0.0672 kilo-ohm (67.2 ohm).
- Area: 56.0 x 0.07 = 3.92 area units.
Net with fanout 6 (beyond the table)
- The last entry is fanout 4 at length 70.0. Two extra fanouts at a slope of 45.0 add 90.0.
- Length: 70.0 + (6 - 4) x 45.0 = 160.0 units.
- Capacitance: 160.0 x 0.00015 = 0.024 pF (24 fF).
- Resistance: 160.0 x 0.0012 = 0.192 kilo-ohm (192 ohm).
The tool adds the wire capacitance to the input pin capacitances of the fanout cells to get the total load on the driver, then looks up the driver's delay and output slew in its delay tables. How the wire resistance is distributed among the loads depends on the tree type set in the library's operating conditions (best case, balanced or worst case).
Setting wire load models manually or automatically
- Manually: the designer names a model for the design or for a block, for example with
set_wire_load_model -name wlm_10kin Synopsys Design Compiler. - Automatically: the library holds a wire load selection table that picks a model from the cell area of the design or block. Larger blocks get models with longer wires per fanout.
- Library default: if nothing is set, the tool uses the default wire load named in the library.
Wire load modes for hierarchical designs
| Mode | Model used for a net that crosses hierarchy |
|---|---|
| top | The model of the top-level design, for every net |
| enclosed | The model of the smallest block that fully contains the net |
| segmented | Each segment of the net uses the model of the block it lies in |
The mode is set with set_wire_load_mode. Top mode is the most pessimistic; segmented is the most optimistic.
Limits of wire load models today
Note added November 2026: wire load models date from process nodes where gate delay dominated and wires were short. At deep submicron nodes, interconnect delay is a large part of every path and depends on where cells are placed, not just on fanout. Two nets with the same fanout can differ in length by a large factor. For that reason most flows at modern nodes use physically aware synthesis, which estimates wires from a placement, and many libraries for those nodes ship with zero or minimal wire load models. Wire load models are still useful for quick estimates, small blocks and mature nodes, and the concept remains a common interview question.
Wire load model versus extracted parasitics
| Wire load model | Extracted parasitics (SPEF) | |
|---|---|---|
| When available | Before placement | After routing |
| Basis | Statistics over many designs, by fanout | Actual geometry of each routed net |
| Accuracy | Approximate | Accurate to the extraction corner |
| Used for | Synthesis and early STA | Sign-off STA |
Common mistakes to avoid
- Trusting pre-layout slack too much. A path with small positive slack under a wire load model can fail after routing.
- Using a model that is too small for the block. Wires are estimated too short and timing looks better than it will be.
- Leaving a wire load model active after loading SPEF. Make sure the sign-off run uses the extracted values for every net.
Related reading
- Slack in static timing analysis
- Setup time and hold time, with fixes for load and slew violations
- Inverters and buffers in STA and timing closure
- RTL synthesis and back-end interview questions with answers
- The Semiconductor Hub glossary: extraction corner and timing closure
Frequently asked questions
What is a wire load model in VLSI?
It is a statistical table in the timing library that estimates a net's length, resistance, capacitance and area from its fanout, so that delays can be calculated before layout.
What is the slope in a wire load model?
The extra wire length added for each fanout beyond the largest fanout listed in the fanout_length table. It lets the tool extrapolate linearly for high-fanout nets.
What is the difference between top, enclosed and segmented wire load modes?
They decide which model a net crossing hierarchy boundaries uses: the top design's model, the model of the smallest block enclosing the net, or a separate model for each segment.
Are wire load models still used?
Less than before. At advanced nodes, physically aware synthesis gives better wire estimates, but wire load models are still found in libraries for mature nodes and in quick early runs.
What replaces a wire load model after layout?
Parasitics extracted from the routed layout, usually read into STA as a SPEF file.


Very informative article.Thank you author for posting this kind of article .
ReplyDeletehttp://www.wikitechy.com/view-article/capacitance-in-digital-electronics-with-example-and-explanation
Both are really good,
Cheers,
Venkat