A low-power 8-bit ALU can reduce the energy demands of digital processing by rethinking the circuits used for arithmetic and logic selection. One proposed design combines an 11-transistor full adder with Gate Diffusion Input (GDI) multiplexers and operates at a reduced supply voltage. Its reported simulations illustrate a familiar design trade-off: lowering voltage can reduce power consumption, but it can also increase propagation delay.
For automation hardware designers working on embedded controllers or other resource-constrained electronics, the design offers a useful case study in circuit-level power optimization. The results are simulation-based and tied to the stated process and test conditions, so they should be treated as a design reference rather than a universal performance guarantee.
Why ALU power matters
An arithmetic logic unit (ALU) is a combinational logic block that performs arithmetic and Boolean operations on digital data. In a processor or controller, it may support functions such as addition, subtraction, incrementing, decrementing, AND, OR, XOR, and XNOR. The ALU’s internal switching activity and circuit structure contribute to the power consumed by the wider digital system.
In CMOS circuits, total power has dynamic, short-circuit, and static components. Dynamic power is associated with charging and discharging capacitive loads as signals switch. Short-circuit power occurs during transitions when current flows between supply rails, while static power is associated with leakage. Reducing supply voltage is one way to lower power, but it requires checking whether the resulting circuit speed remains suitable for its intended use.
Building a low-power 8-bit ALU
The proposed architecture uses an 8-bit ripple-carry adder for arithmetic operations. Each full-adder stage processes operand bits and passes its carry output to the next stage. This structure is straightforward to build, although carry propagation across stages contributes to the overall delay.
At the bit level, the design combines a full adder with multiplexers. The full adder supplies arithmetic results, while mux logic selects among operations and data paths according to the control inputs. Cascading eight full-adder stages forms the arithmetic path; the logic selection is implemented using 2:1 and 4:1 multiplexers.
The operation is selected using three control signals, S2, S1, and S0. The documented function table assigns these combinations to eight functions:
- 000: Operand1 OR Operand2
- 001: XNOR of the operands
- 010: XOR of the operands
- 011: AND of the operands
- 100: increment
- 101: addition
- 110: decrement
- 111: subtraction
How the 11-transistor full adder and GDI muxes contribute
The 11-transistor full adder is the arithmetic building block. The design describes it as using low-power XOR logic and a 2:1 multiplexer to produce sum and carry outputs. An additional NMOS transistor is used in the stated subthreshold operating approach. Subthreshold current refers to the small current that flows when a transistor’s gate-to-source voltage is below its threshold voltage. Using this operating region can support lower-power operation, but circuit behavior and speed must be evaluated under the chosen conditions.
For data selection, the design uses GDI-based multiplexers. A 2:1 mux selects between two inputs based on a select signal; a 4:1 mux can be assembled from three 2:1 muxes. In the proposed ALU, these mux structures direct the selected operands and operation results through the datapath.
GDI is presented as a means of implementing logic functions with fewer transistors than conventional approaches in some circuit structures. The paper also notes that the technique depends on implementation considerations such as twin-well or silicon-on-insulator technology. As a result, GDI’s practical suitability should be assessed against the target fabrication process and circuit requirements.
Simulated power and delay trade-offs
The design was simulated using Tanner EDA tools with a 32 nm BSIM4 process model. The reported ALU measurements show power decreasing as supply voltage falls, while delay rises:
- 0.9 V: 39.23 µW power and 1.42 ns delay
- 0.8 V: 25.48 µW power and 1.49 ns delay
- 0.7 V: 16.34 µW power and 1.63 ns delay
- 0.6 V: 10.12 µW power and 1.92 ns delay
- 0.5 V: 6.16 µW power and 4.66 ns delay
The authors selected 0.7 V as a compromise between power and delay. At that voltage, the reported proposed-design results list 16 µW for the 8-bit ALU and a 2 ns delay. The paper also reports power below that of a conventional CMOS 8-bit ALU, stating a reduction of less than 82%. Such comparisons depend on the reference design and simulation conditions; they should not be generalized without matching the process, voltage, workload, and measurement methodology.
What engineers should consider
For engineers assessing a low-power ALU architecture, the central lesson is that voltage selection is a system-level decision as much as a circuit-level one. A lower supply can reduce simulated power, but the associated increase in delay may affect timing margins or the operating point available to downstream logic.
Before adopting a similar approach, design teams should verify the implementation in the intended process and examine timing across operating conditions. They should also evaluate the ALU as part of the complete datapath, including mux delays and ripple-carry propagation, rather than relying only on an isolated full-adder result. The reported figures describe a specific simulated design, not a measured industrial controller or a guaranteed production outcome.
Conclusion
The proposed 8-bit ALU illustrates how an 11-transistor full adder, GDI-based multiplexers, and reduced-voltage operation can be combined to target lower power consumption. Its simulations show a clear power-delay trade-off as supply voltage changes, with 0.7 V selected as a balance for the reported design. For automation electronics teams, the practical takeaway is to evaluate low-power logic using the timing, process, and operating requirements of the final application—and to validate the complete datapath before drawing conclusions about system-level benefits.