Reference stack

Many of the above transformations are weU-documented in texts on compiler design. The reference stack and combinational logic coalescing techniques are new. They are described in the next two sections. 

The reference stack algorithm efficiently resolves constant and variable propagation, conditional assignment to variables, and multiple assignment to variables. By using the reference stack, many of the control steps that are traditionally needed to implement assignment to variables can be eliminated. This leads to an improvement in the latency of the resulting hardware. 

For the sake of illustration, consider the following code segment. The first column shows the HardwareC code. Temporary variables are inttoduced by the parser to hold the results of any binary or unary operation, denoted by T1 and T2 in the example below. The second column is the first column augmented with temporary variables. The third column shows the code segment after variable references are replaced by the top of reference stack. The underlined entries highlight the references that have been changed. 

Whenever an assignment is made to a variable, subsequent references to that variable results in referencing the value and not the variable itself. In its simplest form, the reference stack algorithm performs constant and variable propagation. 

Variables in HardwareC are categorized into local variables (which are in addition categorized into static, boolean and int variables) and external ports. Whereas local variables are not externally visible, ports connect external signals with computations within the model. Only explicit write commands can change 

---

Reference stack resolution, where multiple and conditional assignments to variables are resolved and eliminated by creating multiplexed values that can subsequently be referenced. 

We present in the remaind of this section the details of the reference stack algorithm. Let the reference stack for a variable v be denoted by RS(v). RS(v)... 

Let M node) denote modified variables in any conditional branch. For each variable v M node), a new entry is pushed onto the reference stack RS v) for each different case of the conditional. For example, if c is an 1-bit Boolean variable ( c = 1), then two new entries are added to the top of the stack one corresponds to the case c = 0 and the otho corresponds to the case c = 1. Likewise, if c is a 3-bit Boolean variable ( c = 3), then eight (2 ) new entries are added each entry corresponds to one of the eight possible values of c. New stack entries are initialized to the ivevious top of stack value. 

The algorithm is recursively applied to each of the conditional branches. For a branch j, the top of the reference stack refers to the newly created stack entry corresponding to the case (c = j). Assignments and references in a branch access and modify the appropriate stack entry only. Once all the branches have been traversed, the algorithm pops off these 2l l stack entries and examines their values. 

An example of the algorithm is shown in Figure 3.2. The first and second columns represent the code segment before and after application of the algorithm, respectively. The reference stack for variable a is shown, which is updated whenever an assignment to a occurs. In the case of the conditional on cond, two entries are pushed onto the top of stack. These stack enuies are initialized... 

Figure 3.2 Example of the reference stack algorithm applied to a code segment The reference stack for variable a is shown.

Semi-submersible rigs are often referred to as semis , and are a floating type of rig. Like the jack-up, a semi is self contained. The structure is supported by large pontoons which are ballasted with water to provide the required stability and height. The rig is held in position by anchors and mooring lines or dynamically positioned by thrusters. A large diameter steel pipe ( riser ) is connected to the sea-bed and serves as a conduit for the drill string. The blowout preventer (BOP) is also located at the sea-bed ( sub sea stack ). 

Selection of pollution control methods is generally based on the need to control ambient air quaUty in order to achieve compliance with standards for critetia pollutants, or, in the case of nonregulated contaminants, to protect human health and vegetation. There are three elements to a pollution problem a source, a receptor affected by the pollutants, and the transport of pollutants from source to receptor. Modification or elimination of any one of these elements can change the nature of a pollution problem. For instance, tall stacks which disperse effluent modify the transport of pollutants and can thus reduce nearby SO2 deposition from sulfur-containing fossil fuel combustion. Although better dispersion aloft can solve a local problem, if done from numerous sources it can unfortunately cause a regional one, such as the acid rain now evident in the northeastern United States and Canada (see Atmospheric models). References 3—15 discuss atmospheric dilution as a control measure. The better approach, however, is to control emissions at the source. 

The autoclave is not the only component of an LDPE plant which may be exposed to a decomposition. Local hot spots in a secondary compressor may initiate a decomposition reaction consequendy it is necessary to protect these units from serious overpressure by pressure relieving devices and to release the products of the decomposition reactions safely. The problem of the aerial decomposition referred to eadier has been largely overcome by rapidly quenching the decomposition products as they enter the vent stack. 

U. S. EPA Regulations on Standards ofPeformanceforNeir Stationay Sources, 40 CER 60, Appendix A, Reference Methods, Washington, D.C., 1993. ASTM D3685-92, Standard Test Methodfor Sampling and Determination of Particulate Matter in Stack Gases, American Society for Testing Materials, Philadelphia, Pa., 1992. 

These structures are commonly referred to as hexagonal Hquid crystals. As the surfactant concentration is further increased, the tubules expand in a second direction to form large, stacked lamellar sheets of surfactants, commonly referred to as lamellar Hquid crystals. These Hquid crystals are very important in soap making. 

Production and consumption of wood products and residues are measured ia various units, based on common usage and their metric equivalents (2—4). Pulpwood logs and fuelwood are commonly measured ia cords. A cord refers to a stacked pile of wood, with outside dimensions of 4 by 4 by 8 ft (1.22 by 1.22 by 2.44 m) and a volume of 128 ft (3.62 m ). The weight of a cord depends on density of wood and bark and on moisture content. In the United States, it can range from 1.3 to 1.7 short tons (1.2 to 1.5 metric tons), air dried. 

Calcium siHcate hydrate is not only variable ia composition, but is very poody crystallised, and is generally referred to as calcium siHcate hydrate gel or tobermorite gel because of the coUoidal sizes (<0.1 fiva) of the gel particles. The calcium siHcate hydrates ate layer minerals having many similarities to the limited swelling clay minerals found ia nature. The layers are bonded together by excess lime and iatedayer water to form iadividual gel particles only 2—3 layers thick. Surface forces, and excess lime on the particle surfaces, tend to bond these particles together iato aggregations or stacks of the iadividual particles to form the porous gel stmcture. 

Stacking faults thereby providing barriers to sHp. If carbides are allowed to precipitate to the point of becoming continuous along the grain boundaries, they often initiate fracture (see Fracture mechanics). A thorough discussion of the mechanical properties of cobalt alloys is given in References 29 and 30 (see also Refractories). 

Research and development efforts have been directed toward improved ceU designs, theoretical electrochemical studies of magnesium ceUs, and improved cathode conditions. A stacked-type bipolar electrode ceU has been operated on a lab scale (112). Electrochemical studies of the mechanism of magnesium ion reduction have determined that it is a two-electron reversible process that is mass-transfer controUed (113). A review of magnesium production is found ia Reference 114. 

Cassettes Cassette is a term used to describe two different cross-flow membrane devices. The less-common design is a usually large stack of membrane separated by a spacer, with flow moving in parallel across the membrane sheets. This variant is sometimes referred to as a flat spiral, since there is some similarity in the way feed and permeate are handled. The more common cassette has long been popular in the pharmaceutical and biotechnical field. It too is a stack of flat-sheet membranes, but the membrane is usually connected so that the feed flows across the membrane elements in series to achieve higher conversion per pass. Their popularity stems from easy direct sc e-up from laboratoiy to plant-scale equipment. Their hmitation is that fluid management is inherently veiy hmited and inefficient. Both types of cassette are veiy compact and capable of automated manufacture. 

Velocity and Volumetric Flow Rate The U.S. EPA has published Method 2 as a reference method for determining stack-gas velocity and volumetric flow rate. At several designated sampling points, which represent equal portions of the stack volume (areas in the stack), the velocity and temperature are measured with instrumentation shown in Fig. 25-27. 

---