# Bytes

As was said in the introduction (Section 2.1), chemical structures are the universal and the most natural language of chemists, but not for computers. Computers woi k with bits packed into words or bytes, and they perceive neither atoms noi bonds. On the other hand, human beings do not cope with bits very well. Instead of thinking in terms of 0 and 1, chemists try to build models of the world of molecules. The models ai e conceptually quite simple 2D plots of molecular sti uctures or projections of 3D structures onto a plane. The problem is how to transfer these models to computers and how to make computers understand them. This communication must somehow be handled by widely understood input and output processes. The chemists way of thinking about structures must be translated into computers internal, machine representation through one or more intermediate steps or representations (sec figure 2-23, The input/output processes defined [c.42]

Figure 5-3. a) Main organization of a database or container the basic units of a field are bits and bytes, b) Example of data organization in a flat-file. [c.229]

Program FOCK This program is designed to read in the LCAO-MO eoeffieient matrix, the one- and two-eleetron AO integrals and to form a elosed shell Foek matrix (i.e., a Foek matrix for speeies with all doubly oeeupied or bitals). With the program limitations deseribed above, FOCK memory usage is 255256 bytes. [c.646]

Program DIAG This program is designed to read in a real symmetrie matrix (but as a square matrix on disk), diagonalize it, and return all eigenvalues and eorresponding eigenveetors. With the program limitations deseribed above, DIAG memory usage is 738540 bytes. [c.646]

HyperChem computes two-electron integrals and saves them in main memory (two-electron integral buffer). The units of this buffer are in double-precision words (8 bytes per double-precision word in Windows). Once this buffer is full, these two-electron integrals are written to a temporary file on a disk (the selected disk can be set in the CHEM.INI file, before you start HyperChem, or with a script command, after you start HyperChem). A large buffer may reduce the processing time through fewer disk accesses. If the two-electron integral buffer size is big enough to hold all the integrals, HyperChem does not use the disk. [c.114]

HyperChem uses 16 bytes (two double-precision words) of storage for each electron repulsion integral. The first 8 bytes save the compressed four indices and the second 8 bytes store the value of the integral. Each index takes 16 bits. Thus the maximum number of basis functions is 65,535. This should satisfy all users of HyperChem for the foreseeable future. [c.263]

The total number of two-electron integrals is proportional to for a molecular system with m basis functions. Some of these integrals may be zero because of the symmetry, and some may be very small in magnitude. Using the regular integral format or the Raffenetti integral format, each integral value and its indices take 16 bytes and all are saved to the computer main memory or disk. All saved two-electron integrals are then used in forming the Fock matrix in every iteration. Those integrals with zero value or with a very small magnitude do not make much contribution to the Fock matrix and to the total energy. Neglecting these integrals may not affect the accuracy of ab initio calculations. Thus, in order to save computer main memory or disk space and speed up the calculation of the SCFprocedure, a two-electron integral cutoff is introduced. HyperChem uses the two-electron integral cutoff to determine which of the two-electron integrals need to be saved. The value of 10" ° (Hartree) generally is good enough for most calculations. A small value is recommended for tight calculations and a large value for loose calculations. [c.265]

Bits are put together as bytes. This example is an 8-bit byte. Faster, more powerful computers have more bits to the byte (16, 32, 64). In reading a byte, the bits flow one after the other out of the byte as electronic pulses (a positive voltage for on and zero for off). [c.306]

The previous discussion concentrated on arithmetical operations by computing in binary numbers represented as bits and bytes. However, other computer functions also use bytes of information. [c.307]

All operations of the computer processor or memory banks revolve around the use of bytes of information in which some of the bits carry the essentials and one or two other bits in each byte carry directional or instructional flags. A central processor of a computer is a piece of hardware that has been preprogrammed with its own memory (not accessible to the user — the so-called ROM or read-only memory) and is used to organize the reception of information (input), initial processing of the information, sending the initial information to a store (memory) or a software program for further processing, and finally sending results or instructions (output). [c.308]

A simple 8-bit device has been described above. These are very limited in what they can handle. For example, in terms of numbers, an 8-bit device (1 byte) alone can deal with numbers only up to 127 or 255, depending on how the eighth bit is used. This device would not be much use for general work, although it might be sufficient for simple instrument control. To get to larger numbers, other bytes must be used in conjunction with each other. Use of several bytes together necessarily slows computation. To get around this problem, larger sized bytes are used. The first major step was the introduction of 16-bit devices, followed soon thereafter by the now-common 32-bit bytes. [c.308]

By electronic engineering, a system of interconnected switching devices is able to respond in one of only two modes (on or off), and these modes can be controlled at the basic level of a bit. Bits are assembled into bytes, as with an 8-bit device, and through programming of the bytes a computer central processor can be made to follow sets of instructions (programs) written in special languages, either at a direct level (machine code) that can be acted upon immediately by a computer or at a high level that is translated for the user into machine code. [c.310]

Movement of information in a computer could be likened to a railway system. Carriers of information (bits or bytes) move together (like a train and wagons) from one location to another along electronic tracks. It is important that no two bits of information are mixed up, and therefore all the moves must be carefully synchronized with a clock. This situation resembles the movement of trains on a railway many trains use the same track but are not all in the same place at the same time. The railways run to a timetable. Similarly, information is moved around the computer under the control of the central processor unit (CPU). [c.311]

The capacity of a computer to carry out various tasks is partly governed by the number of bytes it has. Thus, a one-megabyte memory means there are 1 million locations with 8, 16, 32, or 64 bits in each. [c.419]

Digital computers operate with a binary system whereby all operations consist of a series of on/off electronic switching controlled by a crystal clock. The smallest on/off unit is the bit, and these are assembled into larger units known as bytes. The various functions of a computer are controlled through a processor, the CPU, which deals with incoming and outgoing signals and the execution of instructions (software) dealing with the signals. [c.419]

The number of bytes, each of which contains eight binary bits, assigned to describe each pixel dictates the number of colors that can be represented by the pixel. Assigning one bit to each cell allows two color values, eg, 0 or 1 (black or white). Assigning two bits allows four (2 ) color values, three bits allows eight (2 ), eight bits allows 256, and so on. Whereas 256 is not enough colors for commercially acceptable color photographic images, it is more than enough for commercially acceptable black-and-white photographic images, providing 256 shades of gray. [c.33]

Assigning three eight-bit bytes, one byte for each of the additive primary colors, red, green, and blue, to a single pixel provides 16,777,216 colors (256 ). This arrangement, called 24-bit color, provides commercially acceptable color photographic images. Hence, depending on the number of bytes [c.33]

It is possible to incorporate a Raman or IR spectrometer within a confocal microscope. This allows the spatial resolution of the microscope and compound identification of vibrational spectroscopy to be realized simultaneously. One of the reasons that this is a relatively new development is because of the tremendous volume of data generated. For example, if a Raman microscope has roughly 1 pm spatial resolution, and an area of 100 pmx 100 pm is to be imaged, and the frequency region from 800 cm -3400 is covered with 4 spectral resolution, then the data set has 6 million elements. Assuming each value is represented with a 4 byte number, the image would require 24 M Bytes of storage space. Wliile this is not a problem for current computers, the capacity of a typical hard drive on a PC from around 1985 (IBM 8088) was only 20 M Byte. Also, rapid data transfer is needed to archive and retrieve images. Furthenuore, in order to obtain the spectrum at any spatial position, array detectors (or FT metiiods) are required. A representative experimental set-up is shown in Figure B 1.2.12 [3]. [c.1174]

The database is defined as a self-describing collection of integrated records, mainly stored on hard disk or a CD-ROM. The structure of the database (tables, objects, indices, etc.) is described by metadata (data about data) and is stored within the database as a data dictionary (system catalog). Figure 5-3a presents the units for organizing data in a database. The smallest unit is a bit (0 or I), which is a component of a byte (8 bits = 1 byte). In a database, the bytes express fields of one or more records. These variable-lenght subsets of data of a particular entity consist of fields with unique information or characters (numeric, graphical, etc.). Records or data sets include different attributes, which describe corresponding object proper- [c.228]

IlyperChem uses 16 bytes (two double-precision words) of storage for each electron repulsion integral. The first 8 bytes save thecom-pressed four indices and the second S bytes store the value of the integral. Each index lakes 16 bits. Thus the maximum number of basis fiinctions is 65,535. This should satisfy all users of IlyperChem for the foreseeable future. [c.263]

Program INTEGRAL This program is designed to ealeulate on e- and two-eleetron AO integrals and to write them out to disk in eanonieal order (in Dirae <12 12> eonvention). It is designed to handle only S and P orbitals. With the program limitations deseribed above, INTEGRAL memory usage is 542776 bytes. [c.646]

Program MOCOEFS This program is designed to read in (from the keyboard) the LCAO-MO eoeffieient matrix and write it out to disk. Alternatively, you ean ehoose to have a unit matrix (as your initial guess) put out to disk. With the program 1 imitations deseribed above, MOCOEFS memory usage is 2744 bytes. [c.646]

With the program limitations deseribed above, FNCT MAT memory usage is 1960040 bytes. [c.646]

Program UTMATU This program is designed to read in a real matrix. A, a real transformation matrix, B, perform the transformation X = B(transpose) A B, and output the result. With the program limitations deseribed above, UTMATU memory usage is 1960040 bytes. [c.646]

Program MATXMAT This program is designed to read in two real matriees A and B, and to mul tiply them together AB = A B, and output the result. With the program limitations deseribed above, MATXMAT memory usage is 1470040 bytes. [c.646]

With the program limitations described above, FENERGY memory usage is 1905060 bytes.. [c.647]

Program TRANS This program is designed to read in the ECAO-MO coefficient matrix, the one- and two-elect ron AO integrals (in Dirac <12 12> convention), and to transform the integrals from the AO to the MO basis, and write these MO integrals to a file. With the program limitations described above, TRANS memory usage is 1905060 bytes. [c.647]

Progra m SCF This program is designed to read in the ECAO-MO coefficient matrix (or generate one), the one- and two-electron AO integrals and form a closed shell Fock matrix (i.e., a Fock matrix for species with all doubly occupied orbitals). It then solves the Fock equations iterating until convergence to six significant figures in the energy expression. A modified damping algorithm is used to insure convergence. With the program limitations described above, SCF memory usage is 259780 bytes. [c.647]

Program HAMIETON This program is designed to generate or read in a list of determinants. You can generate determinants for a CAS (Complete Active Space) of orbitals or you can inp ut your own list of determinants. Next, if you wish, you may read in the one- and two-electron MO integrals and form a Hamiltonian matrix over the determinants. Finally, if you so choose, you may diagonalize the Hamiltoni an matrix constructed over the determinants generated. With the program limitations described above, HAMIETON memory usage is 988784 bytes. [c.647]

Specifies the use of a regular format for saving the two-electron integrals. HyperChem uses 16 bytes to store every integral. The first 8 bytes stores the four indices of an integral and the second 8 bytes stores its value. HyperChem only stores an integral and its associated indices when the integral s absolute value is greater than or equal to the two-electron integral cutoff. The two-electron integral and its indices are stored without any modification when you choose this regular two-electron integral format. These two-electron integrals and their indices can be printed out to a log file by choosing a proper setting for QuantumPrintLevel in the CHEM.INI file. [c.114]

There are even available 64-bit and greater bytes. The latter devices can deal directly with numbers up to about 1000 million (10 ) and are very much faster than having to string together four or five 8-bit bytes. However, there is a price to be paid. Constructing 8-bit byte devices (chips) is now straightforward, but, even so, the yield of perfectly functioning chips is not good, and many have to be thrown away. For the 64-bit chips, the engineering complexities are enormous and the yield of perfect devices is quite low, hence the high cost. Except for very large number-cranching computers, most general-purpose computers, such as the PC, work with 16- or 32-bit chips. [c.309]

Blt-M ppedImages. A bit map is a grid pattern composed of tiny cells or picture elements called pixels. Each pixel has two attributes a location and a value or set of values. Location is defined as the address of the cell in a Cartesian, ie, x andjy coordinate, system. Value is defined as the color of the pixel in a specified color system. Geometric quaUties of images are a function of the location attribute, ie, the finer the grid pattern, the more precisely can the geometric quaUties be controlled. Color quaUties are a function of the value attribute, ie, the more bytes of computer memory assigned to describe each pixel, the more precisely can the color quaUties be controlled. [c.33]

The term SAXPY has arisen as a mnemonic for scalar alpha Xplus Y (2). This loop requires two operands and produces one result for each iteration of the loop. In 64 bit, or 8 byte, precision (8 bytes per floating-point number), this is a total of 24 bytes of data being consumed or produced per iteration. If the memory bandwidth were 240 megabytes per second, the memory subsystem could maintain this loop at 10 million iterations per second. Each loop iteration represents two floating-point operations, a multiplication and an addition thus mnning at 10 million iterations per second is only 20 MFLOPS, a small fraction of the peak performance of supercomputers. Most supercomputers have memory subsystems with much higher bandwidths, sometimes with separate pathways for read and write operations. Nevertheless, careful analysis of memory subsystem usage can be an important ingredient of any code optimization. In the preceding example, the system should look for subsequent operations to be performed on Z(I) while it is still near the CPU, before it is returned to main memory. The problem of optimizing CPU performance is not specific to supercomputers. However, given the enormous cost, a great deal more effort is devoted to optimizing codes on supercomputers than on other machines. [c.89]

See pages that mention the term

**Bytes**:

**[c.114] [c.114] [c.265] [c.647] [c.308] [c.286] [c.328] [c.328] [c.328] [c.328] [c.328] [c.328] [c.328] [c.328] [c.328]**

Mass Spectrometry Basics (2003) -- [ c.306 , c.307 , c.308 ]