The present invention relates generally to a manner by which to operate upon data at a modem, or other communication device, that forms part of a radio, or other, communication system. More particularly, the present invention relates to apparatus, and an associated method, for the modem, or other communication device, by which Fourier and Viterbi operations are performed through the use of a unified, vector-architecture processor. Algorithms that perform the Fourier and Viterbi operations are vectorized, forming vectorized algorithms, to facilitate their execution by a vector processor. When implemented at a demodulator part of the modem, Fast Fourier Transforming (FFT) and Viterbi decoding operations are performed. A modem, or other communication device, of improved programmability and scalability is provided.
Advancements in communication and processing technologies have permitted the development, and implementation, of new types of communication systems. Generally, increased data rates of communication are permitted as a result of such advancements. Increasingly large amounts of data are permitted to be communicated within a selected time period.
New types of communication services, requiring high data throughput rates are possible in such new types of communication systems. Multimedia communication services are exemplary of new types of communication services, permitted now due to such communication-technology advancements. And, many other types of communication services are similarly also possible in communication systems constructed to take advantage of the communication technology advancements.
New-generation cellular, and other radio, communication systems are developed to provide for packet-based communications. Most multimedia communication services are predicated upon packet-based communications. Other communication services are analogously also migrating towards packet-based communications. Different types of communication schemes are used, or proposed for use, in such new communication systems. For instance, at least one WLAN (wireless local area network), to be operable pursuant to a variant of the IEEE (Institute of Electrical and Electronic Engineers) 802.11 specification proposes to utilize OFDM (orthogonal frequency division multiplexing) techniques. OFDM effectively forms a hybrid of a multi-carrier modulation (MCM) and frequency shift keying (FSK) modulation. Frequency-divided carrier frequencies are defined in an OFDM system such as the proposed, WLAN, and the carriers are selected to be orthogonal to one another, such as by separating the carriers by integer multiples of the inverses of symbol durations of parallel bit streams that are to be applied thereto. The orthogonal carriers are transmitted simultaneously, thereby permitting an entire allocated channel to be occupied through an aggregated sum of narrow, orthogonal subbands.
The WLAN, operational pursuant to the variant of the IEEE 802.11 standard. Packet-based communication schemes generally utilize digital communication techniques, and communication stations operable in many packet-based communication systems require the utilization of modems (MOdulators/DEModulators). While operation of modems vary, depending upon the type of communication scheme in which the modems are to be operable, certain characteristics of the modems are important. Characteristics include the cost, size, complexity, programmability, scalability, simplicity, and modifiability of the modem.
Existing modem constructions are sometimes categorized according to their general operational features.
One modem-type category forms a data-flow style modem. Such a modem construction is formed of fixed circuits. The fixed circuits are interconnected by way of fixed connections, and the circuits are driven by a common, centrally-controlled timing circuit. The fixed circuits generally are operable in only one manner, e.g., a circuit is capable of executing only a single algorithm. While acceptable for many implementations, a conventional, data-flow construction is unable readily to be modified, is generally of relatively large physical dimensions, is relatively power-consumptive, and exhibits difficulty of design maintenance. Lack of programmability and scalability are disadvantageous from an economic viewpoint and further adaptability.
Another modem-type category forms a DSP-style modem, i.e., a modem that utilizes a DSP (digital signal processor) in its construction. This type of modem is of improved programmability characteristics relative to a data-flow modem type. That is to say, the details of the modem are contained in software code and data arrangement, all storable in a memory device rather than in a particular interconnection of fixed circuits. However, the specialized nature of the DSP, and hardware structures ancillary thereto, limit the modifiability of the modem constructed therefrom. While more modifiable than a data-flow style modem construction, this modem type also lacks full programmability. A modem of this construction type also cannot therefore be modified easily.
Execution of modem-function algorithms at a processor having a vector architecture would be advantageous as the speed at which the operations can be effectuated would be increased. Fast Fourier Transforming and Viterbi decoding in the demodulator part of the modem would particularly be facilitated.
To date, however, a modem construction type has not been provided that fully takes advantage of the features of a processor having a vector architecture. If a modem utilizing such a processor could be provided modem operation would be facilitated and improved programmability and scalability to the modem would be provided.
A modem construction type utilizing such a vector architecture would be particularly advantageously implemented in a communication system that utilizes OFDM.
It is in light of this background information related to modem operation that the significant improvements of the present invention have evolved.
The present invention, accordingly, advantageously provides apparatus, and an associated method, by which to transform and code data that is communicated during operation of a radio, or other, communication system.
Through operation of an embodiment of the present invention, a manner is provided for a modem, or other communication device, by which to perform, e.g., Fourier transform and Viterbi decoding operations through the use of a unified, vector-architecture processor.
Algorithms that perform the Fourier transform and Viterbi decoding operations, as well as others, are vectorized, forming vectorized algorithms to facilitate their execution by parallel, vector processing devices. A modem of improved programmability and scalability is thereby provided.
In one aspect of the present invention, a modem includes a memory storage element for storing indicia associated with algorithms to be executed by parallel processors of a vector processor. The algorithms stored at the memory storage device are vectorized algorithms, vectorized in a manner to permit their execution in a parallel fashion by the processors of the vector processing architecture.
In another aspect of the present invention, the vectorized algorithms stored at the memory device are accessed through operation of a scalar processor. The scalar processor accesses the indicia and, responsive thereto, the indicia is selectably forwarded on to the vector processors. Selection is made at the scalar processor whether to forward on the indicia to the processing devices of the vector processor responsive to, e.g., values contained in selected fields into which the data forming the indicia is formatted. If the values in the field are of first selected values, the indicia is forwarded on to the processing devices. If the values contained in the selected field are of second selected values, the indicia is operated upon by the scalar processor in conventional manner. The fields containing the values, used by the scalar processor to determine whether to forward on the indicia to the processing devices or to operate upon the indicia thereat, are formed of unclassified fields defined in the associated formatting scheme.
In another aspect of the present invention, the vectorized algorithms are formed in manners such that the indicia forwarded on to the parallel processing devices are forwarded to each of the processing devices in unison so that common operations are performed by each of the processing devices of the vector processors.
The algorithms, vectorized and stored at the memory device, when executed, perform fast Fourier transform operations upon data provided to the modem. And, additional algorithms further perform, when executed by the parallel processing devices, Viterbi decoding operations.
Memory banks are positioned to receive data that is to be operated upon by the modem. The data is buffered at storage locations of the memory bank, to be retrieved therefrom and operated upon during execution of the algorithms by the parallel processing devices. The scalar core processor further operates to generate access signals to address selected lines of the memory bank.
In another aspect of the present invention, the vector processor further includes a permutation network for permuting data operated upon during operation of the processing devices of the vector processor as well as a vector register bank/file. The scalar core processor further operates to generate pattern signals for application to the permutation network to control permutation of the data provided thereto. And, the scalar core processor generates register number signals that are applied to the vector register bank.
Because the algorithms stored at the memory device are vectorized algorithms, capable of execution at parallel vector processors of a vector processor instructed pursuant to a unified, vector architecture, improved modem operation is possible. The complex computations required to perform the operations of modem operation are carried more quickly than many conventional modem-construction types. And, because the vectorized algorithms are software-implemented, the algorithms are readily modifiable. Changes to the algorithms are readily implementable merely by substituting, or overwriting, the indicia contained at the memory device. And, scalability of operation of the modem is also much-improved relative to conventional modem devices.
In these and other aspects, therefore, apparatus, and an associated method, is provided for a modem. The modem selectably operates upon digital data, and the digital data is transformed and coded, either by encoding or decoding, during operation of the modem. An instruction memory contains indicia associated with algorithms that operate to effectuate the transforming and coding of the digital data. A vector processor is formed of a first processing device and at least a second processing device. The first and at least second processing devices of the vector processor are positioned in parallel with one another to permit parallel operation. The vector processor executes the algorithms, the indicia which are stored at the instruction memory device, that effectuate the transforming and coding of the digital data.
A more complete appreciation of the present invention and the scope thereof can be obtained from the accompanying drawings that are briefly summarized below, the following detailed description of the presently-preferred embodiments of the present invention, and the appended claims.
Referring first to
The communication system includes a network part 16, here representative of a packet data network and a radio frequency section (RF)). And, a radio link 18 is formed to extend between the network part and the communication station 12. For purposes of illustration, the communication station 12 forms a mobile station operable in the wireless Local Area Network (LAN) communication system, and the communication station 14, as illustrated, forms a network-connected communication station.
Modems 22 of an embodiment of the present invention are coupled to the communication stations 12 and 14. The modems are functionally represented. In an actual implementation, the elements of the modem are implemented in any desired manner and are located at any appropriate location. The functionality of the modems 22 are, for instance, in an actual implementation, embodied within the structure of the respective communication stations 12 and 14, respectively. The modems here each are operable to operate upon data originating at, and terminating at, the communication station with which the modems are respectively associated.
Data is transported upon a communication channel, defined upon the radio link 18, and then delivered to a communication station. The modem at the receiving station operates to perform fast Fourier transforms (FFTs) upon the data. And, then, Viterbi decoding operations are performed upon the data, once transformed by the fast Fourier transforming operations. Thereafter, the data is forwarded to the communication station for further processing thereon.
Data originated at a communication station, i.e., the sending station, to be transported upon the radio link 18, is operated upon by the modem associated therewith to convert the data into a form to permit its communication upon communication channels formed upon the radio link. The data is encoded by convolutional encoding operations. And, then, inverse fast Fourier transforms are performed upon the data, once encoded.
An embodiment of the present invention utilizes a vector processor formed of a unified, vector architecture to facilitate modem operation to perform the modem functions of the modem. Through operation of the modem implemented in the manner of the present invention, improved modifiability and scalability of the modem is permitted.
Turning next to
The modem also includes a scalar core processor 32 and an instruction memory element 34 having memory locations accessible by the scalar core processor. The memory device 34 includes vectorized algorithms of an embodiment of the present invention that are permitting of execution by the parallel-positioned ALUs 28. When forming the receiver part of the modem, the vectorized algorithms stored at the memory device include, at least, an FFT (Fast Fourier Transform) algorithm and a Viterbi decoder algorithm. The indicia is formatted according to a formatting scheme such that the scalar core processor passes on the indicia to the arithmetic logic units. Values contained in an unclassified field, e.g., are used by the scalar core processor when deciding in what manner to make use of the stored indicia. When of appropriate values, the indicia is forwarded to the ALUs, here indicated by way of a bus line 36.
The modem is further shown to include a memory bank cache 38 at which data that is to be operated upon by the modem is cached. Additional instructions generated by the scalar core processor, here by way of a bus line 42, permit selective address of the memory devices of the memory bank cache 38. Thereby, signals generated by the scalar core processor on the control address line 42 determine what portions of the data provided to the modem are operated upon by the ALUs 28.
The vector processor is further shown to include a permutation network 44 coupled to the memory bank cache 38 to receive the data cached thereat and addressed by generation of the signals on the control address line 42. The permutation network operates to permute the data provided thereto by a selected permutation pattern. The permutation network is also coupled by way of a bus line 48 to receive pattern commands to control the permutation pattern by which the permutation network operates. The pattern control command is generated by the scalar core processor on the bus line.
The vector processor shown in the figure further includes a vector register bank/file 52 coupled to the permutation network to receive the data, once permutated thereat and also to the arithmetic logic units. Data, once permutated, is provided to the vector register bank. The vector register bank/file is further coupled to receive a register number command generated on the bus line 54, generated by the scalar core processor 32. Responsive to values of the registration number command, the data contained at the vector register bank/file is provided to the arithmetic logic units to be operated thereon.
By vectorizing the Viterbi decoding and Fast Fourier Transform algorithms, the transforming and decoding operations can be performed by the arithmetic logic units of the vector processor arrangement. Increased speed of transforming and decoding operations are possible through the use of the parallel processor arrangement. And, because the algorithms are software-implemented, the algorithms are easily modified or replaced to modify thereby operation of the modem.
Then, and as indicated by the block 64, the contents of the vector register 2 are subtracted from the contents of the vector register 1 and added to a vector register, indicated as register 4, of the vector register bank 52.
Then, and as indicated by the block 66, the contents of the vector registers 3 and 4 are shuffled. And, as indicated by the block 68, a factor, referred to as a twiddle factor, is loaded to a fifth vector register, designated as vector register 5, of the vector register bank 52.
Thereafter, and as indicated by the block 70, the contents of the vector registers 4 and 5 are multiplied to the contents of the vector register 4.
Then, as indicated by the block 72, the contents of the vector registers 3 and 4 are added to the contents of the vector register 1.
Thereafter, and as indicated by the block 74, the contents of the vector register 4 are subtracted from the contents of vector register 3 and to the vector register 2. Then, and as indicated by the block 76, the contents of the vector registers 1 and 2 are stored to memory. And, the process terminates, indicated by the block 78.
The algorithms that are created are vectorized to permit their efficient execution by parallel processing devices of the vector processor.
The algorithms include, for instance, Fast Fourier Transforming and Viterbi decoding algorithms to be used by a receive part of the modem. And, the algorithms include convolutional encoding and inverse Fast Fourier Transforming of data at the modulator side part of the modem.
Because of the software implementation of the functions of the modem, modifiability of the modem is easily implemented. And, correspondingly, the scalability of the modem is also facilitated.
The previous descriptions are of preferred examples for implementing the invention, and the scope of the invention should not necessarily be limited by this description. The scope of the present invention is defined by the following claims:
Number | Name | Date | Kind |
---|---|---|---|
5473600 | Liu | Dec 1995 | A |
6252902 | Simeon et al. | Jun 2001 | B1 |
6310910 | Shah et al. | Oct 2001 | B1 |
6505290 | Moyer et al. | Jan 2003 | B1 |
6549577 | Florencio et al. | Apr 2003 | B1 |
Number | Date | Country | |
---|---|---|---|
20040001535 A1 | Jan 2004 | US |