Polynomial calculations optimized for programmable integrated circuit device structures

Description

CROSS REFERENCE TO RELATED APPLICATION

This claims the benefit of commonly-assigned U.S. Provisional Patent Application No. 61/729,797, filed Nov. 26, 2012, which is hereby incorporated by reference herein in its entirety.

FIELD OF THE INVENTION

This invention relates to computing floating-point polynomials in programmable integrated circuit devices such as programmable logic devices (PLDs).

BACKGROUND OF THE INVENTION

Certain operations, such as certain memory operations, require evaluation of multiple shifted instances of the same polynomials. This is commonly done by adding multiple shifted partial products for each term, and then adding together all of those sums. However, such an operation may be inefficient when implemented in certain types of devices, particularly in programmable devices such as field-programmable gate arrays (FPGAs) that perform logic operations using arrangements of look-up tables (LUTs). In particular, because at or near both the most-significant bit and the least significant bit there are many empty positions, the addition operations do not pack efficiently into the LUTs.

SUMMARY OF THE INVENTION

The present invention relates to method and circuitry for implementing polynomial calculations using logic structures such as those found in programmable devices such as FPGAs. Instead of summing all of the partial products for each respective multiplier, and then adding together those sums, the partial products are summed across all multipliers in a bit-slice-by-bit-slice order. The sums of the bit slices are then in turn shifted by their relative indices and summed, and then reduced as necessary. Because the initial addition is across a bit slice, all bit positions within the sum will have similar hamming count—i.e., will be similarly populated. Therefore, the sums will pack efficiently into LUTs. Moreover, any level of pipelining can be used.

In accordance with embodiments of the invention, there is provided polynomial circuitry including a respective partial product generator for each bit position of each term of a plurality of terms of a polynomial to be evaluated. A respective plurality of adders for each bit position adds partial products of a respective bit position across all of the plurality of terms to provide a respective bit-slice sum. Resulting bit-slice sums are offset from one another according to their respective bit positions. A final adder adds together the respective offset bit-slice sums to provide a result.

A method of configuring a programmable device as such polynomial circuitry is also provided, and a non-transitory machine-readable data storage medium is provided that is encoded with software for performing the method of configuring such circuitry on a programmable device.

BRIEF DESCRIPTION OF THE DRAWINGS

Further features of the invention, its nature and various advantages will be apparent upon consideration of the following detailed description, taken in conjunction with the accompanying drawings, in which like reference characters refer to like parts throughout, and in which:

FIG. 1 shows a conceptual arrangement for evaluating shifted polynomials;

FIG. 2 shows schematically a circuit arrangement that can serve as one of the multipliers in FIG. 1;

FIG. 3 shows schematically a circuit arrangement that can serve as one of the partial product generators in FIG. 2;

FIG. 4 shows a circuit arrangement in which the arrangement of FIG. 2 is provided in place of each of the multipliers in FIG. 1;

FIG. 5 shows an optimization of the arrangement of FIG. 3;

FIG. 6 shows the order of operations in the arrangements of FIGS. 2-5;

FIG. 7 shows the order of operations in accordance with embodiments of the present invention;

FIG. 8 shows a circuit arrangement in accordance with an embodiment of the present invention;

FIG. 9 is a cross-sectional view of a magnetic data storage medium encoded with a set of machine-executable instructions for performing a method according to the present invention;

FIG. 10 is a cross-sectional view of an optically readable data storage medium encoded with a set of machine executable instructions for performing a method according to the present invention;

FIG. 11 is a simplified block diagram of an illustrative system employing a programmable logic device incorporating the present invention; and

FIG. 12 is a flow diagram of an example of a method according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

One example of the use of embodiments of the current invention is the evaluation of Galois Field (or finite field) polynomials. When large Galois Field multiplier arrays (such as those used for a Chien search in Flash memory error-correcting codes, including BCH and Reed-Solomon codes) are used, the performance on an FPGA is greatly reduced because of the depth of logic. If the calculation is pipelined, the resource count increases significantly because of inefficiencies of packing the partial product calculations into LUTs effectively.

According to embodiments of the current invention, a Galois Field polynomial may be mapped in a new way onto FPGA logic structures. These embodiments can be used to quickly and efficiently evaluate polynomials for such applications such as the aforementioned Chien search used by BCH and Reed Solomon decoders. System performance requirements may require that these structures operate at speeds over 300 MHz, which may not be possible with previously known polynomial structures for FPGAs, particularly for some BCH applications that use larger field sizes (e.g., 13-15 bits).

A Chien search of polynomials requires one finite field multiplier for each polynomial term. If more than one location is checked at a time, the polynomial can be shifted to check another location by simply shifting each polynomial term. Shifting of the polynomial terms can be accomplished by multiplying each term with the required finite field value in multipliers 101, and summing all of the multiplier outputs together in adder 102, as shown in the arrangement in FIG. 1, in which the an terms are Galois Field elements that represent powers of the roots of the field, and the Xn terms are Galois Field elements that represent terms of the error locator polynomial (i.e., the polynomial being searched).

Each finite field multiplier 101 may have the structure 200 shown in FIG. 2, in which there are m partial products 201 having m bits each, where m is the number of bits in the field. Each respective partial product may be generated by ANDing the λ_n input with a respective bit of the an input, as shown in the example of FIG. 3, which shows example of a 4-bit polynomial term (λ) and one bit of a root powers value (α) generating one 4-bit partial product (in this case m=4). Each partial product is left shifted (211) by the bit position index of the bit in the second input to which it corresponds. The partial products are all summed in adder 202, creating a value having 2m−1 bits. That number is then reduced back to m-bits at 203 using the irreducible polynomial for the field, according to any of various known reduction techniques.

Thus, the arrangement of FIG. 1 may have the structure 400 shown in FIG. 4, with each multiplier 101 replaced with a copy of structure 200. One way to improve the efficiency of structure 400 is to eliminate the individual reducers 203, add the outputs of the various adders 202 in a further adder 502, and then apply a single reduction 503, as shown in FIG. 5, and as described in copending, commonly-assigned U.S. patent application Ser. No. 12/719,770, which is hereby incorporated by reference herein in its entirety.

However, even with the structure of FIG. 5, adding together all of the partial products still creates a potentially large number of values each having 2m−1 bits. This may create a very long logic path from the input finite field numbers α, λ to the output of the result. And if pipelining is used (by cutting the paths at any point, whether within or between the individual multipliers), the logic size may increase significantly, because of the uneven packing of the relatively shifted partial product bits into the LUTs, which increases as pipelining depth increases.

A solution in accordance with embodiments of the invention may be appreciated after considering FIGS. 6 and 7.

FIG. 6 is a diagrammatic representation of the matrix 600 of operations performed by arrangements such as that of the circuit arrangement of FIG. 5. Each multiplier is represented by a column 601. There are (n−k)/2 multipliers, corresponding to the number of columns, where n is the number of bits (for BCH coding) or symbols (for Reed-Solomon coding) in the codeword, and k is the number of data bits or symbols in the codeword—i.e., (n−k) is the number of parity bits or symbols in the codeword. There are m partial product generators 602 per multiplier, each m bits wide. Relative shifts are not shown. As indicated by arrow 603, the direction of first addition, corresponding to adders 202, is downward.

It can be shown mathematically that the identical result would be obtained if, as in matrix 700 of FIG. 7, the direction of first addition were horizontal as indicated by arrow 701. This may be implemented, as one example, by the circuit arrangement 750 of FIG. 8.

In arrangement 750, each column 601 of partial products 602 still represents a single term of the polynomial being evaluated (the relative shifts are shown in this drawing). However, unlike in arrangement 500, the partial products 602 in each of columns 601 are not added together in this case. Instead, the partial products 602 in each of rows 751 are added first. As shown here, this may be accomplished with a respective adder 752. It will be appreciated that while the adders 752 are shown as being logically located in each respective row 751 between each respective column 601, they may physically be present in those locations, but they also may be elsewhere on the programmable integrated circuit device and connected by routing resources of the programmable integrated circuit device to the respective partial product generators 602.

After all of the rows 751 have been added, the resulting sums 753, which are aligned with respective 1-bit offsets 754 as shown, are added by adder 755, and then reduced at 756 from 2m−1 bits to m bits. Because the additions of rows 751 occur prior to the offsetting or shifting, they can be efficiently packed into the LUTs of a programmable device. Therefore, even if the addition operations are pipelined (possible cut points where pipelining may occur are shown at dashed lines 702 of FIG. 7), there will be minimal additional programmable logic of the programmable device required to complete the additions. Although the final addition 755 occurs after offsetting or shifting, it is only one operation, so any inefficiency introduced at that point also has minimal effect.

As noted above, when arrangement 750 is constructed in programmable logic, adders 752 may be located elsewhere than between the partial product generators 602. Indeed, both the partial product generators 602 and the adders 752 could be configured by a user completely from programmable logic resources along with programmable interconnect resources. As one alternative, separate dedicated multiplication circuits may be provided in hard-wired form on the programmable device and the user may use the programmable interconnect resources to create addition circuits and connect them to the dedicated multiplication circuits to form arrangement 750.

One example of a method 1200 for configuring a programmable device as circuitry for evaluating a polynomial in accordance with an embodiment of the present invention is diagrammed in FIG. 12. Method 1200 begins at 1201, where, on the programmable device, a respective partial product generator is configured for each bit position of each term of a plurality of terms of a polynomial to be evaluated. Next, at 1202, a respective plurality of adders is configured on the programmable device for each bit position, to add partial products of a respective bit position across all of the plurality of terms to provide a respective bit-slice sum, wherein resulting bit-slice sums are offset from one another according to their respective bit positions. Finally, at 1203, a final adder that adds together said respective offset bit-slice sums to provide a result is configured on the programmable device.

Thus it is seen that circuitry and methods for efficiently performing polynomial calculations have been provided.

Instructions for carrying out a method according to this invention for programming a programmable device to perform polynomial calculations, may be encoded on a machine-readable medium, to be executed by a suitable computer or similar device to implement the method of the invention for programming or configuring PLDs or other programmable devices to perform operations as described above. For example, a personal computer may be equipped with an interface to which a PLD can be connected, and the personal computer can be used by a user to program the PLD using a suitable software tool, such as the QUARTUS® II software available from Altera Corporation, of San Jose, Calif.

FIG. 9 presents a cross section of a magnetic data storage medium 800 which can be encoded with a machine executable program that can be carried out by systems such as the aforementioned personal computer, or other computer or similar device. Medium 800 can be a floppy diskette or hard disk, or magnetic tape, having a suitable substrate 801, which may be conventional, and a suitable coating 802, which may be conventional, on one or both sides, containing magnetic domains (not visible) whose polarity or orientation can be altered magnetically. Except in the case where it is magnetic tape, medium 800 may also have an opening (not shown) for receiving the spindle of a disk drive or other data storage device.

The magnetic domains of coating 802 of medium 800 are polarized or oriented so as to encode, in manner which may be conventional, a machine-executable program, for execution by a programming system such as a personal computer or other computer or similar system, having a socket or peripheral attachment into which the PLD to be programmed may be inserted, to configure appropriate portions of the PLD, including its specialized processing blocks, if any, in accordance with the invention.

FIG. 10 shows a cross section of an optically-readable data storage medium 810 which also can be encoded with such a machine-executable program, which can be carried out by systems such as the aforementioned personal computer, or other computer or similar device. Medium 810 can be a conventional compact disk read-only memory (CD-ROM) or digital video disk read-only memory (DVD-ROM) or a rewriteable medium such as a CD-R, CD-RW, DVD-R, DVD-RW, DVD+R, DVD+RW, or DVD-RAM or a magneto-optical disk which is optically readable and magneto-optically rewriteable. Medium 810 preferably has a suitable substrate 811, which may be conventional, and a suitable coating 812, which may be conventional, usually on one or both sides of substrate 811.

In the case of a CD-based or DVD-based medium, as is well known, coating 812 is reflective and is impressed with a plurality of pits 813, arranged on one or more layers, to encode the machine-executable program. The arrangement of pits is read by reflecting laser light off the surface of coating 812. A protective coating 814, which preferably is substantially transparent, is provided on top of coating 812.

In the case of magneto-optical disk, as is well known, coating 812 has no pits 813, but has a plurality of magnetic domains whose polarity or orientation can be changed magnetically when heated above a certain temperature, as by a laser (not shown). The orientation of the domains can be read by measuring the polarization of laser light reflected from coating 812. The arrangement of the domains encodes the program as described above.

A PLD 90 programmed according to the present invention may be used in many kinds of electronic devices. One possible use is in a data processing system 900 shown in FIG. 11. Data processing system 900 may include one or more of the following components: a processor 901; memory 902; I/O circuitry 903; and peripheral devices 904. These components are coupled together by a system bus 905 and are populated on a circuit board 906 which is contained in an end-user system 907.

System 900 can be used in a wide variety of applications, such as computer networking, data networking, instrumentation, video processing, digital signal processing, or any other application where the advantage of using programmable or reprogrammable logic is desirable. PLD 90 can be used to perform a variety of different logic functions. For example, PLD 90 can be configured as a processor or controller that works in cooperation with processor 901. PLD 90 may also be used as an arbiter for arbitrating access to a shared resources in system 900. In yet another example, PLD 90 can be configured as an interface between processor 901 and one of the other components in system 900. It should be noted that system 900 is only exemplary, and that the true scope and spirit of the invention should be indicated by the following claims.

Various technologies can be used to implement PLDs 90 as described above and incorporating this invention.

It will be understood that the foregoing is only illustrative of the principles of the invention, and that various modifications can be made by those skilled in the art without departing from the scope and spirit of the invention. For example, the various elements of this invention can be provided on a PLD in any desired number and/or arrangement. One skilled in the art will appreciate that the present invention can be practiced by other than the described embodiments, which are presented for purposes of illustration and not of limitation, and the present invention is limited only by the claims that follow.