Abstract: An integrated circuit with specialized processing blocks is provided. A specialized processing block may be optimized for machine learning algorithms and may include a multiplier data path that feeds an adder data path. The multiplier data path may be decomposed into multiple partial product generators, multiple compressors, and multiple carry-propagate adders of a first precision. Results from the carry-propagate adders may be added using a floating-point adder of the first precision. Results from the floating-point adder may be optionally cast to a second precision that is higher or more accurate than the first precision. The adder data path may include an adder of the second precision that combines the results from the floating-point adder with zero, with a general-purpose input, or with other dot product terms. Operated in this way, the specialized processing block provides a technical improvement of greatly increasing the functional density for implementing machine learning algorithms.

Abstract: An integrated circuit that includes very large adder circuitry is provided. The very large adder circuitry receives more than two inputs each of which has hundreds or thousands of bits. The very large adder circuitry includes multiple adder nodes arranged in a tree-like network. The adder nodes divide the input operands into segments, computes the sum for each segment, and computes the carry for each segment independently from the segment sums. The carries at each level in the tree are accumulated using population counters. After the last node in the tree, the segment sums can then be combined with the carries to determine the final sum output. An adder tree network implemented in this way asymptotically approaches the area and performance latency as an adder network that uses infinite speed ripple carry adders.

Abstract: The present embodiments relate to integrated circuits with circuitry that implements floating-point trigonometric functions. The circuitry may include an approximation circuit that generates an approximation of the output of the trigonometric functions, a storage circuit that stores predetermined output values of the trigonometric functions, and a selector circuit that selects between different possible output values based on a control signal from a control circuit. In some embodiments, the circuitry may include a mapping circuit and a restoration circuit. The mapping circuit may map an input value from an original quadrant of the trigonometric circle to a predetermined input interval, and the restoration circuit may map the output value selected by the selection circuit back to the original quadrant of the trigonometric circle. If desired, the circuitry may be implemented in specialized processing blocks.

Abstract: Techniques for designing and implementing networks-on-chip (NoCs) are provided. For example, a computer-implemented method for programming a network-on-chip (NoC) onto an integrated circuit includes determining a first portion of a plurality of registers to potentially be included in a NoC design, determining routing information regarding datapaths between registers of the first portion of the plurality of registers, and determining an expected performance associated with the first portion of the plurality of registers. The method also includes determining whether the expected performance is within a threshold range, including the first portion of the plurality of registers and the datapaths in the NoC design after determining that the expected performance is within the threshold range, and generating instructions configured to cause circuitry corresponding to the NoC design to be implemented on the integrated circuit.

Abstract: An integrated circuit that includes very large adder circuitry is provided. The very large adder circuitry receives more than two inputs each of which has hundreds or thousands of bits. The very large adder circuitry includes multiple adder nodes arranged in a tree-like network. The adder nodes divide the input operands into segments, computes the sum for each segment, and computes the carry for each segment independently from the segment sums. The carries at each level in the tree are accumulated using population counters. After the last node in the tree, the segment sums can then be combined with the carries to determine the final sum output. An adder tree network implemented in this way asymptotically approaches the area and performance latency as an adder network that uses infinite speed ripple carry adders.

Abstract: Based on a Modified Gram-Schmidt (MGS) algorithm, QR decomposition techniques are optimized for parallel structures that provide arithmetic-logic unit (ALU) to ALU connectivity. The techniques utilize a different loop organization, but the dependent functional sequences of the algorithm are unchanged, thereby reducing likelihood of affecting error analysis and/or numerical stability. Some integrated circuit devices (e.g., FPGA) may implement hard floating-point (HFP) circuitry, such as a digital signal processing (DSP) block, distributed memories, and/or flexible internal connectivity, which can support the discussed high performance matrix arithmetic.

Abstract: Integrated circuits with digital signal processing (DSP) blocks are provided. A DSP block may include one or more large multiplier circuits. A large multiplier circuit (e.g., an 18×18 or 18×19 multiplier circuit) may be used to support two or more smaller multiplication operations sharing one or two sets of multiplier operands, a complex multiplication, and a sum of two multiplications. If the multiplier products overflow and interfere with one another, correction operations can be performed. Partial products from two or more larger multiplier circuits can be used to combine decomposed partial products. A large multiplier circuit can also be used to support two floating-point mantissa multipliers.

Abstract: A method for designing a system on a target device is disclosed. The system is synthesized from a register transfer level description. The system is placed on the target device. The system is routed on the target device. A configuration file is generated that reflects the synthesizing, placing, and routing of the system for programming the target device. A modification for the system is identified. The configuration file is modified to effectuate the modification for the system without changing the placing and routing of the system.

Abstract: The present embodiments relate to circuitry that efficiently performs floating-point arithmetic operations and fixed-point arithmetic operations. Such circuitry may be implemented in specialized processing blocks. If desired, the specialized processing blocks may include configurable interconnect circuitry to support a variety of different use modes. For example, the specialized processing block may efficiently perform a fixed-point or floating-point addition operation or a portion thereof, a fixed-point or floating-point multiplication operation or a portion thereof, a fixed-point or floating-point multiply-add operation or a portion thereof, just to name a few. In some embodiments, two or more specialized processing blocks may be arranged in a cascade chain and perform together more complex operations such as a recursive mode dot product of two vectors of floating-point numbers or a Radix-2 Butterfly circuit, just to name a few.

Abstract: The present disclosure describes a digital signal processing (DSP) block that includes a plurality of columns of weight registers and a plurality of inputs configured to receive a first plurality of values and a second plurality of values. The first plurality of values is stored in the plurality of columns of weight registers after being received. Additionally, the DSP block includes a plurality of multipliers configured to simultaneously multiply each value of the first plurality of values by each value of the second plurality of values.

Abstract: The present disclosure describes a digital signal processing (DSP) block that includes a plurality of columns of weight registers and a plurality of inputs configured to receive a first plurality of values and a second plurality of values. The first plurality of values is stored in the plurality of columns of weight registers after being received. In a first mode of operation, the first and second pluralities of values are received via a first portion of the plurality of inputs. In a second mode of operation, the first plurality of values is received via a second portion of the plurality of inputs, and the second plurality of values is received via the first portion of the plurality of inputs. Additionally, the DSP block includes a plurality of multipliers configured to simultaneously multiply each value of the first plurality of values by each value of the second plurality of values.

Abstract: Integrated circuits with programmable logic regions are provided. The programmable logic regions may be organized into smaller logic units sometimes referred to as a logic element. A logic element may include four lookup tables coupled to an adder carry chain. At least some of the lookup tables are configured to output combinatorial outputs, whereas the adder carry chain are used to output sum outputs. Both the combinatorial outputs and the sum outputs may be used simultaneously to support a multiplication operation, three or more logic operations, or arithmetic and combinatorial operations in parallel.

Abstract: A method for implementing a programmable device is provided. The method may include extracting an underlay from an existing routing network on the programmable device and then mapping a user design to the extracted underlay. The underlay may represent a subset of fast routing wires satisfying predetermined constraints. The underlay may be composed of multiple repeating adjacent logic blocks, each implementing some datapath reduction operation. Implementing circuit designs in this way can dramatically improve circuit performance while cutting down compile times by more than half.

Abstract: Integrated circuits that include lightweight processor cores are provided. Each processor core may be configured to execute a series of instructions. At least one of the instructions may include an embedded delay field with a value specifying the amount of time that instruction needs to wait before proceeding to the next instruction to avoid a data hazard. The value of the delay field may be determined by a compiler during software compile time. Such delay field may also be used in conjunction with branch instructions to specify a number of no-operations (NOPs) for one or more associated branch delay slots and may also be used to reduce data forwarding cost.

Abstract: The present embodiments relate to performing reduced-precision floating-point arithmetic operations using specialized processing blocks with higher-precision floating-point arithmetic circuitry. A specialized processing block may receive four floating-point numbers that represent two single-precision floating-point numbers, each separated into an LSB portion and an MSB portion, or four half-precision floating-point numbers. A first partial product generator may generate a first partial product of first and second input signals, while a second partial product generator may generate a second partial product of third and fourth input signals.

Abstract: Integrated circuits with digital signal processing (DSP) blocks are provided. A DSP block may include one or more large multiplier circuits. A large multiplier circuit such as an 18×18 multiplier circuit may be used to support two or more smaller multiplication operations such as two 8×8 integer multiplications or two 9×9 integer multiplications. To implement the two 8×8 or 9×9 unsigned/signed multiplications, the 18×18 multiplier may be configured to support two 8×8 multiplications with one shared operand, two 6×6 multiplications without any shared operand, or two 7×7 multiplications without any shared operand. Any potential overlap of partial product terms may be subtracted out using correction logic. The multiplication of the remaining most significant bits can be computed using associated multiplier extension logic and appended to the other least significant bits using merging logic.

Abstract: Integrated circuits with programmable logic regions are provided. The programmable logic regions may be organized into smaller logic units sometimes referred to as a logic cell. A logic cell may include four 4-input lookup tables (LUTs) coupled to an adder carry chain. Each of the four 4-input LUTs may include two 3-input LUTs and a selector multiplexer. The carry chain may include at three or more full adder circuits. The outputs of the 3-input LUTs may be directly connected to inputs of the full adder circuits in the carry chain. By providing at least the same or more number of full adder circuits as the total number of 4-input LUTs in the logic cell, the arithmetic density of the logic is enhanced.

Abstract: The present embodiments relate to integrated circuits with floating-point arithmetic circuitry that handles normalized and denormalized floating-point numbers. The floating-point arithmetic circuitry may include a normalization circuit and a rounding circuit, and the floating-point arithmetic circuitry may generate a first result in form of a normalized, unrounded floating-point number and a second result in form of a normalized, rounded floating-point number. If desired, the floating-point arithmetic circuitry may be implemented in specialized processing blocks.

Abstract: Integrated circuits with specialized processing blocks are provided. The specialized processing blocks may include floating-point multiplier circuits that can be configured to support variable precision. A multiplier circuit may include a first carry-propagate adder (CPA), a second carry-propagate adder (CPA), and an associated rounding circuit. The first CPA may be wide enough to handle the required precision of the mantissa. In a bridged mode, the first CPA may borrow an additional bit from the second CPA while the rounding circuit will monitor the appropriate bits to select the proper multiplier output. A parallel prefix tree operable in a non-bridged mode or the bridged mode may be used to compute multiple multiplier outputs. The multiplier circuit may also include exponent and exception handling circuitry using various masks corresponding to the desired precision width.

Abstract: Integrated circuits with specialized processing blocks are provided. A specialized processing block may include one real addition stage and one real multiplier stage. The multiplier stage may simultaneously feed its output to the addition stage and directly to an adjacent specialized processing block. The addition stage may also produce sum and difference outputs in parallel. A group of four such specialized processing blocks may be connected in a chain to implement a radix-2 fast Fourier transform (FFT) butterfly. Multiple radix-2 butterflies may be stacked to form yet higher order radix butterflies. If desired, the specialized processing block may also be used to implement a complex multiply operation. Three or four specialized processing blocks may be chained together and along with one or more adders outside the specialized processing blocks, real and imaginary portions of a complex product can be generated.