Abstract: An apparatus and method for performing floating-point operations, particularly a fused multiply add operation. The apparatus includes a arithmetic logic unit adapted to produce both a high-order part (H) and a low-order part (L) of an intermediate extended result according to H, L=A*B+C, where A, B are input operands and C an addend. Each H, L part is formatted the same as the format of the input operands, and alignment of the resulting fractions is not affected by alignment of the inputs. The apparatus includes an architecture for suppressing left-alignment of the intermediate extended result, such that input operands for a subsequent A*B+C operation remain right-aligned.

Abstract: The present invention provides a system, apparatus, and method for detecting and resolving read-after-write hazards encountered in processors following the dispatch of instructions requiring one or more implicit reads in a processor.

Abstract: Arithmetic processing circuits in a circuit in a floating point processor having a fused multiply/ADD circuitry. In order to avoid waiting cycles in the normalizer of the floating point arithmetic, control logic calculates in an extremely early state of the overall Multiply/Add processing. Parts of the intermediate add result are significant and have to be selected in the pre-normalizer multiplexer to be fed to the normalizer by counting the leading zero bits (LAB) of the addend in a dedicated circuit right at the beginning of the pipe. LAB is added to the shift amount (SA) that is calculated to align the addend and is then compared with the width of the incrementer. If the sum of (SA+LAB) is larger than the width of the incrementer, which is a constant value, then no significant bits are in the high-part of the intermediate result, and the pre-normalizer multiplexer selects the data from a second predetermined position, otherwise from a first predetermined position.

Abstract: A method and system for performing floating point additive arithmetic operations of long operands in a narrow dataflow. The operands include first and second floating point numbers having first and second mantissas, respectively, the second operand greater than the first operand. The mantissas are both separated into a low portion and a high portion, the high portions are loaded into N-bit operand registers. The high portion of the first mantissa is aligned with respect to the high portion of the second mantissa, the high portions are then moved into 2N-bit registers. The low portion of the first mantissa is aligned in accordance with the alignment of the first mantissa high portion. The low portions of both mantissas are then concatenated into the registers, the first mantissa concatenated using a hold-function circuit. A 2N-bit-wide adder performs the additive arithmetic operation on the concatenated mantissas.

Abstract: The invention proposes a Floating Point Unit with fused multiply-add, with one addend Method for calculating a result of a division with an A-register and a B-register for two multiplicand operands and a C-register for an addend operand, wherein a divide processor using a subtractive method for calculation with a divisor register and a partial remainder register and a multiplier associated to an subtractor uses the C-register as input, wherein while loading the fraction of the dividend through the divisor register into the partial remainder register of the divide processor a calculated shifting is applied for alignment by using the multiplier associated to the subtractor.

Abstract: An apparatus and method for performing floating-point operations, particularly a fused multiply add operation. The apparatus includes a arithmetic logic unit adapted to produce both a high-order part (H) and a low-order part (L) of an intermediate extended result according to H, L=A*B+C, where A, B are input operands and C an addend. Each H, L part is formatted the same as the format of the input operands, and alignment of the resulting fractions is not affected by alignment of the inputs. The apparatus includes an architecture for suppressing left-alignment of the intermediate extended result, such that input operands for a subsequent A*B+C operation remain right-aligned.

Abstract: Method to convert a hexadecimal floating point number (H) into a binary floating point number by using a Floating Point Unit (FPU) with fused multiply add with an A-register a B-register for two multiplicand operands and a C-register for an addend operand, wherein a leading zero counting unit (LZC) is associated to the addend C-register, wherein the difference of the leading zero result provided by the LZC and the input exponent (E) is calculated by a control unit and determines based on the Raw-Result-Exponent a force signal (F) with special conditions like ‘Exponent Overflow’, ‘Exponent Underflow’, and ‘Zero Result’.

Abstract: A method and system for performing floating point additive arithmetic operations of long operands in a narrow dataflow. The operands include first and second floating point numbers having first and second mantissas, respectively, the second operand greater than the first operand. The mantissas are both separated into a low portion and a high portion, the high portions are loaded into N-bit operand registers. The high portion of the first mantissa is aligned with respect to the high portion of the second mantissa, the high portions are then moved into 2N-bit registers. The low portion of the first mantissa is aligned in accordance with the alignment of the first mantissa high portion. The low portions of both mantissas are then concatenated into the registers, the first mantissa concatenated using a hold-function circuit. A 2N-bit-wide adder performs the additive arithmetic operation on the concatenated mantissas.

Abstract: A method and system for performing integer divisions using subtraction-based division processes in a hardware divide processor primarily dedicated for floating-point division processes. In particular, the method and system involve calculating a quotient of a dividend and a divisor, the dividend and divisor being binary coded integer values, by normalizing the divisor and the dividend, determining a number of binary digits (nV) needed to represent the divisor and a number of binary digits (nD) needed to represent the dividend, determining a number of effective binary digits (nQ) needed to represent the quotient, determining a start bit position to start a subtraction-based divide process, and performing the subtraction-based divide process only for bit positions beginning at the start bit position and at a least significant bit position. In preferred embodiments, the subtraction-based divide process is an SRT (Sweeney, Robinson, Tocher) Divide process.

Abstract: A floating point unit of an in-order-processor having a register array for storing a plurality of operands, a pipeline for executing floating point instructions with a plurality of stages, each stage having a stage register, data input registers (1A, 1B, 1C) for keeping operands to be processed. The data input registers form the first stage register of the pipeline. An input port loads operands from outside said floating point unit into one of said data input registers. A plurality of bypass-registers are provided, the input of which is connected to the input port, and the output of which is provided to the data input registers (1A, 1B, 1C), such that data propagating through the pipeline to be loaded into the register array can be immediately supplied to one or more particular data input registers (1A, 1B, 1C) from a respective bypass-register without a delay caused by additional pipeline stages to be propagated through.

Abstract: Arithmetic processing circuits in a circuit in a floating point processor having a fused multiply/ADD circuitry. In order to avoid waiting cycles in the normalizer of the floating point arithmetic, control logic calculates in an extremely early state of the overall Multiply/Add processing. Parts of the intermediate add result are significant and have to be selected in the pre-normalizer multiplexer to be fed to the normalizer by counting the leading zero bits (LZB) of the addend in a dedicated circuit right at the beginning of the pipe. LZB is added to the shift amount (SA) that is calculated to align the addend and is then compared with the width of the incrementer. If the sum of (SA+LZB) is larger than the width of the incrementer, which is a constant value, then no significant bits are in the high-part of the intermediate result, and the pre-normalizer multiplexer selects the data from a second predetermined position, otherwise from a first predetermined position.

Abstract: A process is provided for monitoring the conversion of numerical values from a first to a second format, where before and after the conversion, the modulo residue of the corresponding numerical value is calculated and compared with the corresponding residue after the conversion. In this way it is possible to effect error-free monitoring of such a conversion, especially of computer data, without great hardware expenditure.

Abstract: A method and system for use in a data processing system is proposed, wherein the Input Exponent is used already in the subblocks of the mantissa addition. Early in the flow of a cycle, there are parts of the Potential exponent result generated and put together using zero detect signals and carry select signals of the Carry Select Adder of the mantissa addition. For the addition of two floating point numbers this reduces the number of required logic gates in the timing critical path. This allows a faster cycle time and/or less latency and/or more complex functions. The method and system according to the invention can be applied to adders of different mantissa widths or different exponent widths as well as power of radix 2.

Abstract: A correction of microprocessor chip design errors is achieved by identifying selected sets of instructions and/or selected sets of instruction sequences in an erroneous control flow of the microprocessor and/or by identifying selected sets of interface control and status signals. A match selectively initiates a corrective action by interfering with the instruction flow in the microprocessor chip or by requesting external control from an associated processor unit. Alternatively, a match is used for a programmable modification of interface control and status signals to adapt the chip to changes of its environment without redesign.