Abstract: A fused floating-point add-subtract unit includes far path logic, close path logic, and selection logic. The far path logic is configured to perform addition and subtraction operations on first and second significands of first and second operands, respectively, to produce a far path sum and a far path difference. The close path logic is configured to perform addition and subtraction operations on the first and second significands of the first and second operands, substantially concurrently with the addition and subtraction operations of the far path logic, to produce a close path sum and a close path difference. The selection logic selectively provides one of the far path sum and the close path sum as a significand of a sum output and one of the far path difference and the close path difference as a significand of a difference output.

Abstract: A fused floating-point add-subtract unit includes far path logic, close path logic, and selection logic. The far path logic is configured to perform addition and subtraction operations on first and second significands of first and second operands, respectively, to produce a far path sum and a far path difference. The close path logic is configured to perform addition and subtraction operations on the first and second significands of the first and second operands, substantially concurrently with the addition and subtraction operations of the far path logic, to produce a close path sum and a close path difference. The selection logic selectively provides one of the far path sum and the close path sum as a significand of a sum output and one of the far path difference and the close path difference as a significand of a difference output.

Abstract: A fused floating-point dot product unit. The fused dot product unit includes an improved alignment scheme that generates smaller significand pairs compared to the traditional alignment due to the reduced shift amount and sticky logic. Furthermore, the fused dot product unit implements early normalization and a fast rounding scheme. By normalizing the significands prior to the significand addition, the length of the adder can be reduced and the round logic can be performed in parallel. Additionally, the fused dot product unit implements a four-input leading zero anticipation unit thereby reducing the overhead of the reduction tree by encoding the four inputs at once. The fused floating-point dot product unit may also employ a dual-path (a far path and a close path) algorithm to improve performance. Pipelining may also be applied to the dual-path fused dot product unit to increase the throughput.