摘要:
A method of an aspect includes receiving a floating point round-off amount determination instruction. The instruction indicates a source of one or more floating point data elements, indicates a number of fraction bits after a radix point, and indicates a destination storage location. A result including one or more result floating point data elements is stored in the destination storage location in response to the floating point round-off amount determination instruction. Each of the one or more result floating point data elements includes a difference between a corresponding floating point data element of the source in a corresponding position, and a rounded version of the corresponding floating point data element of the source that has been rounded to the indicated number of the fraction bits. Other methods, apparatus, systems, and instructions are disclosed.
摘要:
When legacy instructions, that can only operate on smaller registers, are mixed with new instructions in a processor with larger registers, special handling and architecture are used to prevent the legacy instructions from causing problems with the data in the upper portion of the registers, i.e., the portion that they cannot directly access. In some embodiments, the upper portion of the registers are saved to temporary storage while the legacy instructions are operating, and restored to the upper portion of the registers when the new instructions are operating. A special instruction may also be used to disable this save/restore operation if the new instruction are not going to use the upper part of the registers.
摘要:
Methods and apparatus are disclosed using an index array and finite state machine for scatter/gather operations. Embodiment of apparatus may comprise: decode logic to decode scatter/gather instructions and generate micro-operations. An index array holds a set of indices and a corresponding set of mask elements. A finite state machine facilitates the scatter operation. Address generation logic generates an address from an index of the set of indices for at least each of the corresponding mask elements having a first value. Storage is allocated in a buffer for each of the set of addresses being generated. Data elements corresponding to the set of addresses being generated are copied to the buffer. Addresses from the set are accessed to store data elements if a corresponding mask element has said first value and the mask element is changed to a second value responsive to completion of their respective stores.
摘要:
In-lane vector shuffle operations are described. In one embodiment a shuffle instruction specifies a field of per-lane control bits, a source operand and a destination operand, these operands having corresponding lanes, each lane divided into corresponding portions of multiple data elements. Sets of data elements are selected from corresponding portions of every lane of the source operand according to per-lane control bits. Elements of these sets are copied to specified fields in corresponding portions of every lane of the destination operand. Another embodiment of the shuffle instruction also specifies a second source operand, all operands having corresponding lanes divided into multiple data elements. A set selected according to per-lane control bits contains data elements from every lane portion of a first source operand and data elements from every corresponding lane portion of the second source operand. Set elements are copied to specified fields in every lane of the destination operand.
摘要:
A single micro-instruction to perform either an N-bit or a 2N-bit load is provided. A microprocessor having an N-bit load port performs either an N-bit load or a 2N-bit load in a single cycle with the same micro-instruction being used for both the N-bit and the 2N-bit load.
摘要:
A method and apparatus for a two micro-operation flow using source override. In one embodiment, the method includes the identification of a macro-instruction having one or more streaming single instruction multiple data extension type operands. Once received, the macro-instruction is decoded into a first micro-operation (uOP) and a second uOP. Once decoded, a signal is asserted to disable source operand override logic if the first micro-operation updates a logical destination register that matches a logical source register of the micro-operation. Otherwise, the mutual source override is active and executed by a register alias table (RAT) when uOP with matching logic source and destination register are detected in a same clock cycle. In doing so, macro-instructions having 128-bit operands may be processed using, for example, two uOPs (one for the lower half and one for the upper half) in a 64-bit implementation, while preserving the atomicity of the original instruction.
摘要:
A processor may include an instruction decoder to decode macroinstructions into micro-operations. In some embodiments, the instruction decoder may include a first decoder and a second decoder. The first decoder may decode a macroinstruction having SSE data type operands into a laminated micro-operation, and may generate unlamination information for the laminated micro-operation. The second decoder may generate from the laminated micro-operation and the unlamination information two or more micro-operations, where operands of the two or more micro-operations each correspond to a half of one of the SSE operands of the macroinstruction.
摘要:
A micro-operation (uop) fusion technique. More particularly, embodiments of the invention relate to a technique to fuse two or more uops originating from two or more instructions.
摘要:
In some disclosed embodiments instruction execution logic provides conditional memory fault assist suppression. Some embodiments of processors comprise a decode stage to decode one or more instruction specifying: a set of memory operations, one or more register, and one or more memory address. One or more execution units, responsive to the one or more decoded instruction, generate said one or more memory address for the set of memory operations. Instruction execution logic records one or more fault suppress bits to indicate whether one or more portion of the set of memory operations are masked. Fault generation logic is suppressed from considering a memory fault corresponding to a faulting one of the set of memory operations when said faulting one of the set of memory operations corresponds to a portion of the set of memory operations that is indicated as masked by said one or more fault suppress bits.
摘要:
Methods and apparatus are disclosed for using an index array and finite state machine for scatter/gather operations. Embodiment of apparatus may comprise: decode logic to decode a scatter/gather instruction and generate a set of micro-operations, and an index array to hold a set of indices and a corresponding set of mask elements. A finite state machine facilitates the gather operation. Address generation logic generates an address from an index of the set of indices for at least each of the corresponding mask elements having a first value. An address is accessed to load a corresponding data element if the mask element had the first value. The data element is written at an in-register position in a destination vector register according to a respective in-register position the index. Values of corresponding mask elements are changed from the first value to a second value responsive to completion of their respective loads.