Intel Corporation

Gregory Henry

Alexander F Heinecke

Dmitry G Baksheev

Embodiments detailed herein relate to arithmetic operations of float-point values. An exemplary processor includes decoding circuitry to decode an instruction, where the instruction specifies locations of a plurality of operands, values of which being in a floating-point format. The exemplary processor further includes execution circuitry to execute the decoded instruction, where the execution includes to: convert the values for each operand, each value being converted into a plurality of lower precision values, where an exponent is to be stored for each operand; perform arithmetic operations among lower precision values converted from values for the plurality of the operands; and generate a floating-point value by converting a resulting value from the arithmetic operations into the floating-point format and store the floating-point value.

Computer processor for higher precision computations using a mixed-precision decomposition of operations

The present disclosure relates to an apparatus that includes decoding circuitry that decodes a single instruction. The single instruction includes an identifier of a first source operand, an identifier of a second source operand, an identifier of a destination, and an opcode indicative of execution circuitry is to multiply from the identified first source operand and the identified second source operand and store a result in the identified destination. Additionally, the apparatus includes execution circuitry to execute the single decoded instruction to calculate a dot product by calculating a plurality of products using data elements of the identified first and second operands using values less precise than the identified first and second source operands, summing the calculated products, and storing the summed products in the destination.

Systems and methods for combining low-mantissa units to achieve and exceed FP64 emulation of matrix multiplication

Provided is a compile time interface to run-time libraries that can reduce processing overhead in loops. A compile time interface identifies, at compile time, a loop that contains a library function in response to detecting a compiler hint associated with the library function. The compile time interface generates a bypass structure and modifies the loop to include a call to a planning function for a first pass through the loop and a call to a bypass function for one or more subsequent passes though the loop. The planning function sets the bypass function equal to an optimized library if one or more selected arguments to the library function are loop-invariant.

Compile time interface to run-time libraries

An optimizing compiler and method thereof performs a sequence of optimizing changes to an intermediate language representation of a routine, and measures an execution characteristic of each optimization, such as a timing of the machine language representation performed on an architecture similar to the target machine using a user selectable initialization state. The sequence of optimizations is selected according to a criterion that includes a lexicographic search and other methods. The pre-optimized code is also broken into segments wherein discrete optimizations are performed on each segment and measured using a user provided routine. The target routine is tested with the object code in main memory if not the cache if possible and optimizations are chosen only if they improve the target subroutine according to the user defined metric. After a stopping criterion is achieved, the most optimized code is selected.

Growing community of inventors

Hillsboro, OR, United States of America

Gregory Henry