Instructions for efficiently accessing unaligned partial vectors

A method for executing a store-partial instruction, comprising:

• receiving the store-partial instruction to be executed, wherein the store-partial instruction specifies a destination address in memory, which is arbitrarily aligned;
• and
• executing the store-partial instruction, which involves storing a partial-vector-sized datum from a register into a naturally-aligned memory region encompassing the destination address, and in doing so, if the destination address is unaligned, rotating the bytes of the partial-vector-sized datum by swapping a set of bytes residing at addresses lower than the destination address with a set of bytes residing at addresses greater than or equal to the destination address, positions, wherein N is equivalent to either the destination address specified by the instruction modulo the vector length in bytes or the destination address specified by the instruction modulo the vector length in bytes subtracted from the vector length in bytes,
• wherein rotating the bytes of the partial-vector-sized datum occurs after the partial-vector-sized datum moves out of the register, and
• wherein rotating the bytes of the partial-vector-sized datum involves using alignment circuitry which is located along a load-store path between the memory and the register to cause the least significant byte of the partial-vector-sized datum to be stored at the specified destination address on a little-endian processor, or to cause the most significant byte of the partial-vector-sized datum to be stored at the specified destination address on a big-endian processor, or to cause a specified byte to be stored to the destination address in the case of an endian-specific store-partial partial variant.

A method for executing a load-control-vector-partial instruction, comprising:

• receiving a load-control-vector-partial instruction to be executed, wherein the load-control-vector-partial instruction specifies a target address in memory, which is arbitrarily aligned;
• and
• executing the load-control-vector-partial instruction to construct a control vector comprising predicate elements, wherein executing the load-control-vector-partial instruction involves determining a value N, wherein N is the specified target address modulo a length in bytes of a partial-vector-sized datum, wherein the predicate elements comprise a true polarity and a false polarity, wherein the control vector is constructed based on N and an endian-ness of a memory transaction,
• wherein for a big-endian memory transaction the N most-significant elements in the control vector are set to the true polarity and the remaining elements of the vector are set to the false polarity,
• wherein for a little-endian memory transaction the N least-significant elements in the control vector are set to the true polarity and the remaining elements of the vector are set to the false polarity, and
• wherein the control vector is used by a vector select instruction to determine which individual bytes from multiple vectors are selected to merge into a single output vector.

A computer system computer configured to execute a load-partial instruction, comprising:

• a processor;
• a memory;
• an instruction fetch unit within the processor configured to fetch the load-partial instruction to be executed, wherein the load-partial instruction specifies a source address in memory, which is arbitrarily aligned;
• and
• an execution unit within the processor configured to execute the load-partial instruction by loading a partial-vector-sized datum from a naturally-aligned memory region encompassing the source address into a register, and in doing so, if the source address is unaligned, rotating the bytes of the partial-vector-sized datum by swapping a set of bytes residing at addresses greater than or equal to the source address,
• wherein rotating the bytes of the vector involves rotating the bytes N positions, wherein N is equivalent to either the source address specified by the instruction modulo the vector length in bytes or the source address specified by the instruction modulo the vector length in bytes subtracted from the vector length in bytes,
• wherein rotating the bytes of the partial-vector-sized datum occurs before the partial-vector-sized datum reaches the register, and
• wherein rotating the bytes of the partial-vector-sized datum involves using alignment circuitry which is located along a load-store path between the memory and the register to cause the byte at the specified source address to reside at the least-significant byte position within the partial-vector-sized datum for a little-endian memory transaction, or causing the byte to be positioned at the most-significant byte position within the partial-vector-sized datum for a big-endian memory transaction.

A computer system configured to execute a store-partial instruction, comprising:

• a processor;
• a memory;
• an instruction fetch unit within the processor configured to fetch the store-partial instruction to be executed, wherein the store-partial instruction specifics a destination address in memory, which is arbitrarily aligned;
• and
• an execution unit within the processor configured to execute the store-partial instruction by storing a partial-vector-sized datum from a register into a naturally-aligned memory region encompassing the destination address, and in doing so, if the destination address in unaligned, rotating the bytes of the partial-vector-sized datum by swapping a set of bytes residing at addresses lower than the destination address with a set of bytes residing at addresses greater than or equal to the destination address,
• wherein rotating the bytes of the vector involves rotating the bytes N positions, wherein N is equivalent to either the source address specified by the instruction modulo the vector length in bytes or the source address specified by the instruction modulo the vector length in bytes subtracted from the vector length in bytes,
• wherein rotating the bytes of the partial-vector-sized datum occurs after the partial-vector-sized datum moves out of the register, and
• wherein rotating the bytes of the partial-vector-sized datum involves using alignment circuitry which is located along a load-store path between the memory and the register to cause the least significant byte of the partial-vector-sized datum to be stored to at the specified destination address on a little-endian processor, or to cause the most significant byte of the partial-vector-sized datum to be stored to the destination address on a big-endian processor, or to cause the specified byte to be stored to the destination address in the case of an endian-specific store-partial variant.

A computer system configured to execute a load-control-vector-partial instruction, comprising:

• a processor;
• a memory;
• an instruction fetch unit within the processor configured to fetch the load-control-vector-partial instruction to be executed, wherein the load-control-vector-partial instruction specifies a target address in memory, which is arbitrarily aligned;
• and
• an execution unit within the processor configured to execute the load-control-vector-partial instruction to construct a control vector comprising predicate elements, wherein executing the load-control-vector-partial instruction involves determining a value N, wherein N is the specified target address modulo the partial-vector-sized datum length in bytes, wherein the predicate elements comprise a true polarity and a false polarity, wherein the control vector is constructed based on N and an endian-ness of a memory transaction,
• wherein for a big-endian memory transaction the N most-significant elements in the control vector are set to the true polarity and the remaining elements of the vector are set to the false polarity, and
• wherein for a little-endian memory transaction the N least-significant elements in the control vector are set to the true polarity and the remaining elements of the vector are set to the false polarity, and
• wherein the control vector is used by a vector select instruction to determine which individual bytes from multiple vectors are selected to merge into a single output vector.