No-locality hint vector memory access processors, methods, systems, and instructions

Patent metadata

Field	Value
Publication number	US-9600442-B2
Application number	US-201414335006-A
Country	US
Kind code	B2
Filing date	Jul 18, 2014
Priority date	Jul 18, 2014
Publication date	Mar 21, 2017
Grant date	Mar 21, 2017

Abstract

A processor of an aspect includes a plurality of packed data registers, and a decode unit to decode a no-locality hint vector memory access instruction. The no-locality hint vector memory access instruction to indicate a packed data register of the plurality of packed data registers that is to have a source packed memory indices. The source packed memory indices to have a plurality of memory indices. The no-locality hint vector memory access instruction is to provide a no-locality hint to the processor for data elements that are to be accessed with the memory indices. The processor also includes an execution unit coupled with the decode unit and the plurality of packed data registers. The execution unit, in response to the no-locality hint vector memory access instruction, is to access the data elements at memory locations that are based on the memory indices.

First claim

What is claimed is:

1. A processor comprising:
a plurality of packed data registers;
a cache hierarchy;
a decode unit to decode a no-locality hint vector memory access instruction, the no-locality hint vector memory access instruction to indicate a packed data register of the plurality of packed data registers that is to have a source packed memory indices, the source packed memory indices to have a plurality of memory indices, wherein the no-locality hint vector memory access instruction is to provide a no-locality hint to the processor for data elements that are to be accessed with the memory indices; and
an execution unit coupled with the decode unit and the plurality of packed data registers, the execution unit, in response to the no-locality hint vector memory access instruction, to access the data elements at memory locations that are based on the memory indices, wherein the no-locality hint vector memory access instruction comprises a no-locality hint vector load instruction, wherein the execution unit, in response to the no-locality hint vector load instruction, is to load the data elements from the memory locations, and wherein the cache hierarchy, in response to the no-locality hint vector load instruction, upon a cache hit for a data element, is to output no more than half a cache line from the cache hierarchy.

2. The processor of claim 1 , wherein the cache hierarchy, in response to the no-locality hint vector load instruction, is not to cache data elements loaded from the memory locations which do not hit in the cache hierarchy.

3. The processor of claim 1 , wherein the cache hierarchy, in response to the no-locality hint vector load instruction, upon a cache miss for a data element, is not to allocate space in the cache hierarchy for the data element that is to be loaded from memory.

4. The processor of claim 1 , wherein the cache hierarchy, in response to the no-locality hint vector load instruction, upon the cache hit for the data element, is to output no more than a single data element from the cache hierarchy.

5. The processor of claim 1 , wherein the no-locality hint vector memory access instruction comprises a no-locality hint gather instruction, wherein the no-locality hint gather instruction is to indicate a destination packed data register of the plurality of packed data registers, wherein the execution unit, in response to the no-locality hint gather instruction, is to store a packed data result in the destination packed data register, and wherein the packed data result is to include the data elements gathered from the memory locations.

6. The processor of claim 1 , wherein the decode unit is to decode the no-locality hint vector memory access instruction that is to have at least one bit that is to have a first value to indicate the no-locality hint, and is to have a second value to indicate lack of the no-locality hint.

7. The processor of claim 1 , wherein the decode unit is to decode the no-locality hint vector memory access instruction that is to have a plurality of bits that are to have a first value to indicate that the no-locality hint is a no-temporal locality hint, a second value to indicate that the no-locality hint is a no-spatial locality hint, and a third value to indicate that the no-locality hint is a no-temporal and no-spatial locality hint.

8. The processor of claim 1 , wherein the decode unit is to decode the no-locality hint vector memory access instruction that is to indicate a source packed data operation mask.

9. A processor comprising:
a plurality of packed data registers;
a decode unit to decode a no-locality hint vector memory access instruction, the no-locality hint vector memory access instruction to indicate a packed data register of the plurality of packed data registers that is to have a source packed memory indices, the source packed memory indices to have a plurality of memory indices, wherein the no-locality hint vector memory access instruction is to provide a no-locality hint to the processor for data elements that are to be accessed with the memory indices;
an execution unit coupled with the decode unit and the plurality of packed data registers, the execution unit, in response to the no-locality hint vector memory access instruction, to access the data elements at memory locations that are based on the memory indices; and
a memory controller, wherein the no-locality hint vector memory access instruction comprises a no-locality hint vector load instruction, and wherein the memory controller, in response to the no-locality hint vector load instruction, is to load no more than half a cache line amount of data, for each of the data elements loaded from memory.

10. The processor of claim 9 , wherein the memory controller, in response to the no-locality hint vector load instruction, is to load no more than 128-bits for each of the data elements loaded from memory.

11. A processor comprising:
a plurality of packed data registers;
a decode unit to decode a no-locality hint vector memory access instruction, the no-locality hint vector memory access instruction to indicate a packed data register of the plurality of packed data registers that is to have a source packed memory indices, the source packed memory indices to have a plurality of memory indices, wherein the no-locality hint vector memory access instruction is to provide a no-locality hint to the processor for data elements that are to be accessed with the memory indices;
an execution unit coupled with the decode unit and the plurality of packed data registers, the execution unit, in response to the no-locality hint vector memory access instruction, to access the data elements at memory locations that are based on the memory indices; and
a memory controller, wherein the no-locality hint vector memory access instruction comprises a no-locality hint vector write instruction, wherein the execution unit, in response to the no-locality hint vector write instruction, is to write data elements of a source packed data indicated by the instruction over the data elements at the memory locations, and wherein the memory controller, in response to the no-locality hint vector write instruction, is to write no more than half a cache line amount of data, for each of the data elements of the source packed data that is written to memory.

12. A method in a processor comprising:
receiving a no-locality hint vector memory access instruction, the no-locality hint vector memory access instruction indicating a source packed memory indices having a plurality of memory indices, wherein the no-locality hint vector memory access instruction provides a no-locality hint to the processor for data elements that are to be accessed with the memory indices, wherein receiving the no-locality hint vector memory access instruction comprises receiving a no-locality hint vector load instruction; and
accessing the data elements at memory locations that are based on the memory indices in response to the no-locality hint vector memory access instruction, wherein accessing comprises loading the data elements from the memory locations, including loading no more than half a cache line amount of data, for each data element loaded from memory.

13. The method of claim 12 , further comprising omitting caching data elements that are loaded from memory in a cache hierarchy.

14. The method of claim 12 , further comprising, upon a cache hit for a data element in a cache hierarchy, outputting no more than half a cache line from the cache hierarchy.

15. An article of manufacture comprising a non-transitory machine-readable storage medium, the non-transitory machine-

Assignees

• Intel Corp

Inventors

• Hughes Christopher J

Classifications

• G06F15/8069 Primary

  using a cache · CPC title

• G06F9/30

  Arrangements for executing machine instructions, e.g. instruction decode (for executing microinstructions G06F9/22) · CPC title

• G06F12/0897

  with two or more cache hierarchy levels (with multilevel cache hierarchies G06F12/0811) · CPC title

• G06F2212/6028

  Prefetching based on hints or prefetch instructions · CPC title

• G06F12/0862

  with prefetch · CPC title