Parallel texture sampling

One or more computer-readable non-transitory storage media embodying software that is operable when executed to:

• store, by a plurality of first buffer blocks, a plurality of texels organized into a texel array comprising a plurality of N×N texel sub-arrays,
• wherein each texel of the plurality of texels falls within a corresponding N×N texel sub-array and is associated with a two-dimensional sub-array coordinate indicating a position of that texel within the corresponding N×N texel sub-array, and
• wherein each first buffer block of the plurality of first buffer blocks is assigned a particular two-dimensional sub-array coordinate and stores a texel subset of the plurality of texels having the particular two-dimensional sub-array coordinate;
• receive, by a plurality of filter blocks, the plurality of texels from the plurality of first buffer blocks, wherein the plurality of filter blocks comprise a first filter block configured to receive a first texel from each first buffer block of the plurality of first buffer blocks, and wherein the first texels received from the plurality of first buffer blocks are adjacent texels in the texel array comprising the plurality of N×N texel sub-arrays;
• and
• perform, by the plurality of filter blocks, sampling operations parallelly on respective N×N texel sub-arrays.

A method comprising:

• storing, by a plurality of first buffer blocks, a plurality of texels organized into a texel array comprising a plurality of N×N texel sub-arrays,
• wherein each texel of the plurality of texels falls within a corresponding N×N texel sub-array and is associated with a two-dimensional sub-array coordinate indicating a position of that texel within the corresponding N×N texel sub-array, and
• wherein each first buffer block of the plurality of first buffer blocks is assigned a particular two-dimensional sub-array coordinate and stores a texel subset of the plurality of texels having the particular two-dimensional sub-array coordinate;
• receiving, by a plurality of filter blocks, the plurality of texels from the plurality of first buffer blocks, wherein the plurality of filter blocks comprise a first filter block configured to receive a first texel from each first buffer block of the plurality of first buffer blocks, and wherein the first texels received from the plurality of first buffer blocks are adjacent texels in the texel array comprising the plurality of N×N texel sub-arrays;
• and
• performing, by the plurality of filter blocks, sampling operations parallelly on respective N×N texel sub-arrays.