Intel Corporation

Yuyun Liao

Nigel C Paver

David B Roberts

Anthony Jebson

Deli Deng

Steve J Strazdus

James E Quinlan

Tom M Hameenanttila

James Quinlan

A tightly coupled dual 16-bit multiply-accumulate (MAC) unit for performing single-instruction/multiple-data (SIMD) operations may forward an intermediate result to another operation in a pipeline to resolve an accumulating dependency penalty. The MAC unit may also be used to perform 32-bit&#xd7;32-bit operations.

Multiply-accumulate (MAC) unit for single-instruction/multiple-data (SIMD) instructions

In an embodiment, a functional unit including a compressor section and a 36-bit SIMD adder is used to perform a STMD four-pixel averaging instruction. The functional unit generates four four-pixel averages. Four pixel values and a rounding value are compressed into a sum and a carry vector. The two least significant bits of the sum vector and the LSB of the carry vector are dropped before being input to the 36-bit SIMD adder. The two resultant 8-bit vectors are added by the 36-bit adder to directly generate the average pixel value result.

Method and apparatus for performing a pixel averaging instruction

A multiply-accumulate unit, or MAC, may achieve high throughput. The MAC need not use redundant hardware, such as multiple Wallace trees, or pipelining logic, yet may perform Wallace tree and carry look-ahead adder functions simultaneously for different operations.

Processing multiply-accumulate operations in a single cycle

An embodiment of the present invention is a mixed length encoding unit. The mixed length may be a 12/16 bits (12/16-b) encoding algorithm within a multiply-accumulate (MAC). The mixed length encoding unit includes 16-b Booth encoder adapted to produce eight partial product vectors from sixteen bits of data. The 16-b Booth encoder is coupled to a four stage Wallace Tree. During a first cycle of the invention, a multiplex system directs the eight partial products and an accumulation vector to a four stage Wallace Tree. During subsequent cycles, the multiplex system directs six partial product vectors, an accumulation vector, one carry-feedback input vector, and one sum-feedback input vector to the four stage Wallace Tree.

Growing community of inventors

Chandler, AZ, United States of America

Yuyun Liao