# US 2010/0164972 A1 · System, method and apparatus for memory with embedded associative section for computations

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0164972 A1
- Document ID: US20100164972A1
- Status: Published application
- Kind: A1
- Application number: 12/650,604
- Filed: 2009-12-31
- Published: 2010-07-01
- Earliest claimed: 2008-04-02
- Assignees: Zikbit Ltd.
- Inventors: Avidan AKERIB; Eli EHRMAN; Moshe MEYASSED; Oren AGAM
- Web page: https://patentgrep.com/patent/US20100164972A1

## CPC

- G11C 7/1006: {Data managing, e.g. manipulating data before writing or reading out, data bus switches or control circuits therefor}
- G09G 5/06: using colour palettes, e.g. look-up tables
- G09G 5/393: Arrangements for updating the contents of the bit-mapped memory
- G11C: STATIC STORES (semiconductor memory devices H10B)
- G09G: ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION (arrangements for transferring data between digital computers and displays G06F3/14; static indicating arrangements comprising an association of a number of separate sources or light control cells G09F9/00; static indicating arrangements comprising integral associations of a number of light sources H01J, H01K, H10H, H05B33/12; scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission, details thereof H04N1/00)

## Abstract

A system and method for data processing, the method includes: storing input data words in a row-wise manner in a memory that comprises multiple memory cells arranged in rows and columns; and transposing multiple data words by performing a sequence of shift operations and associative operations; wherein an associative operation comprises comparing in parallel multiple columns of associative memory cells to at least one comparand; and storing transposed data words in the memory.

## Claims

### Claim 1

A method for data processing, the method comprising:
storing input data words in a row-wise manner in a memory that comprises multiple memory cells arranged in rows and columns; and
transposing multiple data words by performing a sequence of shift operations and associative operations; wherein an associative operation comprises comparing in parallel multiple columns of associative memory cells to at least one comparand; and storing transposed data words in the memory.

### Claim 2 (depends on 1)

The method according to claim 1, further comprising processing a plurality of transposed data words by applying at least one associative operation.

### Claim 3 (depends on 1)

The method according to claim 1, wherein the transposing comprises repeating the stages of:
writing content of a source row of memory cells to a first row of associative memory section cells;
writing at least a portion of a shift mask to a second row of associative memory section cells;
comparing in parallel multiple columns of associative memory cells to a comparand, wherein each column of the associative memory cells comprises one cell of the first row of associative memory cells and a corresponding cell of the second row of associative memory section cells;
writing a comparison result to a destination row of the memory;
shifting the shift mask; and
selecting a new source row and selecting a new destination row.

### Claim 4 (depends on 1)

The method according to claim 1, comprising storing transposed data words in a column-wise manner.

### Claim 5 (depends on 1)

The method according to claim 1, wherein adjacent transposed data words are transposed versions of adjacent input data words.

### Claim 6 (depends on 1)

The method according to claim 1, wherein adjacent transposed data words are transposed versions of input data words that were spaced apart from each other.

### Claim 7 (depends on 1)

The method according to claim 1, comprising transposing entire pixels input data sets to provide pixels column-wise transposed data sets; wherein a pixel input data set comprises a row-wise sequence of different components of a single pixel.

### Claim 8 (depends on 1)

The method according to claim 1, comprising performing a component-wise transposing of pixel input data sets to provide pixel component-wise transposed data sets; wherein a pixel input data set comprises a row-wise sequence of different components of a single pixel.

### Claim 9 (depends on 1)

The method according to claim 1, comprising transposing groups of input data words to provide column-wise transposed groups of data words; wherein a group of input data words comprises multiple vertically aligned input data words.

### Claim 10 (depends on 9)

The method according to claim 9, wherein each group of input data words comprises multiple pixels.

### Claim 11 (depends on 9)

The method according to claim 9, comprising duplicating multiple input data words and storing transposed versions of duplicated input data words.

### Claim 12 (depends on 9)

The method according to claim 9, comprising duplicating multiple input data words so that transposed data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 13 (depends on 1)

The method according to claim 1, comprising transposing and duplicating code components of a pixel input data set while maintaining a row-wise arrangements of color components of the pixel input data set; wherein the pixel input data set comprises a row-wise sequence of different components of a single pixel and a code component that related to the single pixel.

### Claim 14

A method for data processing, comprising:
storing input data words in a row-wise manner in a memory that comprises multiple memory cells arranged in rows and columns;
re-arranging the input data words in response to parallel computational capabilities of an associative memory section; and
processing re-arranged input-data words by performing at least one associative operation that comprises comparing in parallel multiple columns of associative memory cells to at least one comparand.

### Claim 15 (depends on 14)

The method according to claim 14, comprising re-arranging frames that have a line size that is a fraction of width of row of associative memory section cells; wherein the re-arranging comprises storing multiple frame lines in a single row of the memory.

### Claim 16 (depends on 14)

The method according to claim 14, comprising duplicating multiple input data words so that data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 17 (depends on 14)

The method according to claim 14, further comprising transposing multiple data words by performing a sequence of shift operations and associative operations.

### Claim 18 (depends on 14)

The method according to claim 14, comprising re-arranging frames that have a line size that is larger than a width of row of associative memory section cells; wherein the re-arranging comprises storing vertically partitioning a frame to multiple frame portions, wherein a width of each frame portion does not exceed the width of row of associative memory section cells.

### Claim 19 (depends on 18)

The method according to claim 18, comprising duplicating multiple input data words so that data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 20

A method for data processing, comprising:
storing input data words in a memory that comprises multiple memory cells arranged in rows and columns;
applying at least one associative operation on input data words; wherein the applying comprises comparing in parallel multiple columns of associative memory cells to at least one comparand; and
utilizing address decoders for internal system transfers and outer system transfers; wherein the system comprises the address decoders, and registers rows that are coupled between the columns of the associative memory cells and the multiple memory cells; wherein internal system transfers involve storing data words in registers coupled to the address decoders.

### Claim 21

A method for data processing, comprising:
storing input data words in a row-wise manner in a memory that comprises multiple memory cells arranged in rows and columns; and
processing multiple row-wise data word sets; wherein the processing comprises generating at least one carry prediction vector, wherein the processing comprises multiple shift operations and associative operations, wherein an associative operation comprises comparing in parallel multiple columns of associative memory cells to at least one comparand.

### Claim 22

A system having data processing capabilities, the system comprises:
a semiconductor substrate;
a memory that is formed on the semiconductor substrate and comprises multiple memory cells arranged in rows and columns;
associative memory cells that are arranged in rows and columns; and
a control logic that is coupled to the memory and to the associative memory cells;
wherein the memory stores input data words in a row-wise manner;
wherein the control logic controls a transposition of multiple data words by performing a sequence of shift operations and associative operations that provide transposed data words;
wherein an associative operation comprises comparing in parallel multiple columns of associative memory cells to at least one comparand; and
wherein the memory stores the transposed data words.

### Claim 23 (depends on 22)

The system according to claim 22, wherein the control logic controls a plurality of transposed data words by applying at least one associative operation.

### Claim 24 (depends on 22)

The system according to claim 22, wherein the control logic is adapted to control a transposition of multiple data words by a repetition of:
write a content of a source row of memory cells to a first row of associative memory section cells;
write at least a portion of a shift mask to a second row of associative memory section cells;
compare in parallel multiple columns of associative memory cells to a comparand;
wherein each column of the associative memory cells comprises one cell of the first row of associative memory cells and a corresponding cell of the second row of associative memory section cells;
write a comparison result to a destination row of the memory; shifting the shift mask; and
select a new source row and selecting a new destination row.

### Claim 25 (depends on 22)

The system according to claim 22, wherein the transposed data words are stored in a column-wise manner.

### Claim 26 (depends on 22)

The system according to claim 22, wherein the transposed data words are stored such that adjacent transposed data words are transposed versions of adjacent input data words.

### Claim 27 (depends on 22)

The system according to claim 22, wherein the transposed data words are stored such that adjacent transposed data words are transposed versions of input data words that were spaced apart from each other.

### Claim 28 (depends on 22)

The system according to claim 22 wherein the control logic is adapted to control a transposing of entire pixels input data sets to provide pixels column-wise transposed data sets; wherein a pixel input data set comprises a row-wise sequence of different components of a single pixel.

### Claim 29 (depends on 22)

The system according to claim 22, wherein the control logic is adapted to control a component-wise transposing of pixel input data sets to provide pixel component-wise transposed data sets; wherein a pixel input data set comprises a row-wise sequence of different components of a single pixel.

### Claim 30 (depends on 22)

The system according to claim 22, wherein the control logic is adapted to control a transposing of groups of input data words to provide column-wise transposed groups of data words; wherein a group of input data words comprises multiple vertically aligned input data words.

### Claim 31 (depends on 30)

The system according to claim 30, wherein each group of input data words comprises multiple pixels.

### Claim 32 (depends on 30)

The system according to claim 30, wherein the control logic is adapted to control a duplication of multiple input data words and store transposed versions of duplicated input data words.

### Claim 33 (depends on 30)

The system according to claim 30, wherein the control logic is adapted to control a duplication of multiple input data words so that transposed data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 34 (depends on 22)

The system according to claim 22, wherein the control logic is adapted to control a transposing and a duplicating of code components of a pixel input data set while maintaining a row-wise arrangements of color components of the pixel input data set; wherein the pixel input data set comprises a row-wise sequence of different components of a single pixel and a code component that related to the single pixel.

### Claim 35

A system having data processing capabilities, the system comprises:
a semiconductor substrate;
a memory that is formed on the semiconductor substrate and comprises multiple memory cells arranged in rows and columns;
associative memory cells that are arranged in rows and columns; and
control logic that is coupled to the memory and to the associative memory cells;
wherein the control logic controls a re-arrangement of the input data words in response to parallel computational capabilities of an associative memory section, and controls a process of re-arranged input-data words by instructing the associative memory cells to perform at least one associative operation that comprises comparing in parallel multiple columns of associative memory cells to at least one comparand.

### Claim 36 (depends on 35)

The system according to claim 35, wherein the control logic is adapted to control a re-arrangement of frames that have a line size that is a fraction of width of row of associative memory section cells; wherein the re-arranging comprises storing multiple frame lines in a single row of the memory.

### Claim 37 (depends on 35)

The system according to claim 35, wherein the control logic is adapted to control a duplication of multiple input data words so that data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 38 (depends on 35)

The system according to claim 35, wherein the control logic is adapted to control a transposing of multiple data words by controlling a sequence of shift operations and associative operations.

### Claim 39 (depends on 35)

The system according to claim 35, wherein the control logic is adapted to control a re-arrangement of frames that have a line size that is larger than a width of row of associative memory section cells; wherein the re-arrangement comprises storing vertically partitioning a frame to multiple frame portions, wherein a width of each frame portion does not exceed the width of row of associative memory section cells.

### Claim 40 (depends on 39)

The system according to claim 39, wherein the control logic is adapted to control a duplication of multiple input data words so that data words that store pixel information of adjacent pixels are stored in proximity to each other.

### Claim 41

A system having data processing capabilities, the system comprises:
a semiconductor substrate,
a memory that is formed on the semiconductor substrate and comprises multiple memory cells arranged in rows and columns;
at least one address decoder;
multiple registers coupled to an input output bus that is coupled to additional components;
associative memory cells that are arranged in rows and columns, and
a control logic that is coupled to the memory and to the associative memory cells;
wherein the memory stores input data words;
wherein the associative memory cells apply at least one associative operation on input data words;
wherein the applying comprises comparing in parallel multiple columns of associative memory cells to at least one comparand; and
wherein the system utilizes address decoders for internal system transfers and outer system transfers.

### Claim 42

A system having data processing capabilities, the system comprises:
a semiconductor substrate;
a memory that is formed on the semiconductor substrate and comprises multiple memory cells arranged in rows and columns;
associative memory cells that are arranged in rows and columns; and
a control logic that is coupled to the memory and to the associative memory cells;
wherein the memory stores input data words in a row-wise manner;
wherein the associative memory cells process multiple row-wise data word sets;
wherein a processing of the multiple row-wise data word sets comprises generating at least one carry prediction vector;
wherein the processing comprises multiple shift operations and associative operations;
wherein an associative operation comprises comparing in parallel multiple columns of associative memory cells to at least one comparand.
