# US 2026/0099249 A1 · Self-managed dram modules with built-in adaptive data compression

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

- Number: US 2026/0099249 A1
- Document ID: US20260099249A1
- Status: Published application
- Kind: A1
- Application number: 18/905,527
- Filed: 2024-10-03
- Published: 2026-04-09
- Earliest claimed: 2024-10-03
- Assignees: Scaleflux, Inc.
- Applicants: ScaleFlux, Inc.
- Inventors: Yang Liu; Fei Sun; Tong Zhang
- Web page: https://patentgrep.com/patent/US20260099249A1

## CPC

- G06F 3/0608: {Saving storage space on storage systems}
- G06F 3/064: {Management of blocks}
- G06F 3/0673: {Single storage device}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A self-managed DRAM module and method. The module includes a plurality of DRAM chips; and a controller chip configured to compress a data chunk received from a host according to a process that includes: compressing the data chunk to generate a plurality of compression results, wherein each compression result includes a different block size; identifying an optimal compression result, wherein the optimal compression result has a lowest block size among the plurality of compression results that achieves a threshold compression; and storing the optimal compression result in the plurality of DRAM chips.

## Claims

### Claim 1

A self-managed dynamic random-access memory (DRAM) module, comprising:
a plurality of DRAM chips; and
a controller chip configured to compress a data chunk received from a host according to a process that includes:
compressing the data chunk to generate a plurality of compression results, wherein each compression result includes a different block size;
identifying an optimal compression result, wherein the optimal compression result has a lowest block size among the plurality of compression results that achieves a threshold compression; and
storing the optimal compression result in at least one of the plurality of DRAM chips.

### Claim 2 (depends on 1)

The self-managed DRAM module of claim 1, wherein identifying the optimal compression result includes sequentially evaluating compression ratios for compression results relative to the compression ratio of the compression result having a largest block size.

### Claim 3 (depends on 2)

The self-managed DRAM module of claim 2, wherein sequentially evaluating compression ratios includes:
evaluating the compression ratio of the compression result having a smallest block size; and
evaluating the compression ratio of the compression result having a next smallest block size.

### Claim 4 (depends on 3)

The self-managed DRAM module of claim 3, wherein sequentially evaluating compression ratios for compression results relative to the compression ratio of the compression result having the largest block size includes comparing it to a predefined threshold value.

### Claim 5 (depends on 1)

The self-managed DRAM module of claim 1, wherein compressing the data chunk to generate a plurality of compression results includes utilizes a multi-mode LZ compressor.

### Claim 6 (depends on 5)

The self-managed DRAM module of claim 5, wherein the multi-mode LZ compressor includes:
a single minimum-length match search engine that processes the data chunk using a maximum block size to generate a search result; and
a plurality of post-processing engines that individually post process the search result for each of the different block sizes to generate the plurality of compression results.

### Claim 7

A method of compressing data for a self-managed dynamic random-access memory (DRAM) module, comprising:
compressing a data chunk to generate a plurality of compression results, wherein each compression result includes a different block size;
identifying an optimal compression result, wherein the optimal compression result has a lowest block size among the plurality of compression results that achieves a threshold compression; and
storing the optimal compression result in at least one of a plurality of DRAM chips of the self-managed DRAM module.

### Claim 8 (depends on 7)

The method of claim 7, wherein identifying the optimal compression result includes sequentially evaluating compression ratios for compression results relative to the compression ratio of the compression result having a largest block size.

### Claim 9 (depends on 8)

The method of claim 8, wherein sequentially evaluating compression ratios includes:
evaluating the compression ratio of the compression result having a smallest block size; and
evaluating the compression ratio of the compression result having a next smallest block size.

### Claim 10 (depends on 9)

The method of claim 9, wherein sequentially evaluating compression ratios for compression results relative to the compression ratio of the compression result having the largest block size includes comparing it to a predefined threshold value.

### Claim 11 (depends on 7)

The method of claim 7, wherein compressing the data chunk to generate a plurality of compression results includes utilizes a multi-mode LZ compressor.

### Claim 12 (depends on 11)

The method of claim 11, wherein the multi-mode LZ compressor includes:
a single minimum-length match search engine that processes the data chunk using a maximum block size to generate a search result; and
a plurality of post-processing engines that individually post process the search result for each of the different block sizes to generate the plurality of compression results.

### Claim 13

A self-managed dynamic random-access memory (DRAM) module, comprising:
a plurality of DRAM chips; and
a controller chip configured to compress a data chunk received from a host according to a process that includes:
applying LZ compression on the data chunk to generate an LZ compressed block;
applying entropy coding to the LZ compressed block to generate an LZ/entropy compressed block; and
evaluating a read latency of the LZ compressed block and the LZ/entropy compressed block; and
selecting one of the LZ compressed block or the LZ/entropy compressed block for storage in at least one of the DRAM chips based on the read latencies of the LZ compressed block and the LZ/entropy compressed block.

### Claim 14 (depends on 13)

The self-managed DRAM module of claim 13, wherein evaluating the read latency includes:
determining a first throughput of a decompression engine that performs LZ decompression on the LZ compressed block; and
determining a second throughput of the decompression engine that performs both LZ decompression and entropy decoding on the LZ/entropy compressed block.

### Claim 15 (depends on 14)

The self-managed DRAM module of claim 14, wherein evaluating the read latency further includes determining a size of the LZ compressed block and a size of the LZ/entropy compressed block.

### Claim 16 (depends on 13)

The self-managed DRAM module of claim 13, wherein selecting one of the LZ compressed block or the LZ/entropy compressed block for storage is further based on a size of the LZ compressed block and a size of the LZ/entropy compressed block.
