CHERI Concentrate Compression

Concept

CHERI Concentrate Compression is a fat-pointer compression algorithm introduced by Woodruff et al. to encode the base, top, and address bounds of a CHERI capability in a small number of bits, improving encoding efficiency, mitigating pipeline problems, and easing the semantic restrictions of earlier compressed encodings.

First seen 6/11/2026

Last seen 6/11/2026

Evidence 1 chunks

Wiki v1

WIKI

Overview

CHERI Concentrate Compression is a fat-pointer compression algorithm used in the CHERI capability architecture to represent the memory bounds (base and top) of a capability inside a small, fixed-size bit field. It was introduced by Woodruff et al. and is described in section 2.4.6 of the CHERI-RISC-V VP++ thesis.

The algorithm's stated goals are to:

READ FULL ARTICLE →

NEIGHBORHOOD

No graph connections found for this entity yet. It may appear in future ingestion runs.

explore full graph →

LINKED ENTITIES

1 links

CHERI used_by The evidence states that CHERI uses the CHERI Concentrate Compression algorithm to represent bounds in capabilities.

CITATIONS

6 sources

6 citations — click to expand

[1] CHERI Concentrate Compression was introduced by Woodruff et al. and is a fat-pointer compression scheme that improves encoding efficiency, solves pipeline problems, and eases semantic restrictions of compressed encoding. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...

[2] The uncompressed CHERI-256 format stores permissions/otype/flags, a 256-bit length, a base, and an address, and requires roughly four times the architecture's address size. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...

[3] CHERI Concentrate Compression stores two values B and T together with an internal exponent flag EAF, using about 26 bits for B and T and one bit for EAF. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...

[4] Decompression first determines exponent A from the EAF flag (A=0 if EAF=0, otherwise extracted from the lowest bits of B and T), and reconstructs T from B using a carry-out bit and the exponent, from which the final base and top addresses are calculated. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...

[5] The stored value of T is two bits smaller than the base, because it can be reconstructed from B depending on a carry-out bit and the exponent. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...

[6] Bounds in CHERI represent the lower and upper memory addresses the capability is authorized to access, and dereferencing an out-of-bounds capability raises an exception, which prevents buffer overflows and out-of-bounds execution. CHERI-RISC-V VP++: A Virtual Prototyping Platform Enabling Fine ...