Skip to content

Re-release (Public Domain)

Latest
Latest
Compare
Choose a tag to compare
@GiantDarth GiantDarth released this 19 Oct 00:14
v1.0.0
9fddf0f
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Verified
Learn about vigilant mode.

This is a re-release of v1.0.0 with the CC0 license in place of the original Apache License. Functionally everything else is the same.

Features

  • Added support for the following symmetric keys (ciphers):

    • AES-256-ECB
    • ChaCha20

  • Added support for the following public keys:

    • ECC Secp256r1

  • Added support for the following hash functions:

    • MD5
    • SHA1
    • SHA224
    • SHA256
    • SHA384
    • SHA512
    • SHA3-224
    • SHA3-256
    • SHA3-384
    • SHA3-512
    • SHAKE128
    • SHAKE256
    • KangarooTwelve

  • Added support for the following modes:

    • Direct: Pass in a known host seed and known client output to derive the client's seed
    • --random: Generate the seed and errors randomly. Useful for benchmarking that models
      uniformly random distributions.
    • --benchmark: Generate the seed and errors such that the errors are exactly halfway through a
      core's/rank's workload. Useful for benchmarking but you're short on time or want
      reproducibility.

  • Added support for both OpenMP and OpenMPI implementations

  • Added the following options:

    • --all: Enforces search to continue through an entire Hamming distance's search space even if a
      matching seed is found
    • --fixed: Only searches a given Hamming distance's search space (no more nor less)

  • Added the ability to choose between OpenSSL vs. a faster custom AES256 implementation using
    ALWAYS_EVP_AES. Defaults to the custom implementation

  • Added the ability to choose between EVP vs. low level OpenSSL implementation of MD5, SHA1, and
    SHA2 using ALWAYS_EVP_HASH. Defaults to the low level implementations.

  • Added the ability to choose between OpenSSL vs. XKCP implementations for SHA3 and SHAKE
    using ALWAYS_EVP_SHA3. Defaults to OpenSSL implementations. (Note: Testing has found that the
    OpenSSL tends to be faster than the generic64, SSSE3, and AVX XKCP compilations on Zen 2.
    AVX2 is about on par. AVX512 XKCP is significantly faster than OpenSSL on Xeon Skylake
    CPUs.)

  • Added the ability to manually select the architecture to build XKCP against between:

    • reference
    • reference32bits
    • compact
    • generic32
    • generc32lc
    • generic64
    • generic64lc
    • SSSE3
    • AVX
    • AVX2
    • AVX512

    Defaults to generic64 to maximize compatibility while balancing for basic optimizations.

Assets 3
Loading