Clean-room, portable C++17 implementation of the PlanB IPv6 LPM algorithm.
Includes:
- AVX-512 SIMD path + scalar fallback
- Wait-free lookups with rebuild-and-swap dynamic FIB
- Benchmarks on synthetic data and real RIPE RIS BGP (~254K prefixes)
Interesting result: on real BGP + uniform random lookups, a plain Patricia trie can sometimes match or beat the SIMD tree due to cache locality and early exits.
Would love feedback, especially comparisons with PopTrie / CP-Trie.
The obvious question, I guess: How much faster are you than whatever is in the Linux kernel's FIB? (Although I assume they need RCU overhead and such. I have no idea what it all looks like internally.)
IPv6 longest-prefix-match (LPM).
Clean-room, portable C++17 implementation of the PlanB IPv6 LPM algorithm.
Includes: - AVX-512 SIMD path + scalar fallback - Wait-free lookups with rebuild-and-swap dynamic FIB - Benchmarks on synthetic data and real RIPE RIS BGP (~254K prefixes)
Interesting result: on real BGP + uniform random lookups, a plain Patricia trie can sometimes match or beat the SIMD tree due to cache locality and early exits.
Would love feedback, especially comparisons with PopTrie / CP-Trie.
The obvious question, I guess: How much faster are you than whatever is in the Linux kernel's FIB? (Although I assume they need RCU overhead and such. I have no idea what it all looks like internally.)
I likewise wonder from time to time whether I should replace WireGuard's allowedips.c trie with something better: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...
I wonder if this would port nicely over to rustybgp.
I wonder how this would look like in risc-v vector instructions
Why detect avx512 in build system instead of using #ifdef ?