Decoding Base64 at 4 GB/s with SIMD
How vectorized lookup and a couple of bit-packing tricks turn a memory-bound loop into a memcpy-fast monster.
i write here to think out loud — about systems, tools, things i've built, and occasionally something that has nothing to do with any of that.
How vectorized lookup and a couple of bit-packing tricks turn a memory-bound loop into a memcpy-fast monster.
What writing a tiny SQL engine taught me about indexes, storage, and trade-offs.