Expand description
Vectorized SIMD implementations for x86_64 (AVX2).
Functions§
- decode_
avx2 ⚠ - Decodes 32 Base64 ASCII bytes to 24 raw bytes using AVX2. Returns (input_bytes_consumed, output_bytes_written). Breaks out if padding (‘=’) or invalid characters are encountered.
- decode_
avx512 ⚠ - Decodes 64 Base64 ASCII bytes to 48 raw bytes using AVX-512BW. Returns (input_bytes_consumed, output_bytes_written).
- decode_
ssse3 ⚠ - Decodes 16 Base64 ASCII bytes to 12 raw bytes using SSSE3 (128-bit SIMD).
- encode_
avx2 ⚠ - Encodes 24 bytes to 32 Base64 ASCII bytes using AVX2. Returns (input_bytes_consumed, output_bytes_written).
- encode_
avx512 ⚠ - Encodes 48 bytes to 64 Base64 ASCII bytes using AVX-512BW. Returns (input_bytes_consumed, output_bytes_written).
- encode_
ssse3 ⚠ - Encodes 12 bytes to 16 Base64 ASCII bytes using SSSE3 (128-bit SIMD).
- has_
avx2 - Checks if the host x86_64 CPU supports AVX2.
- has_
avx512 - Checks if the host x86_64 CPU supports AVX-512 (specifically AVX-512F and AVX-512BW).
- has_
neon - Checks if the host CPU supports ARM NEON (always true on AArch64).
- has_
ssse3 - Checks if the host x86_64 CPU supports SSSE3.
- has_
wasm_ simd128 - Checks if the host CPU supports WebAssembly 128-bit SIMD.