Module simd 

Raw Pointer SIMD Operations for JIT

These functions operate directly on f64 data buffers with zero boxing overhead. Signature: simd_op(ptr_a: *const f64, ptr_b: *const f64, len: u64) -> *mut f64

The JIT compiler extracts Series data pointers and lengths, then calls these functions directly for maximum performance.

Functions

jit_simd_add

SIMD-accelerated vector addition: result[i] = a[i] + b[i] Returns a newly allocated buffer that must be freed by the caller

jit_simd_add_scalar

SIMD-accelerated scalar addition: result[i] = a[i] + scalar

jit_simd_div

SIMD-accelerated vector division: result[i] = a[i] / b[i]

jit_simd_div_scalar

SIMD-accelerated scalar division: result[i] = a[i] / scalar

jit_simd_eq

SIMD-accelerated equality: result[i] = (a[i] == b[i]) ? 1.0 : 0.0

jit_simd_free

Free a SIMD result buffer allocated by jit_simd_* functions

jit_simd_gt

SIMD-accelerated greater-than: result[i] = (a[i] > b[i]) ? 1.0 : 0.0

jit_simd_gte

SIMD-accelerated greater-than-or-equal: result[i] = (a[i] >= b[i]) ? 1.0 : 0.0

jit_simd_lt

SIMD-accelerated less-than: result[i] = (a[i] < b[i]) ? 1.0 : 0.0

jit_simd_lte

SIMD-accelerated less-than-or-equal: result[i] = (a[i] <= b[i]) ? 1.0 : 0.0

jit_simd_max

SIMD-accelerated element-wise max: result[i] = max(a[i], b[i])

jit_simd_min

SIMD-accelerated element-wise min: result[i] = min(a[i], b[i])

jit_simd_mul

SIMD-accelerated vector multiplication: result[i] = a[i] * b[i]

jit_simd_mul_scalar

SIMD-accelerated scalar multiplication: result[i] = a[i] * scalar

jit_simd_neq

SIMD-accelerated inequality: result[i] = (a[i] != b[i]) ? 1.0 : 0.0

jit_simd_sub

SIMD-accelerated vector subtraction: result[i] = a[i] - b[i]

jit_simd_sub_scalar

SIMD-accelerated scalar subtraction: result[i] = a[i] - scalar