Module jit 

Structs

ExternFnSig

Extern function signature info for FFI

JitCompiler

JIT Compiler for Sigil

SigilValue

Runtime value representation

Functions

sigil_abs

sigil_ackermann

Iterative Ackermann function using explicit stack Much faster than recursive version - no stack overflow, O(result) space

sigil_add

sigil_array_add

Element-wise add two arrays into a new array (SIMD-friendly)

sigil_array_choice

Get random element of array (χ morpheme)

sigil_array_dot

Dot product of two arrays (SIMD-friendly)

sigil_array_fill

Fill array with a value (SIMD-friendly)

sigil_array_first

Get first element of array (α morpheme)

sigil_array_get

sigil_array_last

Get last element of array (ω morpheme)

sigil_array_len

sigil_array_max

Find maximum value in array (SIMD-friendly)

sigil_array_middle

Get middle element of array (μ morpheme)

sigil_array_min

Find minimum value in array (SIMD-friendly)

sigil_array_mul

Element-wise multiply two arrays into a new array (SIMD-friendly)

sigil_array_new

sigil_array_next

Get next element (iterator advance) - currently returns first element (ξ morpheme)

sigil_array_nth

Get nth element of array (ν morpheme) - same as sigil_array_get but clearer semantics

sigil_array_offset

Add a scalar to all elements (in-place, SIMD-friendly)

sigil_array_product

Product of all elements in array (Π morpheme)

sigil_array_push

sigil_array_scale

Multiply all elements by a scalar (in-place, SIMD-friendly)

sigil_array_set

sigil_array_sort

Sort array in ascending order (σ morpheme) - returns new sorted array

sigil_array_sum

Sum all elements in an array using SIMD-friendly loop

sigil_ceil

sigil_cos

sigil_div

sigil_exp

sigil_floor

sigil_ge

sigil_gpu_filter

GPU filter operation with parallel stream compaction Uses scan-based compaction algorithm

sigil_gpu_map

GPU map operation - would compile to WGSL/SPIR-V compute shader Shader structure:

sigil_gpu_reduce

GPU reduce operation - uses tree reduction in shared memory Achieves O(log n) parallel steps

sigil_gt

sigil_le

sigil_ln

sigil_lt

sigil_memo_free

Free a memoization cache

sigil_memo_get_1

Lookup a single-argument function result in cache Returns the cached value, or MEMO_NOT_FOUND if not found

sigil_memo_get_2

Lookup a two-argument function result in cache

sigil_memo_new

Create a new memoization cache

sigil_memo_set_1

Store a single-argument function result in cache

sigil_memo_set_2

Store a two-argument function result in cache

sigil_mul

sigil_now

sigil_parallel_filter

Parallel filter operation - filters elements in parallel Uses parallel predicate evaluation with stream compaction

sigil_parallel_map

Parallel map operation - applies a transformation in parallel across array elements For now, returns the array unchanged as full closure parallelization requires more complex infrastructure. In production, this would:

sigil_parallel_reduce

Parallel reduce operation - tree reduction for associative operations Achieves O(log n) depth with O(n) work

sigil_pow

sigil_print

sigil_print_float

sigil_print_int

sigil_print_str

sigil_simd_add

SIMD add - uses AVX when available

sigil_simd_cross

SIMD cross product (3D, ignores w component)

sigil_simd_div

SIMD divide

sigil_simd_dot

SIMD dot product (returns scalar) - optimized for auto-vectorization

sigil_simd_extract

Extract element from SIMD vector

sigil_simd_free

Free SIMD vector (for memory management)

sigil_simd_hadd

SIMD horizontal add (sum all lanes)

sigil_simd_length

SIMD length - uses FMA for length calculation

sigil_simd_length_sq

SIMD length squared - uses FMA for better performance

sigil_simd_max

SIMD max - element-wise maximum

sigil_simd_min

SIMD min - element-wise minimum

sigil_simd_mul

SIMD multiply

sigil_simd_new

Create a new Vec4 SIMD vector

sigil_simd_normalize

SIMD normalize - fast reciprocal sqrt pattern

sigil_simd_splat

Create Vec4 by splatting a scalar to all lanes

sigil_simd_sub

SIMD subtract

sigil_sin

sigil_sqrt

sigil_sub

sigil_tak

Iterative Tak (Takeuchi) function using explicit stack