-
Notifications
You must be signed in to change notification settings - Fork 0
Pull requests: bugparty/cpu_math_kernels_pri
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
⚡ Thunderbolt: Softmax — 8x unrolled AVX2 with split-ln2 exp256
#78
opened Jul 15, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — 8x Unrolling and FMA-optimized exp256
#77
opened Jul 14, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — 8x unrolled AVX2 Softmax with single-FMA exp256
#76
opened Jul 12, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — AVX2 8x Unrolled Softmax
#75
opened Jul 10, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — 8x unroll and single-FMA range reduction
#74
opened Jul 9, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — 8x unrolled AVX2 single-FMA exp256
#73
opened Jul 8, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — 8x unroll via single-FMA range reduction
#72
opened Jul 6, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — 8x unroll with single-FMA exp
#71
opened Jul 5, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — 8x unroll and single-FMA exp approximation
#70
opened Jul 4, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — Single-FMA range reduction + 8x unrolling
#69
opened Jul 3, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — 8x unroll via single-FMA exp256 optimization
#68
opened Jul 1, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — AVX2 8x unroll with single-FMA exp256
#67
opened Jun 30, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — 8x Unrolled Single-FMA exp256
#66
opened Jun 28, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — AVX2 8x Unrolling and single FMA exp256
#65
opened Jun 27, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — 8x Unrolling + Single-FMA Range Reduction
#64
opened Jun 26, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax AVX2 — FMA exp256 reduction and 8x unrolling
#63
opened Jun 25, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: max_v4 — 16x unrolled AVX2 max reduction
#61
opened Jun 23, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — Single-FMA range reduction optimization
#60
opened Jun 22, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — AVX2 single-FMA exp256 optimization
#59
opened Jun 21, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — Fused FMA transcendental reduction for exp256
#58
opened Jun 20, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — Heterogeneous Unrolling (8x-4x-8x)
#57
opened Jun 19, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: softmax_v6 — Heterogeneous AVX2 Loop Unrolling
#56
opened Jun 17, 2026 by
bugparty
Owner
Loading…
⚡ Thunderbolt: Softmax — single FMA range reduction
#55
opened Jun 16, 2026 by
bugparty
Owner
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.