External/cglm
1
0
Fork 0
mirror of https://github.com/recp/cglm.git synced 2025-10-03 16:51:35 +00:00
Code Issues Packages Projects Releases Wiki Activity

drop swizzling helpers fro now for simplicity

Browse Source
This commit is contained in:
Recep Aslantas
2021-04-25 02:35:55 +03:00
parent 92151c6328
commit 1fb941a41b
2 changed files with 0 additions and 49 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
CREDITS
View File

@@ -75,11 +75,6 @@ Link to paper: http://webserver2.tecgraf.puc-rio.br/~mgattass/cg/trbRR/Fast%20Mi
14. ARM NEON: Matrix Vector Multiplication
https://stackoverflow.com/a/57793352/2676533
15. ARM NEON Vector Swizzling and Permute
https://stackoverflow.com/questions/32536265/how-to-convert-mm-shuffle-ps-sse-intrinsic-to-neon-intrinsic
http://github.com/microsoft/DirectXMath
16. ARM NEON Div
http://github.com/microsoft/DirectXMath

44
include/cglm/simd/arm.h
View File

@@ -25,50 +25,6 @@
#define glmm_splat_z(x) vdupq_lane_f32(vget_high_f32(x), 0)
#define glmm_splat_w(x) vdupq_lane_f32(vget_high_f32(x), 1)
#define SWIZZLE(NAME) \
static inline float32x4_t NAME(glmm_128 v)
/* TODO:
* I'm not sure if glmm_xxxx(), glmm_yyyy()... is better than glmm_0000()...
*/
/* Memory layout Register layout (glmm)
0 1 2 3 -> 3 2 1 0
*/
SWIZZLE(glmm_0000) { return vdupq_lane_f32(vget_low_f32(v), 0); }
SWIZZLE(glmm_1111) { return vdupq_lane_f32(vget_low_f32(v), 1); }
SWIZZLE(glmm_2222) { return vdupq_lane_f32(vget_high_f32(v), 0); }
SWIZZLE(glmm_3333) { return vdupq_lane_f32(vget_high_f32(v), 1); }
SWIZZLE(glmm_2301) { return vrev64q_f32(v); }
SWIZZLE(glmm_1010) { float32x2_t vt = vget_low_f32(v); return vcombine_f32(vt, vt); }
SWIZZLE(glmm_3232) { float32x2_t vt = vget_high_f32(v); return vcombine_f32(vt, vt); }
SWIZZLE(glmm_0101) { float32x2_t vt = vrev64_f32(vget_low_f32(v)); return vcombine_f32(vt, vt); }
SWIZZLE(glmm_2323) { float32x2_t vt = vrev64_f32(vget_high_f32(v)); return vcombine_f32(vt, vt); }
SWIZZLE(glmm_2310) { return vcombine_f32(vget_low_f32(v), vrev64_f32(vget_high_f32(v))); }
SWIZZLE(glmm_3201) { return vcombine_f32(vrev64_f32(vget_low_f32(v)), vget_high_f32(v)); }
SWIZZLE(glmm_0132) { return vcombine_f32(vget_high_f32(v), vrev64_f32(vget_low_f32(v))); }
SWIZZLE(glmm_1023) { return vcombine_f32(vrev64_f32(vget_high_f32(v)), vget_low_f32(v)); }
SWIZZLE(glmm_0123) { return vcombine_f32(vrev64_f32(vget_high_f32(v)), vrev64_f32(vget_low_f32(v))); }
SWIZZLE(glmm_2200) { return vtrnq_f32(v, v).val[0]; }
SWIZZLE(glmm_3311) { return vtrnq_f32(v, v).val[1]; }
SWIZZLE(glmm_1100) { return vzipq_f32(v, v).val[0]; }
SWIZZLE(glmm_3322) { return vzipq_f32(v, v).val[1]; }
SWIZZLE(glmm_2020) { return vuzpq_f32(v, v).val[0]; }
SWIZZLE(glmm_3131) { return vuzpq_f32(v, v).val[1]; }
SWIZZLE(glmm_0321) { return vextq_f32(v, v, 1); }
SWIZZLE(glmm_1032) { return vextq_f32(v, v, 2); }
SWIZZLE(glmm_2103) { return vextq_f32(v, v, 3); }
#undef SWIZZLE
#define glmm_xor(a, b) \
vreinterpretq_f32_s32(veorq_s32(vreinterpretq_s32_f32(a), \
vreinterpretq_s32_f32(b)))