/external/XNNPACK/src/f32-argmaxpool/ |
D | 9x-neon-c4.c | 86 uint32x4_t vidx = vmovq_n_u32(0); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 90 vidx = vbslq_u32(vm1, vmovq_n_u32(1), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 94 vidx = vbslq_u32(vm2, vmovq_n_u32(2), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 98 vidx = vbslq_u32(vm3, vmovq_n_u32(3), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 102 vidx = vbslq_u32(vm4, vmovq_n_u32(4), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 106 vidx = vbslq_u32(vm5, vmovq_n_u32(5), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 110 vidx = vbslq_u32(vm6, vmovq_n_u32(6), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 114 vidx = vbslq_u32(vm7, vmovq_n_u32(7), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 118 vidx = vbslq_u32(vm8, vmovq_n_u32(8), vidx); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() 135 uint32x4_t vidx = vmovq_n_u32(0); in xnn_f32_argmaxpool_ukernel_9x__neon_c4() [all …]
|
D | 4x-neon-c4.c | 56 uint32x4_t vidx = vmovq_n_u32(0); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 60 vidx = vbslq_u32(vm1, vmovq_n_u32(1), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 64 vidx = vbslq_u32(vm2, vmovq_n_u32(2), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 68 vidx = vbslq_u32(vm3, vmovq_n_u32(3), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 80 uint32x4_t vidx = vmovq_n_u32(0); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 84 vidx = vbslq_u32(vm1, vmovq_n_u32(1), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 88 vidx = vbslq_u32(vm2, vmovq_n_u32(2), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4() 92 vidx = vbslq_u32(vm3, vmovq_n_u32(3), vidx); in xnn_f32_argmaxpool_ukernel_4x__neon_c4()
|
D | 9p8x-neon-c4.c | 67 uint32x4_t vidx = vmovq_n_u32(0); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 71 vidx = vbslq_u32(vm1, vmovq_n_u32(1), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 75 vidx = vbslq_u32(vm2, vmovq_n_u32(2), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 79 vidx = vbslq_u32(vm3, vmovq_n_u32(3), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 83 vidx = vbslq_u32(vm4, vmovq_n_u32(4), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 87 vidx = vbslq_u32(vm5, vmovq_n_u32(5), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 91 vidx = vbslq_u32(vm6, vmovq_n_u32(6), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 95 vidx = vbslq_u32(vm7, vmovq_n_u32(7), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 99 vidx = vbslq_u32(vm8, vmovq_n_u32(8), vidx); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() 105 const uint32x4_t v1 = vmovq_n_u32(1); in xnn_f32_argmaxpool_ukernel_9p8x__neon_c4() [all …]
|
/external/XNNPACK/src/f32-vrnd/gen/ |
D | vrndu-neon-x8.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndu_ukernel__neon_x8() 43 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 44 vrndmask4567 = vbicq_u32(vrndmask4567, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 55 vadjmask0123 = vorrq_u32(vadjmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 56 vadjmask4567 = vorrq_u32(vadjmask4567, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 69 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 73 vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 82 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8() 86 vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x8()
|
D | vrndu-neon-x4.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndu_ukernel__neon_x4() 39 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x4() 47 vadjmask0123 = vorrq_u32(vadjmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x4() 58 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x4() 62 vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndu_ukernel__neon_x4()
|
D | vrndz-neon-x8.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndz_ukernel__neon_x8() 42 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x8() 43 vrndmask4567 = vbicq_u32(vrndmask4567, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x8() 56 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x8() 65 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x8()
|
D | vrndne-neon-x8.c | 28 const float32x4_t vmagic_number = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndne_ukernel__neon_x8() 41 vrndmask0123 = vorrq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x8() 42 vrndmask4567 = vorrq_u32(vrndmask4567, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x8() 58 vrndmask = vorrq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x8() 68 vrndmask = vorrq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x8()
|
D | vrndd-neon-x8.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndd_ukernel__neon_x8() 43 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x8() 44 vrndmask4567 = vbicq_u32(vrndmask4567, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x8() 66 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x8() 78 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x8()
|
D | vrndz-neon-x4.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndz_ukernel__neon_x4() 38 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x4() 49 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndz_ukernel__neon_x4()
|
D | vrndne-neon-x4.c | 28 const float32x4_t vmagic_number = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndne_ukernel__neon_x4() 37 vrndmask0123 = vorrq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x4() 50 vrndmask = vorrq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndne_ukernel__neon_x4()
|
D | vrndd-neon-x4.c | 28 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); in xnn_f32_vrndd_ukernel__neon_x4() 39 vrndmask0123 = vbicq_u32(vrndmask0123, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x4() 56 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); in xnn_f32_vrndd_ukernel__neon_x4()
|
/external/XNNPACK/src/f32-vrnd/ |
D | vrndu-neon.c.in | 27 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); 43 vrndmask${ABC[N:N+4]} = vbicq_u32(vrndmask${ABC[N:N+4]}, vmovq_n_u32(UINT32_C(0x80000000))); 55 vadjmask${ABC[N:N+4]} = vorrq_u32(vadjmask${ABC[N:N+4]}, vmovq_n_u32(UINT32_C(0x80000000))); 69 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); 73 vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000))); 82 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); 86 vadjmask = vorrq_u32(vadjmask, vmovq_n_u32(UINT32_C(0x80000000)));
|
D | vrndz-neon.c.in | 27 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); 42 vrndmask${ABC[N:N+4]} = vbicq_u32(vrndmask${ABC[N:N+4]}, vmovq_n_u32(UINT32_C(0x80000000))); 56 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); 65 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000)));
|
D | vrndne-neon.c.in | 27 const float32x4_t vmagic_number = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); 40 vrndmask${ABC[N:N+4]} = vorrq_u32(vrndmask${ABC[N:N+4]}, vmovq_n_u32(UINT32_C(0x80000000))); 57 vrndmask = vorrq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); 67 vrndmask = vorrq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000)));
|
D | vrndd-neon.c.in | 27 const float32x4_t vintegral_threshold = vreinterpretq_f32_u32(vmovq_n_u32(UINT32_C(0x4B000000))); 43 vrndmask${ABC[N:N+4]} = vbicq_u32(vrndmask${ABC[N:N+4]}, vmovq_n_u32(UINT32_C(0x80000000))); 66 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000))); 78 vrndmask = vbicq_u32(vrndmask, vmovq_n_u32(UINT32_C(0x80000000)));
|
/external/XNNPACK/src/math/ |
D | roundd-neon-cvt.c | 26 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundd__neon_cvt() 28 const uint32x4_t vone = vmovq_n_u32(UINT32_C(0x3F800000)); in xnn_math_f32_roundd__neon_cvt()
|
D | roundd-neon-addsub.c | 28 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundd__neon_addsub() 30 const uint32x4_t vone = vmovq_n_u32(UINT32_C(0x3F800000)); in xnn_math_f32_roundd__neon_addsub()
|
D | roundz-neon-addsub.c | 28 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundz__neon_addsub() 31 const uint32x4_t vone = vmovq_n_u32(UINT32_C(0x3F800000)); in xnn_math_f32_roundz__neon_addsub()
|
D | roundz-neon-cvt.c | 26 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundz__neon_cvt()
|
D | roundne-neon-addsub.c | 28 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundne__neon_addsub()
|
D | roundu-neon-cvt.c | 26 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundu__neon_cvt()
|
D | roundu-neon-addsub.c | 28 const uint32x4_t vsign_mask = vmovq_n_u32(UINT32_C(0x80000000)); in xnn_math_f32_roundu__neon_addsub()
|
/external/llvm-project/clang/test/CodeGen/ |
D | arm_neon_intrinsics.c | 8195 return vmovq_n_u32(a); in test_vmovq_n_u32()
|
/external/clang/test/CodeGen/ |
D | arm_neon_intrinsics.c | 10076 return vmovq_n_u32(a); in test_vmovq_n_u32()
|
/external/neon_2_sse/ |
D | NEON_2_SSE.h | 1697 _NEON2SSESTORAGE uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0 12348 _NEON2SSESTORAGE uint32x4_t vmovq_n_u32(uint32_t value); // VDUP.32 q0,r0 12349 #define vmovq_n_u32 vdupq_n_u32 macro
|