Searched refs:va0x0e (Results 1 – 10 of 10) sorted by relevance
/external/XNNPACK/src/bf16-gemm/gen/ |
D | 1x4c8-minmax-neonfma-zip.c | 106 const float32x4_t va0x0e = vreinterpretq_f32_u16(vzip1q_u16(vzero, va0x0)); in xnn_bf16_gemm_minmax_ukernel_1x4c8__neonfma_zip() local 111 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_1x4c8__neonfma_zip()
|
D | 1x4c8-minmax-neonfma-shland.c | 106 … const float32x4_t va0x0e = vreinterpretq_f32_u32(vshlq_n_u32(vreinterpretq_u32_u16(va0x0), 16)); in xnn_bf16_gemm_minmax_ukernel_1x4c8__neonfma_shland() local 111 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_1x4c8__neonfma_shland()
|
D | 2x4c8-minmax-neonfma-shland.c | 132 … const float32x4_t va0x0e = vreinterpretq_f32_u32(vshlq_n_u32(vreinterpretq_u32_u16(va0x0), 16)); in xnn_bf16_gemm_minmax_ukernel_2x4c8__neonfma_shland() local 141 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_2x4c8__neonfma_shland()
|
D | 2x4c8-minmax-neonfma-zip.c | 132 const float32x4_t va0x0e = vreinterpretq_f32_u16(vzip1q_u16(vzero, va0x0)); in xnn_bf16_gemm_minmax_ukernel_2x4c8__neonfma_zip() local 141 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_2x4c8__neonfma_zip()
|
D | 3x4c8-minmax-neonfma-shland.c | 158 … const float32x4_t va0x0e = vreinterpretq_f32_u32(vshlq_n_u32(vreinterpretq_u32_u16(va0x0), 16)); in xnn_bf16_gemm_minmax_ukernel_3x4c8__neonfma_shland() local 171 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_3x4c8__neonfma_shland()
|
D | 3x4c8-minmax-neonfma-zip.c | 158 const float32x4_t va0x0e = vreinterpretq_f32_u16(vzip1q_u16(vzero, va0x0)); in xnn_bf16_gemm_minmax_ukernel_3x4c8__neonfma_zip() local 171 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_3x4c8__neonfma_zip()
|
D | 4x4c8-minmax-neonfma-zip.c | 184 const float32x4_t va0x0e = vreinterpretq_f32_u16(vzip1q_u16(vzero, va0x0)); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_zip() local 201 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_zip()
|
D | 4x4c8-minmax-neonfma-shland.c | 184 … const float32x4_t va0x0e = vreinterpretq_f32_u32(vshlq_n_u32(vreinterpretq_u32_u16(va0x0), 16)); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_shland() local 201 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_4x4c8__neonfma_shland()
|
D | 5x4c8-minmax-neonfma-zip.c | 210 const float32x4_t va0x0e = vreinterpretq_f32_u16(vzip1q_u16(vzero, va0x0)); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_zip() local 231 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_zip()
|
D | 5x4c8-minmax-neonfma-shland.c | 210 … const float32x4_t va0x0e = vreinterpretq_f32_u32(vshlq_n_u32(vreinterpretq_u32_u16(va0x0), 16)); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_shland() local 231 vacc0x0 = vfmaq_f32(vacc0x0, va0x0e, vb0e); in xnn_bf16_gemm_minmax_ukernel_5x4c8__neonfma_shland()
|