Home
last modified time | relevance | path

Searched refs:batchreduce_kernelb (Results 1 – 21 of 21) sorted by relevance

/dports/math/libxsmm/libxsmm-1.16.3/src/template/
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_kcck_core_bf16.tpl.c210 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, diB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
218 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dciB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
226 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dfB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
234 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dpB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
250 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, diB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
254 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dciB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
258 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dfB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
262 batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, dpB, ikb, 0, 0, 0, 0, nBlocks, bn_lp, bk, lpb),
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck_diffused_bf16.tpl.c118batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ri, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
131batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rc, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
144batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rf, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
157batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ro, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
230batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ri, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
243batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rc, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
256batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rf, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
269batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ro, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
H A Dlibxsmm_dnn_rnncell_st_rnn_fwd_nc_kcck.tpl.c43 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
112batchreduce_kernelb(A_array2, B_array2, &LIBXSMM_VLA_ACCESS(3, z, i, in*bn, ik*bk, N, K), &blocks);
121batchreduce_kernelb(A_array2, B_array2, &LIBXSMM_VLA_ACCESS(3, z, i, in*bn, ik*bk, N, K), &blocks);
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck_fused_bf16.tpl.c56batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ri, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
84batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rc, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
112batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rf, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
140batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ro, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
227batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ri, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
255batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rc, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
283batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, rf, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
311batchreduce_kernelb(&LIBXSMM_VLA_ACCESS(5, ro, ikb, CB*KB_BLOCKS, 0, 0, 0, kBlocks, bk_lp, bk, lpb…
H A Dlibxsmm_dnn_rnncell_st_rnn_fwd_ncnc_kcck.tpl.c43 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
152batchreduce_kernelb(&A_array2[ii][jj][0], &B_array2[ii][jj][0], &LIBXSMM_VLA_ACCESS(5, z, i, in, i…
209batchreduce_kernelb(A_array2, B_array2, &LIBXSMM_VLA_ACCESS(5, z, i, in, ik, 0, 0, nBlocks, kBlock…
218batchreduce_kernelb(A_array2, B_array2, &LIBXSMM_VLA_ACCESS(5, z, i, in, ik, 0, 0, nBlocks, kBlock…
H A Dlibxsmm_dnn_rnncell_st_gru_fwd_nc_kcck.tpl.c61 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
138 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, i, j, in, ik, N, K), &blocks);
163 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, c, j, in, ik, N, K), &blocks);
201 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, f, j, in, ik, N, K), &blocks);
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck_fused.tpl.c57 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, i, j, in, ik, N, K), &blocks);
100 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, ci, j, in, ik, N, K), &blocks);
143 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, f, j, in, ik, N, K), &blocks);
186 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, o, j, in, ik, N, K), &blocks);
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck_diffused.tpl.c130 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, i, j, in, ik, N, K), &blocks);
155 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, ci, j, in, ik, N, K), &blocks);
180 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, f, j, in, ik, N, K), &blocks);
204 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, o, j, in, ik, N, K), &blocks);
H A Dlibxsmm_dnn_rnncell_st_gru_fwd_nc_ck_generic.tpl.c76 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
201 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, i, j, in, ik, N, K), &blocks);
226 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, c, j, in, ik, N, K), &blocks);
264 batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(3, f, j, in, ik, N, K), &blocks);
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_kcck_core.tpl.c356batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, dri, ikb, icb, 0, 0, kBlocks, bk, bk)…
368batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drc, ikb, icb, 0, 0, kBlocks, bk, bk)…
380batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drf, ikb, icb, 0, 0, kBlocks, bk, bk)…
392batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, dro, ikb, icb, 0, 0, kBlocks, bk, bk)…
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck.tpl.c75 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
H A Dlibxsmm_dnn_rnncell_st_gru_bwdupd_nc_kcck.tpl.c118 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_a…
525batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, dri, ikb, icb, 0, 0, kBlocks, bk, bk)…
537batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drc, ikb, icb, 0, 0, kBlocks, bk, bk)…
549batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drf, ikb, icb, 0, 0, kBlocks, bk, bk)…
H A Dlibxsmm_dnn_rnncell_st_gru_bwdupd_nc_ck_generic.tpl.c133 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_a…
505batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, dri, ikb, icb, 0, 0, kBlocks, bk, bk)…
517batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drc, ikb, icb, 0, 0, kBlocks, bk, bk)…
529batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, drf, ikb, icb, 0, 0, kBlocks, bk, bk)…
H A Dlibxsmm_dnn_rnncell_st_rnn_bwdupd_nc_kcck.tpl.c71 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
383batchreduce_kernelb(A_array, B_array, &LIBXSMM_VLA_ACCESS(4, dr, ikb, icb, 0, 0, kBlocks, bk, bk),…
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_kcck_bf16.tpl.c136 const libxsmm_bsmmfunction_reducebatch_strd batchreduce_kernelb = handle->fwd_kernelb; variable
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_ck_generic.tpl.c93 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
H A Dlibxsmm_dnn_rnncell_st_lstm_fwd_nc_ck_generic_bf16.tpl.c152 const libxsmm_bsmmfunction_reducebatch_strd batchreduce_kernelb = handle->fwd_kernelb; variable
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_kcck.tpl.c147 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_ck_generic.tpl.c160 const libxsmm_smmfunction_reducebatch_addr batchreduce_kernelb = libxsmm_smmdispatch_reducebatch_ad… variable
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_ck_generic_bf16.tpl.c177 const libxsmm_bsmmfunction_reducebatch_strd batchreduce_kernelb = handle->bwdupd_kernelb; variable
H A Dlibxsmm_dnn_rnncell_st_lstm_bwdupd_nc_kcck_bf16.tpl.c182 const libxsmm_bsmmfunction_reducebatch_strd batchreduce_kernelb = handle->bwdupd_kernelb; variable