You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

KERNEL.POWER10 7.7 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264
  1. #SGEMM_BETA = ../generic/gemm_beta.c
  2. #DGEMM_BETA = ../generic/gemm_beta.c
  3. #CGEMM_BETA = ../generic/zgemm_beta.c
  4. #ZGEMM_BETA = ../generic/zgemm_beta.c
  5. SBGEMM_BETA = ../generic/gemm_beta.c
  6. SBGEMMKERNEL = sbgemm_kernel_power10.c
  7. SBGEMMINCOPY = sbgemm_ncopy_16_power10.c
  8. SBGEMMITCOPY = sbgemm_tcopy_16_power10.c
  9. SBGEMMONCOPY = sbgemm_ncopy_8_power10.c
  10. SBGEMMOTCOPY = sbgemm_tcopy_8_power10.c
  11. SBGEMMINCOPYOBJ = sbgemm_incopy$(TSUFFIX).$(SUFFIX)
  12. SBGEMMITCOPYOBJ = sbgemm_itcopy$(TSUFFIX).$(SUFFIX)
  13. SBGEMMONCOPYOBJ = sbgemm_oncopy$(TSUFFIX).$(SUFFIX)
  14. SBGEMMOTCOPYOBJ = sbgemm_otcopy$(TSUFFIX).$(SUFFIX)
  15. STRMMKERNEL = sgemm_kernel_power10.c
  16. DTRMMKERNEL = dgemm_kernel_power10.c
  17. ifeq ($(OSNAME), AIX)
  18. CTRMMKERNEL = ctrmm_kernel_8x4_power8.S
  19. ZTRMMKERNEL = ztrmm_kernel_8x2_power8.S
  20. else
  21. CTRMMKERNEL = cgemm_kernel_power10.S
  22. ZTRMMKERNEL = zgemm_kernel_power10.S
  23. endif
  24. SGEMMKERNEL = sgemm_kernel_power10.c
  25. SGEMMINCOPY = sgemm_ncopy_16_power.c
  26. SGEMMITCOPY = sgemm_tcopy_16_power8.S
  27. SGEMMONCOPY = ../generic/gemm_ncopy_8.c
  28. SGEMMOTCOPY = sgemm_tcopy_8_power8.S
  29. SGEMMINCOPYOBJ = sgemm_incopy$(TSUFFIX).$(SUFFIX)
  30. SGEMMITCOPYOBJ = sgemm_itcopy$(TSUFFIX).$(SUFFIX)
  31. SGEMMONCOPYOBJ = sgemm_oncopy$(TSUFFIX).$(SUFFIX)
  32. SGEMMOTCOPYOBJ = sgemm_otcopy$(TSUFFIX).$(SUFFIX)
  33. SGEMM_SMALL_M_PERMIT = gemm_small_kernel_permit_power10.c
  34. SGEMM_SMALL_K_NN = sgemm_small_kernel_nn_power10.c
  35. SGEMM_SMALL_K_B0_NN = sgemm_small_kernel_nn_power10.c
  36. SGEMM_SMALL_K_NT = sgemm_small_kernel_nt_power10.c
  37. SGEMM_SMALL_K_B0_NT = sgemm_small_kernel_nt_power10.c
  38. SGEMM_SMALL_K_TN = sgemm_small_kernel_tn_power10.c
  39. SGEMM_SMALL_K_B0_TN = sgemm_small_kernel_tn_power10.c
  40. SGEMM_SMALL_K_TT = sgemm_small_kernel_tt_power10.c
  41. SGEMM_SMALL_K_B0_TT = sgemm_small_kernel_tt_power10.c
  42. DGEMMKERNEL = dgemm_kernel_power10.c
  43. DGEMMINCOPY =
  44. DGEMMITCOPY =
  45. DGEMMONCOPY = dgemm_ncopy_8_power10.c
  46. DGEMMOTCOPY = ../generic/gemm_tcopy_8.c
  47. DGEMMINCOPYOBJ =
  48. DGEMMITCOPYOBJ =
  49. DGEMMONCOPYOBJ = dgemm_oncopy$(TSUFFIX).$(SUFFIX)
  50. DGEMMOTCOPYOBJ = dgemm_otcopy$(TSUFFIX).$(SUFFIX)
  51. DGEMM_SMALL_M_PERMIT = gemm_small_kernel_permit_power10.c
  52. DGEMM_SMALL_K_NT = dgemm_small_kernel_nt_power10.c
  53. DGEMM_SMALL_K_B0_NT = dgemm_small_kernel_nt_power10.c
  54. DGEMM_SMALL_K_NN = dgemm_small_kernel_nn_power10.c
  55. DGEMM_SMALL_K_B0_NN = dgemm_small_kernel_nn_power10.c
  56. DGEMM_SMALL_K_TT = dgemm_small_kernel_tt_power10.c
  57. DGEMM_SMALL_K_B0_TT = dgemm_small_kernel_tt_power10.c
  58. DGEMM_SMALL_K_TN = dgemm_small_kernel_tn_power10.c
  59. DGEMM_SMALL_K_B0_TN = dgemm_small_kernel_tn_power10.c
  60. ifeq ($(OSNAME), AIX)
  61. CGEMMKERNEL = cgemm_kernel_8x4_power8.S
  62. else
  63. CGEMMKERNEL = cgemm_kernel_power10.S
  64. endif
  65. #CGEMMKERNEL = cgemm_kernel_8x4_power8.S
  66. CGEMMINCOPY = ../generic/zgemm_ncopy_8.c
  67. ifeq ($(OSNAME), AIX)
  68. CGEMMITCOPY = cgemm_tcopy_8_power8.S
  69. else
  70. CGEMMITCOPY = ../generic/zgemm_tcopy_8.c
  71. endif
  72. CGEMMONCOPY = ../generic/zgemm_ncopy_4.c
  73. CGEMMOTCOPY = ../generic/zgemm_tcopy_4.c
  74. CGEMMONCOPYOBJ = cgemm_oncopy$(TSUFFIX).$(SUFFIX)
  75. CGEMMOTCOPYOBJ = cgemm_otcopy$(TSUFFIX).$(SUFFIX)
  76. CGEMMINCOPYOBJ = cgemm_incopy$(TSUFFIX).$(SUFFIX)
  77. CGEMMITCOPYOBJ = cgemm_itcopy$(TSUFFIX).$(SUFFIX)
  78. ifeq ($(OSNAME), AIX)
  79. ZGEMMKERNEL = zgemm_kernel_8x2_power8.S
  80. else
  81. ZGEMMKERNEL = zgemm_kernel_power10.S
  82. endif
  83. ZGEMMONCOPY = ../generic/zgemm_ncopy_2.c
  84. ZGEMMOTCOPY = ../generic/zgemm_tcopy_2.c
  85. ZGEMMINCOPY = ../generic/zgemm_ncopy_8.c
  86. ZGEMMITCOPY = zgemm_tcopy_8_power8.S
  87. ZGEMMONCOPYOBJ = zgemm_oncopy$(TSUFFIX).$(SUFFIX)
  88. ZGEMMOTCOPYOBJ = zgemm_otcopy$(TSUFFIX).$(SUFFIX)
  89. ZGEMMINCOPYOBJ = zgemm_incopy$(TSUFFIX).$(SUFFIX)
  90. ZGEMMITCOPYOBJ = zgemm_itcopy$(TSUFFIX).$(SUFFIX)
  91. STRSMKERNEL_LN = trsm_kernel_LN_power10.c
  92. STRSMKERNEL_LT = trsm_kernel_LT_power10.c
  93. STRSMKERNEL_RN = trsm_kernel_RN_power10.c
  94. STRSMKERNEL_RT = trsm_kernel_RT_power10.c
  95. DTRSMKERNEL_LN = trsm_kernel_LN_power10.c
  96. DTRSMKERNEL_LT = trsm_kernel_LT_power10.c
  97. DTRSMKERNEL_RN = trsm_kernel_RN_power10.c
  98. DTRSMKERNEL_RT = trsm_kernel_RT_power10.c
  99. CTRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
  100. CTRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
  101. CTRSMKERNEL_RN = ../generic/trsm_kernel_RN.c
  102. CTRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
  103. ZTRSMKERNEL_LN = ../generic/trsm_kernel_LN.c
  104. ZTRSMKERNEL_LT = ../generic/trsm_kernel_LT.c
  105. ZTRSMKERNEL_RN = ../generic/trsm_kernel_RN.c
  106. ZTRSMKERNEL_RT = ../generic/trsm_kernel_RT.c
  107. #Todo: CGEMM3MKERNEL should be 4x4 blocksizes.
  108. #CGEMM3MKERNEL = zgemm3m_kernel_8x4_sse3.S
  109. #ZGEMM3MKERNEL = zgemm3m_kernel_4x4_sse3.S
  110. #Pure C for other kernels
  111. #SAMAXKERNEL = ../arm/amax.c
  112. #DAMAXKERNEL = ../arm/amax.c
  113. #CAMAXKERNEL = ../arm/zamax.c
  114. #ZAMAXKERNEL = ../arm/zamax.c
  115. #
  116. #SAMINKERNEL = ../arm/amin.c
  117. #DAMINKERNEL = ../arm/amin.c
  118. #CAMINKERNEL = ../arm/zamin.c
  119. #ZAMINKERNEL = ../arm/zamin.c
  120. #
  121. #SMAXKERNEL = ../arm/max.c
  122. #DMAXKERNEL = ../arm/max.c
  123. #
  124. #SMINKERNEL = ../arm/min.c
  125. #DMINKERNEL = ../arm/min.c
  126. #
  127. ifeq ($(C_COMPILER), GCC)
  128. ifneq ($(GCCVERSIONGTEQ9),1)
  129. ISAMAXKERNEL = isamax_power9.S
  130. else
  131. ISAMAXKERNEL = isamax.c
  132. endif
  133. IDAMAXKERNEL = idamax.c
  134. ifneq ($(GCCVERSIONGTEQ9),1)
  135. ICAMAXKERNEL = icamax_power9.S
  136. else
  137. ICAMAXKERNEL = icamax.c
  138. endif
  139. IZAMAXKERNEL = izamax.c
  140. #
  141. ifneq ($(GCCVERSIONGTEQ9),1)
  142. ISAMINKERNEL = isamin_power9.S
  143. else
  144. ISAMINKERNEL = isamin.c
  145. endif
  146. IDAMINKERNEL = idamin.c
  147. ifneq ($(GCCVERSIONGTEQ9),1)
  148. ICAMINKERNEL = icamin_power9.S
  149. else
  150. ICAMINKERNEL = icamin.c
  151. endif
  152. else
  153. ISAMAXKERNEL = isamax.c
  154. IDAMAXKERNEL = idamax.c
  155. ICAMAXKERNEL = icamax.c
  156. IZAMAXKERNEL = izamax.c
  157. ISAMINKERNEL = isamin.c
  158. IDAMINKERNEL = idamin.c
  159. ICAMINKERNEL = icamin.c
  160. endif
  161. IZAMINKERNEL = izamin.c
  162. #
  163. #ISMAXKERNEL = ../arm/imax.c
  164. #IDMAXKERNEL = ../arm/imax.c
  165. #
  166. #ISMINKERNEL = ../arm/imin.c
  167. #IDMINKERNEL = ../arm/imin.c
  168. #
  169. SASUMKERNEL = sasum.c
  170. DASUMKERNEL = dasum.c
  171. CASUMKERNEL = casum.c
  172. ZASUMKERNEL = zasum.c
  173. #
  174. SAXPYKERNEL = saxpy_power10.c
  175. DAXPYKERNEL = daxpy_power10.c
  176. CAXPYKERNEL = caxpy_power10.c
  177. ZAXPYKERNEL = zaxpy_power10.c
  178. #
  179. SCOPYKERNEL = scopy_power10.c
  180. DCOPYKERNEL = dcopy_power10.c
  181. CCOPYKERNEL = ccopy_power10.c
  182. ZCOPYKERNEL = zcopy_power10.c
  183. #
  184. SDOTKERNEL = sdot_power10.c
  185. DDOTKERNEL = ddot_power10.c
  186. DSDOTKERNEL = sdot_power10.c
  187. CDOTKERNEL = cdot.c
  188. ZDOTKERNEL = zdot.c
  189. #
  190. SNRM2KERNEL = ../arm/nrm2.c
  191. DNRM2KERNEL = ../arm/nrm2.c
  192. CNRM2KERNEL = ../arm/znrm2.c
  193. ZNRM2KERNEL = ../arm/znrm2.c
  194. #
  195. SROTKERNEL = srot.c
  196. DROTKERNEL = drot.c
  197. CROTKERNEL = crot.c
  198. ZROTKERNEL = zrot.c
  199. #
  200. SSCALKERNEL = sscal.c
  201. DSCALKERNEL = dscal.c
  202. ifeq ($(C_COMPILER), PGI)
  203. CSCALKERNEL = ../arm/zscal.c
  204. ZSCALKERNEL = ../arm/zscal.c
  205. else
  206. CSCALKERNEL = zscal.c
  207. ZSCALKERNEL = zscal.c
  208. endif
  209. #
  210. SSWAPKERNEL = sswap.c
  211. DSWAPKERNEL = dswap.c
  212. CSWAPKERNEL = cswap.c
  213. ZSWAPKERNEL = zswap.c
  214. #
  215. SGEMVNKERNEL = sgemv_n.c
  216. DGEMVNKERNEL = dgemv_n_power10.c
  217. CGEMVNKERNEL = cgemv_n.c
  218. ZGEMVNKERNEL = zgemv_n_power10.c
  219. #
  220. SGEMVTKERNEL = sgemv_t.c
  221. DGEMVTKERNEL = dgemv_t_power10.c
  222. CGEMVTKERNEL = cgemv_t.c
  223. ZGEMVTKERNEL = zgemv_t_4.c
  224. #SSYMV_U_KERNEL = ../generic/symv_k.c
  225. #SSYMV_L_KERNEL = ../generic/symv_k.c
  226. #DSYMV_U_KERNEL = ../generic/symv_k.c
  227. #DSYMV_L_KERNEL = ../generic/symv_k.c
  228. #QSYMV_U_KERNEL = ../generic/symv_k.c
  229. #QSYMV_L_KERNEL = ../generic/symv_k.c
  230. #CSYMV_U_KERNEL = ../generic/zsymv_k.c
  231. #CSYMV_L_KERNEL = ../generic/zsymv_k.c
  232. #ZSYMV_U_KERNEL = ../generic/zsymv_k.c
  233. #ZSYMV_L_KERNEL = ../generic/zsymv_k.c
  234. #XSYMV_U_KERNEL = ../generic/zsymv_k.c
  235. #XSYMV_L_KERNEL = ../generic/zsymv_k.c
  236. #ZHEMV_U_KERNEL = ../generic/zhemv_k.c
  237. #ZHEMV_L_KERNEL = ../generic/zhemv_k.c
  238. LSAME_KERNEL = ../generic/lsame.c
  239. SCABS_KERNEL = ../generic/cabs.c
  240. DCABS_KERNEL = ../generic/cabs.c
  241. QCABS_KERNEL = ../generic/cabs.c
  242. #Dump kernel
  243. CGEMM3MKERNEL = ../generic/zgemm3mkernel_dump.c
  244. ZGEMM3MKERNEL = ../generic/zgemm3mkernel_dump.c