Don't know if this is a concern. The patch was built & tested with intel mesa CI and locally. Does this need attention or can be ignored? Please advise.
-Yogesh. >From: mesa-dev [mailto:[email protected]] On >Behalf Of AppVeyor >Sent: Tuesday, November 21, 2017 4:23 AM >To: [email protected] >Subject: [Mesa-dev] [AppVeyor] mesa master #6216 failed > >Build mesa 6216 failed >Commit 971b3c019b by Aravindan Muthukumar on 11/9/2017 5:45 AM: >i965: Optimize bucket index calculation\n\nReducing Bucket index >calculation to O(1).\n\nThis algorithm calculates the index using >matrix method. Assuming\nPAGE_SIZE is 4096, matrix arrangement is as >below:\n\n 1*4096 2*4096 3*4096 4*4096\n 5*4096 6*4096 7*4096 8*4096\n >10*4096 12*4096 14*4096 16*4096\n 20*4096 24*4096 28*4096 32*4096\n ... >... ... ...\n ... ... ... ...\n ... ... ... max_cache_size\n\nFrom this >matrix its clearly seen that every row follows the below way:\n\n ... >... ... n\n n+(1/4)n n+(1/2)n n+(3/4)n 2n\n\nRow is calculated as >log2(size/PAGE_SIZE) Column is calculated as\nconverting the difference >between the elements to fit into power size of\ntwo and indexing >it.\n\nFinal Index is (row*4)+(col-1)\n\nTested with Intel Mesa >CI.\n\nImproves performance of 3DMark on BXT by 0.705966% +/- 0.229767% >(n=20)\n\nv4: Review comments on style and code comments implemented >(Ian).\nv3: Review comments implemented (Ian).\nv2: Review comments >implemented (Jason).\n\nSigned-off-by: Aravindan Muthukumar ><[email protected]>\nSigned-off-by: Kedar Karanje ><[email protected]>\nReviewed-by: Yogesh Marathe ><[email protected]>\nSigned-off-by: Ian Romanick ><[email protected]> Configure your notification preferences > _______________________________________________ mesa-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/mesa-dev
