From b3e037bf1b279878277a09d60885d0ee7662ffb2 Mon Sep 17 00:00:00 2001 From: Zengtudor Date: Fri, 20 Sep 2024 10:39:14 +0800 Subject: [PATCH] update --- README.md | 41 +++++++++++++++++++++++++++++++---------- xmake.lua | 5 ++--- 2 files changed, 33 insertions(+), 13 deletions(-) diff --git a/README.md b/README.md index 50d3a92..da7a6da 100644 --- a/README.md +++ b/README.md @@ -69,23 +69,44 @@ xmake project -k vsxmake > perf ``` -Samples: 31K of event 'task-clock:ppp', Event count (approx.): 7866750000 -Overhead Command Shared Object Symbol - 90.39% test [unknown] [k] 0xffffffffa84435e1 - 5.48% test test [.] reverseComplement(std::array&, unsigned long) - 1.58% test [unknown] [k] 0xffffffffc06abd30 - 0.59% test [unknown] [k] 0xffffffffa83ab787 - 0.51% test [unknown] [k] 0xffffffffa842aee0 + 31.88% test test [.] reverseComplement(char*, char*) [clone ._omp_fn.1] + 10.27% test [unknown] [k] 0xffffffffc06abd30 + 7.32% test libgomp.so.1.0.0 [.] 0x0000000000024c6a + 4.39% test [unknown] [k] 0xffffffffa84435e1 + 3.32% test libgomp.so.1.0.0 [.] 0x0000000000024ab2 + 3.20% test [unknown] [k] 0xffffffffa8443ee5 + 2.35% test [unknown] [k] 0xffffffffa72d138b + 2.30% test [unknown] [k] 0xffffffffa7309ed4 + 1.98% test [unknown] [k] 0xffffffffa7a5ba37 + 1.74% test [unknown] [k] 0xffffffffa760ecee + 1.63% test test [.] reverseComplement(char*, char*) [clone ._omp_fn.0] + 1.48% test [unknown] [k] 0xffffffffa83ab787 + 1.32% test [unknown] [k] 0xffffffffa842aee0 + 1.31% test [unknown] [k] 0xffffffffa766d76a + 0.83% test [unknown] [k] 0xffffffffa7a5ad88 + 0.80% test libc.so.6 [.] __memset_evex_unaligned_erms + 0.55% test [unknown] [k] 0xffffffffa766d747 + 0.47% test [unknown] [k] 0xffffffffa76c08f8 ``` > 800MB fastq DNA 序列处理性能展示 ``` -[Timer: All spent] Start timing Open input file stream to value [input_file_stream] ok , from ["filteredReads.txt"] Open output file stream to value [output_file_stream] ok , from ["reversedSequence.txt"] -Undergoing transformation -[Timer: All spent] Stop timing , using 5960ms +Chunk size :4294967296 bytes +[Timer: All spent] Start timing +[Timer: chunk_id:[1]] Start timing +[Timer: read_chunk_id:[1]] Start timing +[Timer: read_chunk_id:[1]] Stop timing , used 1102ms +buf_len : 897963094 +[Timer: calculate_chunk_id:[1]] Start timing +omp_get_num_threads() : 12 +[Timer: calculate_chunk_id:[1]] Stop timing , used 463ms +[Timer: write_chunk_id:[1] , [Wrote bytes] start_pos : 897963094] Start timing +[Timer: write_chunk_id:[1] , [Wrote bytes] start_pos : 897963094] Stop timing , used 1287ms +[Timer: chunk_id:[1]] Stop timing , used 2854ms +[Timer: All spent] Stop timing , used 2855ms ``` ## 关于版权 diff --git a/xmake.lua b/xmake.lua index 3ff5ff1..132ed5e 100644 --- a/xmake.lua +++ b/xmake.lua @@ -6,11 +6,10 @@ if is_mode("release")then -- set_optimize("fastest") --上面和下面的二选一,--表示注释 if is_plat("windows")then - -- 注意,下面是启用AVX512指令集矢量化加速,只有新2010年后的CPU支持,但是可以大大加快批量指令加速,可以试试将下面的--删除 - -- 但是似乎这个程序目前任然是IO密集型,性能差不多,目前注释掉了 + -- 矢量化加速似乎没用 -- add_cxxflags("/arch:AVX512") elseif is_plat("linux") or is_plat("mingw") or is_plat("clang")then - -- 启用AVX512指令集矢量化加速,可能会导致无法在除了编译本程序的电脑上运行,出现兼容问题 + -- 矢量化加速似乎没用 -- add_cxxflags("-march=native") end end