zzh
/
sw_perf
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
将bwa的sw extend代码部分抽出来，做benchmark
20 Commits
1 Branch
0 Tags
498 KiB
C 99.1%
Makefile 0.5%
Shell 0.4%
 
 
 
Go to file
zzh b4615f890f implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
.vscode implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
.gitignore sc消融实验版本 2024-04-08 20:00:48 +08:00
Makefile sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
README.md sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
align.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
align.h implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
align_avx2_i16.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
align_avx2_u8.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
align_sse_i16.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
align_sse_u8.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
byte_alloc.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
byte_alloc.h sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
debug.c sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
debug.h sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
extend.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
extend.h sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
extend_avx2_i16.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
extend_avx2_i16_sp.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
extend_avx2_u8.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
extend_scalar.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
get_line.c 解决了读文件的bug，原因是target有些行是0，没有碱基，保留了测试代码 2023-09-03 23:59:24 +08:00
global.c sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
global.h sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
global_avx2_i16.c sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
global_avx2_u8.c sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
global_scalar.c sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
kvec.h sw benchmark，对extend部分做了全部的优化，去除target倒序，减少FIND_MAX次数 2024-04-11 13:29:28 +08:00
main.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
profiling.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
profiling.h implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
run.sh implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
utils.c implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00
utils.h implement ksw_align with avx2 instructions 2025-09-09 16:07:45 +08:00

README.md

sw_perf

  • 对bwa的sw相关代码进行抽取和优化实现，做成benchmark
  • bwa有三个部分包含sw算法，分别是seed-extension阶段，生成sam阶段的计算pair-end align分值，以及最后生成cigar的global
  • 实现avx2和cuda版本，并进行性能对比