On Wed, Apr 6, 2011 at 01:21, Eric Sandeen <sandeen@xxxxxxxxxx> wrote:
> Zeev, if you'd like to test that patch above with your profiling, that'd be awesome.
My wall-clock timings are strange and I disregard them.
On a desktop with X turned off and perf recording system-wide, I trust
the profile to show everything.
In the no-stride case, I don't think you've made it any worse.
In the stripe=stride=1, it went from:
[ perf record: Woken up 13 times to write data ]
[ perf record: Captured and wrote 3.163 MB perf.data (~138211 samples) ]
# Events: 14K cycles
#
# Overhead Command Shared Object
Symbol
# ........ .............. .....................
.......................................
#
13.08% gzip gzip [.] zip
12.65% flush-253:0 [kernel.kallsyms] [k] ext4_mb_scan_aligned
12.46% gzip gzip [.] treat_file.part.4.2264
9.34% flush-253:0 [csnappy_compress] [k] snappy_compress_fragment
5.47% md5sum md5sum [.] digest_file.isra.2.2089
4.70% md5sum [csnappy_decompress] [k]
snappy_decompress_noheader
4.29% md5sum md5sum [.] 0x2fe7
2.16% gzip libc-2.13.so [.] __memcpy_ssse3
1.79% flush-253:0 [kernel.kallsyms] [k] memcpy
1.52% tar [kernel.kallsyms] [k] copy_user_generic_string
1.08% tar [kernel.kallsyms] [k] _raw_spin_lock
1.00% gzip [kernel.kallsyms] [k] copy_user_generic_string
0.87% flush-253:0 [kernel.kallsyms] [k] __memset
0.58% md5sum md5sum [.] __libc_csu_init
0.54% md5sum [zram] [k] zram_make_request
0.50% md5sum [kernel.kallsyms] [k] copy_user_generic_string
0.45% tar [kernel.kallsyms] [k] ext4_mark_iloc_dirty
0.44% tar [kernel.kallsyms] [k] __memset
0.39% gzip gzip [.] treat_stdin.2262
0.35% gzip gzip [.] treat_file.2267
0.33% swapper [kernel.kallsyms] [k] mwait_idle
0.28% dd [kernel.kallsyms] [k] system_call
0.26% md5sum [kernel.kallsyms] [k] memcpy
0.25% flush-253:0 [zram] [k] zram_make_request
0.25% flush-253:0 [kernel.kallsyms] [k] _raw_spin_lock
0.23% tar [kernel.kallsyms] [k] ext4_mb_scan_aligned
0.22% gzip gzip [.] compress_block.2644.2190
0.20% tar [kernel.kallsyms] [k] __find_get_block
to:
[ perf record: Woken up 12 times to write data ]
[ perf record: Captured and wrote 3.091 MB perf.data (~135042 samples) ]
# Events: 13K cycles
#
# Overhead Command Shared Object
Symbol
# ........ ........... .....................
.....................................
#
15.10% gzip gzip [.] zip
14.70% gzip gzip [.] treat_file.part.4.2264
11.10% flush-253:0 [csnappy_compress] [k] snappy_compress_fragment
6.12% md5sum md5sum [.] digest_file.isra.2.2089
5.44% md5sum [csnappy_decompress] [k] snappy_decompress_noheader
5.10% md5sum md5sum [.] 0x3262
2.42% gzip libc-2.13.so [.] __memcpy_ssse3
2.03% flush-253:0 [kernel.kallsyms] [k] memcpy
1.52% tar [kernel.kallsyms] [k] copy_user_generic_string
1.13% gzip [kernel.kallsyms] [k] copy_user_generic_string
0.80% flush-253:0 [kernel.kallsyms] [k] __memset
0.62% md5sum md5sum [.] __libc_csu_init
0.60% md5sum [zram] [k] zram_make_request
0.57% gzip gzip [.] treat_stdin.2262
0.56% md5sum [kernel.kallsyms] [k] copy_user_generic_string
0.55% tar [kernel.kallsyms] [k] __memset
0.49% tar [kernel.kallsyms] [k] ext4_mark_iloc_dirty
0.39% md5sum [kernel.kallsyms] [k] memcpy
0.37% gzip gzip [.] treat_file.2267
0.37% swapper [kernel.kallsyms] [k] mwait_idle
0.28% dd [kernel.kallsyms] [k] system_call
0.28% flush-253:0 [kernel.kallsyms] [k] _raw_spin_lock
0.25% gzip gzip [.] compress_block.2644.2190
0.22% flush-253:0 [kernel.kallsyms] [k] ext4_mark_iloc_dirty
0.22% tar [kernel.kallsyms] [k] __ext4_get_inode_loc
0.22% tar [kernel.kallsyms] [k] _raw_spin_lock
0.21% tar [kernel.kallsyms] [k] mark_page_accessed
0.20% tar [kernel.kallsyms] [k] __find_get_block
0.20% tar tar [.] contains_dot_dot
So that's great, but I still don't want mke2fs to set a stripe and
stride of 1, because that's silly - no code improvement in the
allocator is ever going to be able to use that information.
> Thanks,
> -Eric
Thank you!
-Z.T.
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
