Created
March 31, 2013 16:47
-
-
Save ssvb/5281249 to your computer and use it in GitHub Desktop.
Branch prediction rate doing 'dd if=/dev/sda of=/dev/null' (with the same usb flash) for Atom, Cortex-A8 and Cortex-A15
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
======= Intel Atom N450 ========== | |
atom ~ # perf stat dd if=/dev/sdb of=/dev/null bs=1M | |
7839+1 records in | |
7839+1 records out | |
8220834816 bytes (8.2 GB) copied, 411.443 s, 20.0 MB/s | |
Performance counter stats for 'dd if=/dev/sdb of=/dev/null bs=1M': | |
35503.879851 task-clock # 0.086 CPUs utilized | |
251,314 context-switches # 0.007 M/sec | |
1,042 CPU-migrations # 0.029 K/sec | |
507 page-faults # 0.014 K/sec | |
54,831,835,725 cycles # 1.544 GHz | |
<not supported> stalled-cycles-frontend | |
<not supported> stalled-cycles-backend | |
16,841,062,310 instructions # 0.31 insns per cycle | |
3,226,037,494 branches # 90.864 M/sec | |
152,386,299 branch-misses # 4.72% of all branches | |
411.461097543 seconds time elapsed | |
======= Allwinner A10 (Mele A2000) ======== | |
sun4i ~ # perf stat dd if=/dev/sda of=/dev/null bs=1M | |
7839+1 records in | |
7839+1 records out | |
8220834816 bytes (8.2 GB) copied, 441.431 s, 18.6 MB/s | |
Performance counter stats for 'dd if=/dev/sda of=/dev/null bs=1M': | |
60620.000000 task-clock # 0.137 CPUs utilized | |
295066 context-switches # 0.005 M/sec | |
0 CPU-migrations # 0.000 K/sec | |
466 page-faults # 0.008 K/sec | |
49514491451 cycles # 0.817 GHz | |
7468409937 stalled-cycles-frontend # 15.08% frontend cycles idle | |
<not supported> stalled-cycles-backend | |
18304965462 instructions # 0.37 insns per cycle | |
# 0.41 stalled cycles per insn | |
2050618041 branches # 33.827 M/sec | |
213548800 branch-misses # 10.41% of all branches | |
441.446989758 seconds time elapsed | |
perf record -e branch-misses dd if=/dev/sda of=/dev/null bs=1M count=1000 && perf report -n | |
# Events: 110K branch-misses | |
# | |
# Overhead Samples Command Shared Object Symbol | |
# ........ .......... ....... ................. ..................................... | |
# | |
5.12% 5206 dd [kernel.kallsyms] [k] blkdev_get_block | |
4.28% 4477 dd [kernel.kallsyms] [k] bio_attempt_back_merge | |
4.24% 4158 dd [kernel.kallsyms] [k] get_request | |
4.12% 4437 dd [kernel.kallsyms] [k] generic_make_request_checks | |
4.05% 4457 dd [kernel.kallsyms] [k] __slab_alloc.isra.58.constprop.63 | |
4.04% 3652 dd [kernel.kallsyms] [k] get_page_from_freelist | |
3.70% 4404 dd [kernel.kallsyms] [k] submit_bh | |
3.66% 4176 dd [kernel.kallsyms] [k] kmem_cache_alloc | |
3.58% 3788 dd [kernel.kallsyms] [k] block_read_full_page | |
3.44% 3820 dd [kernel.kallsyms] [k] sub_preempt_count | |
3.43% 3021 dd [kernel.kallsyms] [k] scsi_request_fn | |
3.37% 4177 dd [kernel.kallsyms] [k] add_preempt_count | |
3.00% 3212 dd [kernel.kallsyms] [k] __do_page_cache_readahead | |
2.53% 2613 dd [kernel.kallsyms] [k] create_empty_buffers | |
2.10% 3365 dd [kernel.kallsyms] [k] __blk_recalc_rq_segments | |
1.97% 1965 dd [kernel.kallsyms] [k] __copy_to_user_std | |
1.93% 2195 dd [kernel.kallsyms] [k] add_to_page_cache_locked | |
1.85% 3159 dd [kernel.kallsyms] [k] ll_back_merge_fn | |
1.82% 1918 dd [kernel.kallsyms] [k] generic_file_aio_read | |
1.79% 2906 dd [kernel.kallsyms] [k] finish_task_switch.constprop.75 | |
1.66% 2094 dd [kernel.kallsyms] [k] drive_stat_acct | |
1.59% 1977 dd [kernel.kallsyms] [k] blk_queue_bio | |
1.58% 1612 dd [kernel.kallsyms] [k] __memzero | |
1.34% 1448 dd [kernel.kallsyms] [k] alloc_buffer_head | |
1.18% 1117 dd [kernel.kallsyms] [k] put_page | |
1.06% 1380 dd [kernel.kallsyms] [k] mempool_alloc | |
0.93% 1604 dd [kernel.kallsyms] [k] __rcu_read_unlock | |
0.89% 998 dd [kernel.kallsyms] [k] bio_alloc_bioset | |
0.88% 1903 dd [kernel.kallsyms] [k] generic_make_request | |
0.78% 866 dd [kernel.kallsyms] [k] radix_tree_lookup_element | |
0.76% 768 dd [kernel.kallsyms] [k] __alloc_pages_nodemask | |
0.74% 771 dd [kernel.kallsyms] [k] pagevec_lru_move_fn | |
0.70% 717 dd [kernel.kallsyms] [k] prepare_to_wait_exclusive | |
========= Exynos 5250 (ARM Chromebook) ============= | |
a15 ~ # perf stat dd if=/dev/sda of=/dev/null bs=1M | |
7839+1 records in | |
7839+1 records out | |
8220834816 bytes (8.2 GB) copied, 350.353 s, 23.5 MB/s | |
Performance counter stats for 'dd if=/dev/sda of=/dev/null bs=1M': | |
27822.467420 task-clock # 0.079 CPUs utilized | |
252294 context-switches # 0.009 M/sec | |
31 CPU-migrations # 0.001 K/sec | |
465 page-faults # 0.017 K/sec | |
45755749191 cycles # 1.645 GHz | |
<not supported> stalled-cycles-frontend | |
<not supported> stalled-cycles-backend | |
20143388608 instructions # 0.44 insns per cycle | |
3972022678 branches # 142.763 M/sec | |
180404037 branch-misses # 4.54% of all branches | |
350.364623292 seconds time elapsed | |
perf record -e branch-misses dd if=/dev/sda of=/dev/null bs=1M count=1000 && perf report -n | |
# Events: 10K branch-misses | |
# | |
# Overhead Samples Command Shared Object Symbol | |
# ........ .......... ....... ................. ............................... | |
# | |
27.06% 2928 dd [kernel.kallsyms] [k] _raw_spin_unlock_irq | |
6.48% 703 dd [kernel.kallsyms] [k] blkdev_max_block | |
5.58% 605 dd [kernel.kallsyms] [k] kmem_cache_alloc | |
5.24% 571 dd [kernel.kallsyms] [k] generic_make_request_checks | |
3.78% 393 dd [kernel.kallsyms] [k] _test_and_set_bit | |
2.80% 285 dd [kernel.kallsyms] [k] get_page_from_freelist | |
2.39% 265 dd [kernel.kallsyms] [k] block_read_full_page | |
2.33% 274 dd [kernel.kallsyms] [k] _set_bit | |
2.06% 228 dd [kernel.kallsyms] [k] __copy_to_user_std | |
2.00% 164 dd [kernel.kallsyms] [k] put_page | |
1.97% 147 dd [kernel.kallsyms] [k] _clear_bit | |
1.70% 178 dd [kernel.kallsyms] [k] __blk_recalc_rq_segments | |
1.65% 170 dd [kernel.kallsyms] [k] _raw_spin_unlock_irqrestore | |
1.42% 164 dd [kernel.kallsyms] [k] bio_put | |
1.36% 140 dd [kernel.kallsyms] [k] __slab_alloc.constprop.49 | |
1.18% 123 dd [kernel.kallsyms] [k] _cond_resched | |
1.14% 128 dd [kernel.kallsyms] [k] blk_queue_bio | |
1.13% 93 dd [kernel.kallsyms] [k] do_raw_spin_unlock | |
1.11% 119 dd [kernel.kallsyms] [k] generic_file_aio_read | |
1.08% 109 dd [kernel.kallsyms] [k] alloc_buffer_head | |
1.02% 97 dd [kernel.kallsyms] [k] recalc_bh_state | |
1.00% 103 dd [kernel.kallsyms] [k] drive_stat_acct | |
1.00% 72 dd [kernel.kallsyms] [k] free_hot_cold_page | |
0.98% 92 dd [kernel.kallsyms] [k] find_get_page | |
0.95% 106 dd [kernel.kallsyms] [k] submit_bio | |
0.79% 82 dd [kernel.kallsyms] [k] radix_tree_lookup_element | |
0.78% 85 dd [kernel.kallsyms] [k] page_address | |
0.76% 63 dd [kernel.kallsyms] [k] _test_and_clear_bit | |
0.70% 69 dd [kernel.kallsyms] [k] submit_bh | |
0.63% 55 dd [kernel.kallsyms] [k] bio_alloc | |
0.61% 54 dd [kernel.kallsyms] [k] __alloc_pages_nodemask | |
0.60% 59 dd [kernel.kallsyms] [k] blkdev_get_block |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment