[lkml]   [2017]   [Jan]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH v4 0/9] mm/swap: Regular page swap optimizations
Hi, Minchan,

Minchan Kim <> writes:
> The patchset has used several techniqueus to reduce lock contention, for example,
> batching alloc/free, fine-grained lock and cluster distribution to avoid cache
> false-sharing. Each items has different complexity and benefits so could you
> show the number for each step of pathchset? It would be better to include the
> nubmer in each description. It helps how the patch is important when we consider
> complexitiy of the patch.

Here is the test data.

We test the vm-scalability swap-w-seq test case with 32 processes on a
Xeon E5 v3 system. The swap device used is a RAM simulated PMEM
(persistent memory) device. To test the sequential swapping out, the
test case created 32 processes, which sequentially allocate and write to
the anonymous pages until the RAM and part of the swap device is used

The patchset is rebased on v4.9-rc8. So the baseline performance is as

"": 1428002,
"perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list": 13.94,
"perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg": 13.75,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.swap_info_get.swapcache_free.__remove_mapping.shrink_page_list": 7.05,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.swap_info_get.page_swapcount.try_to_free_swap.swap_writepage": 7.03,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.__swap_duplicate.swap_duplicate.try_to_unmap_one.rmap_walk_anon": 7.02,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_swap_page.add_to_swap.shrink_page_list.shrink_inactive_list": 6.83,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.page_check_address_transhuge.page_referenced_one.rmap_walk_anon.rmap_walk": 0.81,

>> Patch 1 is a clean up patch.
> Could it be separated patch?
>> Patch 2 creates a lock per cluster, this gives us a more fine graind lock
>> that can be used for accessing swap_map, and not lock the whole
>> swap device

After patch 2, the result is as follow,

"": 1481704,
"perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list": 27.53,
"perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg": 27.01,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.free_pcppages_bulk.drain_pages_zone.drain_pages.drain_local_pages": 1.03,

The swap out throughput is at the same level, but the lock contention on
swap_info_struct->lock is eliminated.

>> Patch 3 splits the swap cache radix tree into 64MB chunks, reducing
>> the rate that we have to contende for the radix tree.

After patch 3,

"": 2050097,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_swap_page.add_to_swap.shrink_page_list.shrink_inactive_list": 43.27,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_vma.handle_mm_fault": 4.84,

The swap out throughput is improved about ~43% compared with baseline.
The lock contention on swap cache radix tree lock is eliminated.
swap_info_struct->lock in get_swap_page() becomes the most heavy
contended lock.

>> Patch 4 eliminates unnecessary page allocation for read ahead.
> Could it be separated patch?
>> Patch 5-9 create a per cpu cache of the swap slots, so we don't have
>> to contend on the swap device to get a swap slot or to release
>> a swap slot. And we allocate and release the swap slots
>> in batches for better efficiency.

After patch 9,

"": 4170746,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.swapcache_free_entries.free_swap_slot.free_swap_and_cache.unmap_page_range": 13.91,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_page_from_freelist.__alloc_pages_nodemask.alloc_pages_vma.handle_mm_fault": 8.56,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_page_from_freelist.__alloc_pages_slowpath.__alloc_pages_nodemask.alloc_pages_vma": 2.56,
"perf-profile.calltrace.cycles-pp._raw_spin_lock.get_swap_pages.get_swap_page.add_to_swap.shrink_page_list": 2.47,

The swap out throughput is improved about 192% compared with the
baseline. There are still some lock contention for
swap_info_struct->lock, but the pressure begins to shift to buddy system

Best Regards,
Huang, Ying

 \ /
  Last update: 2017-01-05 02:34    [W:0.344 / U:0.032 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site