Dynamic eviction strategy #57
Draft
Conversation
This file contains hidden or 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
It's implementation is mostly based on PosixShmSegment. Also, extend ShmManager and ShmSegmentOpts to support this new segment type.
After introducing file segment type, nameToKey_ does not provide enough information to recover/remove segments on restart. This commit fixes that by replacing nameToKey_ with nameToOpts_. Previously, the Key from nameToKey_ map was only used in a single DCHECK().
* New class MemoryTierCacheConfig allows to configure a memory tier. Setting tier size and location of a file for file-backed memory are supported in this initial implementation; * New member, vector of memory tiers, is added to class CacheAllocatorConfig. * New test suite, chelib/allocator/tests/MemoryTiersTest.cpp, demonstrates the usage of and tests extended config API.
to allow using new configureMemoryTiers() API with legacy behavior. Move validation code for memory tiers to validate() method and convert ratios to sizes lazily (on get)..
It wrongly assumed that the only possible segment type is PosixSysV segment.
… handling when NVM cache state is not enabled
Now it's size is 8 bytes intead of 4. Original CompressedPtr stored only some offset with a memory Allocator. For multi-tier implementation, this is not enough. We must also store tierId and when uncompressing, select a proper allocator. An alternative could be to just resign from CompressedPtr but they are leveraged to allow the cache to be mapped to different addresses on shared memory. Changing CompressedPtr impacted CacheItem size - it increased from 32 to 44 bytes.
Without this fix removeCb called even in case when Item is moved between tiers.
It fails because CentOS is EOL. We might want to consider using CentOS Streams but for now, just remove it. Right now, we rely on build-cachelib-centos workflow anyway.
Disabled test suite allocator-test-AllocatorTypeTest to skip sporadically failing tests.
…m#43) Compensation results in ratios being different than originially specified.
Do not use shallow clone to make sure Docker rebuild logic works correctly.
Added required packages to install Intel ittapi
Remove the item from mmContainer and drop the lock before attempting eviction. Use moving bit for synchronization in findEviction moving bit is used to give exclusive right to evict the item to a particular thread. Originially, there was an assumption that whoever marked the item as moving will try to free it until he succeeds. Since we don't want to do that in findEviction (potentially can take a long time) we need to make sure that unmarking is safe. This patch checks the flags after unmarking (atomically) and if ref is zero it also recyles the item. This is needed as there might be some concurrent thread releasing the item (and decrementing ref count). If moving bit is set, that thread would not free the memory back to allocator, resulting in memory leak on unmarkMoving().
Get tier id of item before calling any function on allocator (which needs the tierID).
Fix slab release code
…etero Reduce critical section in findEviction (for heterogenous memory cache)
The main purpose of this patch is to better simulate workloads in cachebench. Setting touchValue to true allows to see performance impact of using different mediums for memory cache.
Extend cachbench with touch value
* pmem#75: Use actual tier sizes (rounded down to slab size and decreased by header size) when creating new memory pools * Added getPoolSize method to calculate combined pool size for all tiers; added pool size validation to tests * Explicitly specified type for totalCacheSize to avoid overflow * Minor test change * Reworked tests * Minor change * Deleted redundant tests * Deleted unused constant * First set of changes to cache configuration API to enable multi-tier caches (facebook#138) Summary: These changes introduce per-tier cache configuration required to implement features discussed here: facebook#102. These specific changes enable single DRAM tier configs only which are compatible with the current version of cachelib. Configuration API will be expanded as multi-tier changes in other parts of the library are introduced. Pull Request resolved: facebook#138 Reviewed By: therealgymmy Differential Revision: D36189766 Pulled By: jiayuebao fbshipit-source-id: 947aa0cd800ea6accffc1b7b6b0c9693aa7fc0a5 Co-authored-by: Victoria McGrath <[email protected]>
and add additional parameters to control allocation and eviction of items.
instead of relying on transparent synchronization mechanims (wait context).
updated stats map
expose p99
byrnedj
force-pushed
the
develop
branch
2 times, most recently
from
f6ad2ae to
09d7bab
Compare
byrnedj
force-pushed
the
develop
branch
2 times, most recently
from
09d7bab to
ebfca17
Compare
byrnedj
force-pushed
the
develop
branch
4 times, most recently
from
9ba4e79 to
c1ff397
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
7 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I'm planning to archive pmem/CacheLib and I'm moving all potentially useful PRs to intel/CacheLib.
Ref: pmem#92
This change is