The hierarchy now uses a request barrier after the first synchronous request completion. Async requests could occasionally complete fast enough that DispatchRequests would be left to handle the completion. In that scenario, there would not be a barrier in place when beginning any subsequent requests, which would lead to a failed assertion in the request owner.
#preflight 63ceb5c63a03cb0bbff95362
#rb Zousar.Shaker
[CL 23848713 by Devin Doucette in ue5-main branch]
Without using StopGetStore, get hits on the in-flight cache would propagate as puts to subsequent nodes in the hierarchy. Since the in-flight cache is only used as temporary storage while a put is in flight, this caused duplicates of the in-flight puts, which manifests as move collision log spam from the file system cache store.
#preflight 625d772d691f49969ea93249
#rb Zousar.Shaker
#rnx
[CL 19785647 by Devin Doucette in ue5-main branch]
Without using StopStore, get hits on the in-flight cache would propagate as puts to subsequent nodes in the hierarchy. Since the in-flight cache is only used as temporary storage while a put is in flight, this caused duplicates of the in-flight puts, which manifests as move collision log spam from the file system cache store.
#preflight 625835d97f628a9018d8acf6
#rb Zousar.Shaker
#rnx
[CL 19757142 by Devin Doucette in ue5-main branch]
The wrong response status was being checked, and caused subsequent nodes to be skipped even if the StopStore node did not contain the data, if any previous node contained the data.
#preflight 62582d67946114248db62d58
#rb Zousar.Shaker
#rnx
[CL 19756375 by Devin Doucette in ue5-main branch]
- Legacy puts are now executing asynchronously.
- Memory cache now merges partial cache records.
- Memory cache is used to store data temporarily while it is being written by an async put.
- Expanded the deprecation of persisted boot/memory caches, which no longer offer the performance benefit that they used to.
- Fixed the pak file cache to skip data with a compressed size of over 2 GiB.
#jira UE-141307
#preflight 620d85f93609e19371510fb1
#lockdown Aurel.Cordonnier
#rb Zousar.Shaker
#rnx
#ROBOMERGE-OWNER: Devin.Doucette
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-COMMAND: _robomerge UE5-Main
#ROBOMERGE-SOURCE: CL 19076205 in //UE5/Release-5.0/... via CL 19094550
#ROBOMERGE-BOT: UE5 (Release-Engine-Staging -> Main) (v921-19075845)
[CL 19095974 by devin doucette in ue5-main branch]
The cache stores can individually control their "legacy mode" now, which offers a choice of ValueOnly, ValueWithLegacyFallback, LegacyOnly.
- Using ValueOnly will forward every legacy request through the Value API, which compresses values by default and supports values larger than 2 GiB.
- Using ValueWithLegacyFallback behaves like ValueOnly, but misses from GetValue are forwarded to GetCachedData, and stored with PutValue if found.
- Using LegacyOnly will forward every legacy request through the Legacy API.
FLegacyCacheValue uses shared state to ensure that each value is compressed or decompressed at most once.
#jira UE-134381
#preflight 62054b430c64e1822f41231b
#lockdown Mark.Lintott
#rb Zousar.Shaker
#rnx
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18940270 in //UE5/Release-5.0/... via CL 18941016 via CL 18941392
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v917-18934589)
[CL 18941401 by devin doucette in ue5-main branch]
Also adds the intended inline allocators for arrays that were missing them.
#jira UE-134381
#lockdown Mark.Lintott
#preflight 61fd998ff370b0d3111fad17
#rb Matt.Peters
#rnx
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18874174 in //UE5/Release-5.0/... via CL 18874208 via CL 18874924
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v910-18824042)
[CL 18874951 by devin doucette in ue5-main branch]
The async cache hierarchy:
- Is required to add compression to the legacy cache by forwarding LegacyPut/LegacyGet to PutValue/GetValue.
- Is always present in the graph, unlike the previous cache hierarchy, which will allow significant simplification of leaf cache stores.
- Allows for the leaf cache store nodes to operate asynchronously without blocking a worker thread like the previous cache hierarchy.
- Shifts from controlling cache behavior by speed class to controlling cache behavior by local/remote classification, which is a critical distinction for a cache like Zen that can identify as both local and remote.
- Respects the local/remote and query/store flags in the cache policy.
- Does not fully implement the partial record cache policy at this time.
- Does not propagate records or values in GetChunks, which will be added in a future release.
The verify wrapper was previously missing an implementation for the new cache interface. This version is more efficient than the previous because requests through the new cache interface can compare data without loading it from storage, and load it only when a mismatch has been detected, to dump it to disk.
#jira UE-134381
#lockdown Mark.Lintott
#preflight 61fc32ac0a50c2606f266388
#rb Zousar.Shaker
#rnx
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18851861 in //UE5/Release-5.0/... via CL 18851943 via CL 18852169
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v910-18824042)
[CL 18852184 by devin doucette in ue5-main branch]
Includes policy fixes for other functions.
#rb Zousar.Shaker
#rnx
#preflight 61f9b31c9e4d23cd93b1dad3
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18817404 in //UE5/Release-5.0/... via CL 18817411 via CL 18822862
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v910-18824042)
[CL 18824693 by devin doucette in ue5-main branch]
When a node with StopStore is present and contains the data, but comes later in the hierarchy than the node that first found the data, puts to subsequent nodes were not being disabled.
#rb Zousar.Shaker
#rnx
#preflight 61f433b16b5aea38e5bf86ca
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18778611 in //UE5/Release-5.0/... via CL 18780649 via CL 18780924
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v903-18687472)
[CL 18781004 by devin doucette in ue5-main branch]
This new cache hierarchy does not require the cache stores within it to complete every operation synchronously. When the child cache stores start to do async I/O, this hierarchy will support that instead of blocking like the existing hierarchy does.
This is a snapshot of the work in progress that is only implemented for the functions in ILegacyCacheStore.
#rb Zousar.Shaker
#rnx
#preflight 61f37e487a7f0d39ddd2d38b
#ROBOMERGE-AUTHOR: devin.doucette
#ROBOMERGE-SOURCE: CL 18768547 in //UE5/Release-5.0/... via CL 18768556 via CL 18768581
#ROBOMERGE-BOT: UE5 (Release-Engine-Test -> Main) (v903-18687472)
[CL 18768594 by devin doucette in ue5-main branch]
