Release 0.54.0

CHANGELOG.md

@@ -2,6 +2,20 @@
 
 ## [Unreleased]
 
+## [0.54.0] - 2026-09-06
+
+Headline: **every `ExecutionContext` gets a `Schedule` — and the CI run that exercised it found a
+real deadlock.** `sk.ainet.context.schedule.Schedule` splits *what* an op computes from *how its
+independent chunks spread across cores*: `scaledDotProductAttention` is the first scheduled op,
+`parallelChunks` no longer hides a `runBlocking(Dispatchers.Default)` island, and a JVM
+`CoroutineSchedule` spreads chunks across a shared pool. Turning it on deadlocked
+`skainet-backend-cpu:jvmTest` on CI's 4-vCPU runner for three of the last four `test (jvm)` runs —
+not the OOM a first pass assumed, but a `coroutineScope` waiting on children the pool had no thread
+left to run. A region is now a shared chunk queue instead: whatever the pool is doing, the caller
+can always finish its own region alone. Also in this release: `SafeTensorsParametersLoader`'s
+`tensorFilter` reaches parity with the sharded loader, and the `sk.ainet.lang.memory` API drops its
+`ExperimentalMemoryApi` opt-in gate now that SKEEP-003's M0–M2 have shipped.
+
 ### Added
 
 - **Schedules — the compute-level algorithm/schedule split (SKEEP-005)**
@@ -40,6 +54,23 @@
   complete in 0.49.0. `ExperimentalMemoryApi` is deleted along with every `@OptIn`/
   `@ExperimentalMemoryApi` annotation referencing it.
 
+### Fixed
+
+- **`CoroutineSchedule` deadlock when a region is entered from its own pool**
+  ([#1262](https://github.com/SKaiNET-developers/SKaiNET/issues/1262),
+  [#1264](https://github.com/SKaiNET-developers/SKaiNET/issues/1264),
+  [#1266](https://github.com/SKaiNET-developers/SKaiNET/pull/1266)): `CoroutineSchedule.forRange`
+  ran a region as `runBlocking { coroutineScope { launch(Dispatchers.Default) … } }`, and a
+  `coroutineScope` waits for every child — including ones the pool never got a thread for. Once
+  every `Dispatchers.Default` worker was itself inside a region (routine on a 4-vCPU CI runner;
+  never reproduced on a many-core laptop), nobody was left to run the children and the JVM parked
+  forever — the intermittent `test (jvm)` timeout that #1264's "force fully serial" fix mistook for
+  an OOM hang. A region is now a shared chunk queue: `tasks - 1` helpers dispatch to the pool, the
+  caller runs chunk 0 and then drains the queue itself, and waits only for chunks a thread has
+  already claimed — a caller can always finish its own region alone, whatever the pool is doing.
+  Contract unchanged (first failure wins, nested regions run inline, writes happen-before return);
+  public API unchanged.
+
 ## [0.53.0] - 2026-09-02
 
 Headline: **the export pipeline emits billion-parameter models.** Tracing a 4.5B-parameter

README.md

@@ -51,7 +51,7 @@ Add the core dependencies (Gradle Kotlin DSL):
 ```kotlin
 dependencies {
     // Recommended: import the umbrella BOM and drop versions on the engine modules.
-    implementation(platform("sk.ainet:skainet-bom:0.53.0"))
+    implementation(platform("sk.ainet:skainet-bom:0.54.0"))
 
     implementation("sk.ainet.core:skainet-lang-core")
     implementation("sk.ainet.core:skainet-backend-cpu")
@@ -308,27 +308,26 @@ val withoutLabel = dataPipeline<RawDataset>()
 
 ---
 
-## What's New in 0.53.0
-
-The export pipeline emits billion-parameter models:
-
-- **Full Gemma 3n E2B export** — the DSL → tape → StableHLO path traced a 4.5B-parameter model
-  into an OOM at a 46 GB heap; the same repro now exports the whole model in under a minute.
-  Shape-only tracing no longer materializes zero buffers, graph constants alias the live weights
-  instead of copying them, and ≥2 GiB constants travel as `BufferHandle.Floats` — an aliased
-  float array — because the tied embedding is exactly one byte over what a JVM byte array holds.
-- **Conversion fails loudly** — `StableHloConverter` is strict by default: an unconvertible node
-  throws `HloConversionException`, an unresolved operand throws `MissingOperandException`, and a
-  packed weight reaching constant extraction throws `PackedConstantException` instead of silently
-  becoming a function argument. `ConversionErrorPolicy.LENIENT` restores the old
-  comment-and-continue behavior for inspection.
-- **Sharded SafeTensors in the engine** — `ShardedSafeTensorsParametersLoader` consumes
-  `model.safetensors.index.json` with the single-file loader's BF16/FP16 policies, a fail-fast
-  dtype pre-scan, and a `tensorFilter` hook; the per-family hand-rolled loaders downstream can
-  collapse onto it.
-- **Registry gaps the strictness surfaced** — `clamp` and the camelCase `indexSelect` the tracer
-  actually emits now lower, and `createBasic` registers the neural-net converter like
-  `createExtended` does.
+## What's New in 0.54.0
+
+Structured concurrency lands as a first-class citizen — and the CI run that exercised it found a
+real deadlock:
+
+- **`Schedule` on every `ExecutionContext`** (SKEEP-005) — `sk.ainet.context.schedule.Schedule`
+  splits *what* an op computes from *how its independent chunks spread across cores*.
+  `scaledDotProductAttention` is the first scheduled op, `parallelChunks` no longer hides a
+  `runBlocking(Dispatchers.Default)` island, and the JVM `CoroutineSchedule.hardware()` default
+  spreads chunks across cores while `Schedule.Sequential` keeps every kernel single-threaded.
+- **A real deadlock, found by turning scheduling on** — `CoroutineSchedule.forRange`'s region
+  waited on children the pool had no thread left to run, once every worker was itself inside a
+  region; it looked like the OOM hang a first CI fix assumed. A region is now a shared chunk
+  queue, so a caller can always finish its own region alone, whatever the pool is doing.
+- **`tensorFilter` on the single-file `SafeTensorsParametersLoader`** — parity with the sharded
+  loader; lets a family load selectively from a checkpoint that carries tensors the requested
+  dtype can't accept.
+- **`ExperimentalMemoryApi` opt-in gate removed** — SKEEP-003's M0–M2 shipped complete back in
+  0.49.0, so `Storage`, `Scope`, `Format`, `TensorView`, `WeightForm`, and the rest of
+  `sk.ainet.lang.memory` no longer need `@OptIn`.
 
 See [CHANGELOG.md](CHANGELOG.md) for full release notes, including every prior release.
 
@@ -354,6 +353,13 @@ We love contributions! Whether it's a new operator, documentation, or a bug fix:
 
 Browse the full codebase documentation on [DeepWiki](https://deepwiki.com/SKaiNET-developers/SKaiNET).
 
+### Contributors (0.54.0)
+
+- **Michal Harakal** ([@michalharakal](https://github.com/michalharakal)) — the SKEEP-005
+  structured-concurrency Schedule API, the coroutine-pool deadlock it uncovered on CI and its
+  shared-chunk-queue fix, `SafeTensorsParametersLoader` `tensorFilter` parity, and retiring the
+  `ExperimentalMemoryApi` opt-in gate now that SKEEP-003 has shipped
+
 ### Contributors (0.53.0)
 
 - **Michal Harakal** ([@michalharakal](https://github.com/michalharakal)) — the billion-parameter

docs/antora.yml

@@ -15,7 +15,7 @@ asciidoc:
     framework_name: SKaiNET
     # Current SKaiNET release — bump once per release; referenced as
     # {skainet_version} in dependency snippets (blocks need subs="attributes+").
-    skainet_version: 0.53.0
+    skainet_version: 0.54.0
     ksp_version: 2.2.21-2.0.5
     dokka_version: 2.1.0
     asciidoctorj_version: 3.0.0

docs/modules/ROOT/pages/reference/kernel-support-matrix.adoc

@@ -1,7 +1,7 @@
 = Kernel × platform support matrix
 :description: Which compute-kernel provider serves each weight format on each KMP target.
 
-Generated from `kernel-support.json` (version `0.53.0`) by `KernelSupportMatrixTest` — registry introspection of the registered `KernelProvider` implementations. Do not edit by hand; run `./gradlew generateKernelMatrix` to refresh.
+Generated from `kernel-support.json` (version `0.54.0`) by `KernelSupportMatrixTest` — registry introspection of the registered `KernelProvider` implementations. Do not edit by hand; run `./gradlew generateKernelMatrix` to refresh.
 
 Each cell is the best (highest-priority) provider that serves `Float32 × format` `matmul` on that platform: *native-ffm* (100) → *panama-vector* (50) → *scalar* (0). An empty cell (`—`) means no provider carries a kernel there (the format is dequant-to-FP32 only).
 

docs/modules/ROOT/pages/reference/operators/generated/index.adoc

@@ -1,6 +1,6 @@
 = AI-NET Operators Reference
 
-Generated from version `0.53.0` on 2026-09-02
+Generated from version `0.54.0` on 2026-09-06
 
 == Operators by Modality
 

docs/modules/ROOT/pages/reference/ops-status-matrix.adoc

@@ -1,7 +1,7 @@
 = Operator Coverage Matrix
 :description: Cross-backend status for every operator function in SKaiNET.
 
-Generated from `operators.json` version `0.53.0` on 2026-09-02.
+Generated from `operators.json` version `0.54.0` on 2026-09-06.
 
 Rows are `Operator.function` pairs. The `Validated` column shows whether the function's documentation has been DARC-validated by a reviewer (see xref:contributing/darc-workflow.adoc[DARC workflow]). Remaining columns are backends that appear in any function's `statusByBackend` map — a missing entry means the backend makes no claim about the function (treat it as "unknown", not "not supported").
 

docs/modules/ROOT/pages/tutorials/android-classifier-getting-started.adoc

@@ -19,12 +19,12 @@ An Android project with Kotlin. Add the SKaiNET modules:
 [source,kotlin]
 ----
 dependencies {
-    implementation("sk.ainet.core:skainet-lang-core:0.53.0")     // tensors, DSL, training
-    implementation("sk.ainet.core:skainet-backend-cpu:0.53.0")   // CPU ops
-    implementation("sk.ainet.core:skainet-compile-dag:0.53.0")   // autograd (training context)
-    implementation("sk.ainet.core:skainet-data-api:0.53.0")      // Dataset / DataBatch
-    implementation("sk.ainet.core:skainet-data-simple:0.53.0")   // embedded Iris
-    runtimeOnly("sk.ainet.core:skainet-backend-jni-cpu:0.53.0")  // NEON kernels (see below)
+    implementation("sk.ainet.core:skainet-lang-core:0.54.0")     // tensors, DSL, training
+    implementation("sk.ainet.core:skainet-backend-cpu:0.54.0")   // CPU ops
+    implementation("sk.ainet.core:skainet-compile-dag:0.54.0")   // autograd (training context)
+    implementation("sk.ainet.core:skainet-data-api:0.54.0")      // Dataset / DataBatch
+    implementation("sk.ainet.core:skainet-data-simple:0.54.0")   // embedded Iris
+    runtimeOnly("sk.ainet.core:skainet-backend-jni-cpu:0.54.0")  // NEON kernels (see below)
 }
 ----
 

gradle.properties

@@ -1,5 +1,5 @@
 GROUP=sk.ainet.core
-VERSION_NAME=0.53.0
+VERSION_NAME=0.54.0
 POM_DESCRIPTION=SKaiNET
 
 POM_URL=https://github.com/SKaiNET-developers/skainet/