Skip to content

Steam Machine and SteamOS support - #6201

Merged
shai-almog merged 6 commits into
masterfrom
steam-machine-support
Nov 10, 2026
Merged

shai-almog merged 6 commits into
masterfrom
steam-machine-support

Conversation

@shai-almog

Copy link
Copy Markdown
Collaborator

Ship Codename One games and apps to the Steam Machine, the Steam Deck and Linux players on Steam.

  • Controllers everywhere: a generic Gamepad/Gamepads API. Any form is usable from a couch with no code: D-pad and stick move focus, A presses, B goes back. Linux reads controllers directly from the input devices; Windows feeds its Xbox controller input into the same API.
  • Steamworks without an SDK in the tree: GameServices gives achievements, stats, the overlay, cloud saves and Steam's floating keyboard. The Linux port loads libsteam_api.so at run time; every other port answers "unavailable".
  • Runs in the Steam Linux Runtime: no direct libGLESv2 link, plus an OpenSSL 1.1 fallback. CI runs the Linux demo inside Valve's runtime image.
  • Depot packaging: set steam.appId, run cn1:buildSteamLocal or cn1:buildSteam, and get a depot with Valve's upload scripts; cn1:steamUpload uploads it with your own SteamCMD.
  • New developer-guide chapter: Publishing to Steam.

Companion PRs: BuildDaemon (packaging and SteamCMD upload on the build servers) and BuildCloud (Steam as a submission store).

🤖 Generated with Claude Code

shai-almog and others added 3 commits November 9, 2026 22:04
- Generic controller and game-services API in com.codename1.gaming
  (Gamepad, Gamepads, GameServices); a controller drives focus on any form.
- Linux port: evdev gamepads with rumble, fullscreen, Steamworks through the
  flat API loaded at run time (achievements, stats, overlay, cloud saves,
  floating keyboard). No Steamworks file is committed.
- Runs in the Steam Linux Runtime: no direct libGLESv2 link, OpenSSL 1.1
  fallback for crypto.
- Windows: Xbox controller input feeds the same API.
- Steam depot packaging (steam.* hints), cn1:buildSteamLocal, cn1:buildSteam,
  cn1:steamUpload and the Gradle tasks; Steam chapter in the developer guide.
- CI runs the Linux demo inside the Steam runtime image.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…public

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…g tests

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

Developer Guide build artifacts are available for download from this workflow run:

Developer Guide quality checks:

  • AsciiDoc linter: No issues found (report)
  • Vale: No alerts found (report)
  • Paragraph capitalization: No paragraph capitalization issues (report)
  • LanguageTool: No grammar matches (report)
  • Image references: No unused images detected (report)

@github-actions

Copy link
Copy Markdown
Contributor

Cloudflare Preview

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

✅ Continuous Quality Report

Test & Coverage

Static Analysis

  • SpotBugs [Report archive]
    • ✅ ByteCodeTranslator: 0 findings (no issues)
    • ✅ android: 0 findings (no issues)
    • ✅ android-compat: 0 findings (no issues)
    • ✅ android-res-compiler: 0 findings (no issues)
    • ✅ backend: 0 findings (no issues)
    • ✅ backend-test: 0 findings (no issues)
    • ✅ build-engine: 0 findings (no issues)
    • ✅ build-hint-catalog: 0 findings (no issues)
    • ✅ build-hint-tools: 0 findings (no issues)
    • ✅ codenameone-gradle-plugin: 0 findings (no issues)
    • ✅ codenameone-maven-plugin: 0 findings (no issues)
    • ✅ compose-compat: 0 findings (no issues)
    • ✅ core-unittests: 0 findings (no issues)
    • ✅ desktop-compat-runtime: 0 findings (no issues)
    • ✅ fxml-compiler: 0 findings (no issues)
    • ✅ ios: 0 findings (no issues)
    • ✅ javac: 0 findings (no issues)
    • ✅ javafx-compat: 0 findings (no issues)
    • ✅ project-model: 0 findings (no issues)
    • ✅ python-runtime: 0 findings (no issues)
    • ✅ python-transpiler: 0 findings (no issues)
    • ✅ swing-compat: 0 findings (no issues)
  • ✅ PMD: 0 findings (no issues) [Report archive]
  • ✅ Checkstyle: 0 findings (no issues) [Report archive]

Generated automatically by the PR CI workflow.

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 231 screenshots: 231 matched.
✅ JavaScript-port screenshot tests passed.

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

✅ ByteCodeTranslator Quality Report

Test & Coverage

  • ✅ Tests: 933 total, 0 failed, 55 skipped

Benchmark Results

  • Execution Time: 8634 ms

  • Hotspots (Top 20 sampled methods):

    • 10.57% java.lang.System.identityHashCode (136 samples)
    • 5.98% java.util.ArrayList.indexOf (77 samples)
    • 5.75% java.lang.Object.hashCode (74 samples)
    • 5.28% com.codename1.tools.translator.BytecodeMethod.equals (68 samples)
    • 2.80% java.util.AbstractCollection.containsAll (36 samples)
    • 2.72% java.lang.StringBuilder.append (35 samples)
    • 2.18% com.codename1.tools.translator.JavascriptReachability.enqueueResolved (28 samples)
    • 1.94% java.lang.String.equals (25 samples)
    • 1.79% com.codename1.tools.translator.ByteCodeClass.markDependencies (23 samples)
    • 1.63% com.codename1.tools.translator.classfile.analysis.Frame.execute (21 samples)
    • 1.63% com.codename1.tools.translator.ByteCodeClass.fillVirtualMethodTable (21 samples)
    • 1.55% com.codename1.tools.translator.classfile.analysis.Frame.merge (20 samples)
    • 1.48% java.io.FileOutputStream.writeBytes (19 samples)
    • 1.48% java.util.HashMap.hash (19 samples)
    • 1.32% com.codename1.tools.translator.bytecodes.Invoke.resolveDirectTarget (17 samples)
    • 1.32% com.codename1.tools.translator.Parser.cullMethods (17 samples)
    • 1.09% com.codename1.tools.translator.bytecodes.Invoke.findMethodUp (14 samples)
    • 1.01% java.io.FileOutputStream.open0 (13 samples)
    • 0.93% com.codename1.tools.translator.ByteCodeClass.generateCCode (12 samples)
    • 0.85% com.codename1.tools.translator.ReachabilityCull.run (11 samples)
  • ⚠️ Coverage report not generated.

Static Analysis

  • ✅ SpotBugs: no findings (report was not generated by the build).
  • ⚠️ PMD report not generated.
  • ⚠️ Checkstyle report not generated.

Generated automatically by the PR CI workflow.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 196 screenshots: 196 matched.
✅ Native Mac screenshot tests passed.

Benchmark Results

  • VM Translation Time: 0 seconds
  • Compilation Time: 760 seconds

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 54ms / native 3ms = 18.0x speedup
SIMD float-mul (64K x300) java 54ms / native 3ms = 18.0x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 native bridge unavailable (CN1 + SIMD + image benchmarks only)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path active (NEON-accelerated)
Base64 CN1 encode 52.000 ms
Base64 CN1 decode 54.000 ms
Image encode benchmark iterations 100
Image createMask (SIMD off) 7.000 ms
Image createMask (SIMD on) 4.000 ms
Image createMask ratio (SIMD on/off) 0.571x (42.9% faster)
Image applyMask (SIMD off) 71.000 ms
Image applyMask (SIMD on) 54.000 ms
Image applyMask ratio (SIMD on/off) 0.761x (23.9% faster)
Image modifyAlpha (SIMD off) 46.000 ms
Image modifyAlpha (SIMD on) 57.000 ms
Image modifyAlpha ratio (SIMD on/off) 1.239x (23.9% slower)
Image modifyAlpha removeColor (SIMD off) 49.000 ms
Image modifyAlpha removeColor (SIMD on) 45.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 0.918x (8.2% faster)

ParparVM vs HotSpot (JDK 25): macOS arm64

Runner CPU: Apple M1 (Virtual) (baseline macos-arm64)

Ratios are ParparVM / JDK 25: below 1.00x ParparVM is faster (time) or smaller (RAM). Median of 5 interleaved, paired rounds; every run's output was verified. A ratio more than 20% (time) / 25% (RAM) above its baseline in vm/selfhost/perf-baseline/ is a regression and fails; one that far below is an improvement, reported but not failed (a row whose calibration runs were noisier carries a wider tolerance), and for RAM the change must also exceed 0.05x in absolute terms. Both run unpinned on all of the runner's CPUs, with their own default thread counts.

Benchmark Cores Time RAM Status
hello (HelloCodenameOne) 3 1.02x (base 0.96x, +6.8%) 0.88x (base 0.93x, -4.5%) ok
translator (self) 3 0.66x (base 0.62x, +6.2%) 0.68x (base 0.78x, -13.1%) ok
intArithmetic 3 1.02x (base 1.03x, -0.8%) 0.12x (base 0.12x, +0.5%) ok
longArithmetic 3 0.99x (base 1.03x, -3.2%) 0.12x (base 0.07x, +65.0%) ok
mathTranscendental 3 0.96x (base 1.01x, -4.7%) 0.12x (base 0.12x, +1.8%) ok
arraySequential 3 0.44x (base 0.42x, +4.0%) 0.53x (base 0.53x, -0.2%) ok
arrayRandom 3 0.94x (base 0.99x, -5.4%) 0.33x (base 0.32x, +0.5%) ok
objectAllocation 3 1.00x (base 1.11x, -9.7%) 0.70x (base 0.72x, -2.8%) ok
valueEscape 3 0.51x (base 0.51x, -0.3%) 0.08x (base 0.09x, -9.9%) ok
hashMapChurn 3 1.11x (base 1.18x, -5.9%) 0.07x (base 0.07x, +1.3%) ok
stringBuilding 3 0.90x (base 0.84x, +6.5%) 0.39x (base 0.46x, -15.1%) ok
recursion 3 1.22x (base 1.24x, -0.8%) 0.12x (base 0.12x, +0.5%) ok
quicksort 3 0.99x (base 1.01x, -2.4%) 0.11x (base 0.11x, -0.3%) ok

Result: no regression

shai-almog and others added 2 commits November 10, 2026 07:56
glibc 2.38 headers rename sscanf to __isoc23_sscanf, which the glibc 2.28
link target lacks; the arm64 suite failed to link. Also a contraction in
the steam.branch hint and gamepad in the guide's accepted words.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 195 screenshots: 195 matched.

Native Android coverage

  • 📊 Line coverage: 19.18% (56169/292885 lines covered) [HTML preview] (artifact android-coverage-report, jacocoAndroidReport/html/index.html)
    • Other counters: instruction 18.13% (270599/1492353), branch 12.38% (19678/158943), complexity 10.36% (15848/153017), method 15.74% (11411/72518), class 27.93% (2247/8045)
    • Lowest covered classes
      • kotlin.collections.unsigned.kotlin.collections.unsigned.UArraysKt___UArraysKt – 0.00% (0/2396 lines covered)
      • org.jacoco.agent.rt.internal_0e20598.asm.org.jacoco.agent.rt.internal_0e20598.asm.ClassReader – 0.00% (0/1524 lines covered)
      • androidx.compose.runtime.androidx.compose.runtime.LinkComposer – 0.00% (0/1430 lines covered)
      • org.jacoco.agent.rt.internal_0e20598.asm.org.jacoco.agent.rt.internal_0e20598.asm.MethodWriter – 0.00% (0/922 lines covered)
      • kotlinx.coroutines.channels.kotlinx.coroutines.channels.BufferedChannel – 0.00% (0/863 lines covered)
      • kotlin.sequences.kotlin.sequences.SequencesKt___SequencesKt – 0.00% (0/744 lines covered)
      • com.google.common.cache.com.google.common.cache.LocalCache$Segment – 0.00% (0/726 lines covered)
      • okio.okio.Buffer – 0.00% (0/687 lines covered)
      • kotlin.text.kotlin.text.StringsKt___StringsKt – 0.00% (0/630 lines covered)
      • com.codename1.androidcompat.rt.com.codename1.androidcompat.rt.TextXmlPullParser – 0.00% (0/623 lines covered)

✅ Native Android screenshot tests passed.

Native Android coverage

  • 📊 Line coverage: 19.18% (56169/292885 lines covered) [HTML preview] (artifact android-coverage-report, jacocoAndroidReport/html/index.html)
    • Other counters: instruction 18.13% (270599/1492353), branch 12.38% (19678/158943), complexity 10.36% (15848/153017), method 15.74% (11411/72518), class 27.93% (2247/8045)
    • Lowest covered classes
      • kotlin.collections.unsigned.kotlin.collections.unsigned.UArraysKt___UArraysKt – 0.00% (0/2396 lines covered)
      • org.jacoco.agent.rt.internal_0e20598.asm.org.jacoco.agent.rt.internal_0e20598.asm.ClassReader – 0.00% (0/1524 lines covered)
      • androidx.compose.runtime.androidx.compose.runtime.LinkComposer – 0.00% (0/1430 lines covered)
      • org.jacoco.agent.rt.internal_0e20598.asm.org.jacoco.agent.rt.internal_0e20598.asm.MethodWriter – 0.00% (0/922 lines covered)
      • kotlinx.coroutines.channels.kotlinx.coroutines.channels.BufferedChannel – 0.00% (0/863 lines covered)
      • kotlin.sequences.kotlin.sequences.SequencesKt___SequencesKt – 0.00% (0/744 lines covered)
      • com.google.common.cache.com.google.common.cache.LocalCache$Segment – 0.00% (0/726 lines covered)
      • okio.okio.Buffer – 0.00% (0/687 lines covered)
      • kotlin.text.kotlin.text.StringsKt___StringsKt – 0.00% (0/630 lines covered)
      • com.codename1.androidcompat.rt.com.codename1.androidcompat.rt.TextXmlPullParser – 0.00% (0/623 lines covered)

Benchmark Results

Detailed Performance Metrics

Metric Duration
SIMD kernel backend scalar fallback (no native SIMD)
SIMD int-add (64K x300) java 175ms / native 94ms = 1.8x speedup
SIMD float-mul (64K x300) java 98ms / native 61ms = 1.6x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path gated to scalar (CPU autovectorizes scalar; explicit SIMD not beneficial here)
Base64 CN1 encode 97.000 ms
Base64 CN1 decode 86.000 ms
Base64 native encode 715.000 ms
Base64 encode ratio (CN1/native) 0.136x (86.4% faster)
Base64 native decode 363.000 ms
Base64 decode ratio (CN1/native) 0.237x (76.3% faster)
Image encode benchmark status skipped (SIMD unsupported)

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 210 screenshots: 210 matched.
Native Windows port, REAL shipping pipeline: the hellocodenameone screenshot suite rendered by a binary CROSS-COMPILED on Linux (clang-cl + xwin, WebView2 linked) and RUN on a Windows x64 runner. Compared against the in-repo baseline in scripts/windows/screenshots.

Benchmark Results

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 27ms / native 2ms = 13.5x speedup
SIMD float-mul (64K x300) java 26ms / native 2ms = 13.0x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 native bridge unavailable (CN1 + SIMD + image benchmarks only)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path gated to scalar (CPU autovectorizes scalar; explicit SIMD not beneficial here)
Base64 CN1 encode 43.000 ms
Base64 CN1 decode 45.000 ms
Base64 SIMD encode 51.000 ms
Base64 encode ratio (SIMD/CN1) 1.186x (18.6% slower)
Base64 SIMD decode 48.000 ms
Base64 decode ratio (SIMD/CN1) 1.067x (6.7% slower)
Image encode benchmark iterations 100
Image createMask (SIMD off) 6.000 ms
Image createMask (SIMD on) 5.000 ms
Image createMask ratio (SIMD on/off) 0.833x (16.7% faster)
Image applyMask (SIMD off) 17.000 ms
Image applyMask (SIMD on) 16.000 ms
Image applyMask ratio (SIMD on/off) 0.941x (5.9% faster)
Image modifyAlpha (SIMD off) 14.000 ms
Image modifyAlpha (SIMD on) 10.000 ms
Image modifyAlpha ratio (SIMD on/off) 0.714x (28.6% faster)
Image modifyAlpha removeColor (SIMD off) 15.000 ms
Image modifyAlpha removeColor (SIMD on) 8.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 0.533x (46.7% faster)

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 210 screenshots: 210 matched.
Native Windows port (x64 / Intel-AMD): full hellocodenameone screenshot suite rendered offscreen with Direct2D/DirectWrite, plus the real benchmarks (base64 native/CN1/SIMD, image createMask/applyMask/modifyAlpha/PNG/JPEG, SSE2 SIMD kernels). Compared against the in-repo baseline in scripts/windows/screenshots.

Benchmark Results

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 29ms / native 2ms = 14.5x speedup
SIMD float-mul (64K x300) java 28ms / native 2ms = 14.0x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 native bridge unavailable (CN1 + SIMD + image benchmarks only)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path gated to scalar (CPU autovectorizes scalar; explicit SIMD not beneficial here)
Base64 CN1 encode 43.000 ms
Base64 CN1 decode 48.000 ms
Base64 SIMD encode 55.000 ms
Base64 encode ratio (SIMD/CN1) 1.279x (27.9% slower)
Base64 SIMD decode 48.000 ms
Base64 decode ratio (SIMD/CN1) 1.000x (0.0% slower)
Image encode benchmark iterations 100
Image createMask (SIMD off) 5.000 ms
Image createMask (SIMD on) 1.000 ms
Image createMask ratio (SIMD on/off) 0.200x (80.0% faster)
Image applyMask (SIMD off) 15.000 ms
Image applyMask (SIMD on) 10.000 ms
Image applyMask ratio (SIMD on/off) 0.667x (33.3% faster)
Image modifyAlpha (SIMD off) 15.000 ms
Image modifyAlpha (SIMD on) 12.000 ms
Image modifyAlpha ratio (SIMD on/off) 0.800x (20.0% faster)
Image modifyAlpha removeColor (SIMD off) 17.000 ms
Image modifyAlpha removeColor (SIMD on) 9.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 0.529x (47.1% faster)

ParparVM vs HotSpot (JDK 25): Windows x64

Runner CPU: AMD64 Family 26 Model 2 Stepping 1, AuthenticAMD (baseline windows-x64@amd64-family-26-model-2-authenticamd)

Ratios are ParparVM / JDK 25: below 1.00x ParparVM is faster (time) or smaller (RAM). Median of 5 interleaved, paired rounds; every run's output was verified. A ratio more than 20% (time) / 25% (RAM) above its baseline in vm/selfhost/perf-baseline/ is a regression and fails; one that far below is an improvement, reported but not failed (a row whose calibration runs were noisier carries a wider tolerance), and for RAM the change must also exceed 0.05x in absolute terms. Both run unpinned on all of the runner's CPUs, with their own default thread counts.

Benchmark Cores Time RAM Status
hello (WinHelloMain) 4 2.06x (base 1.41x, +45.8%) 0.70x (base 0.68x, +1.9%) ok
translator (self) 4 0.91x (base 0.90x, +1.7%) 0.48x (base 0.47x, +3.0%) ok
intArithmetic 4 1.00x (base 0.99x, +0.9%) 0.06x (base 0.05x, +2.8%) ok
longArithmetic 4 1.00x (base 0.99x, +0.9%) 0.06x (base 0.05x, +3.0%) ok
mathTranscendental 4 0.93x (base 0.94x, -0.7%) 0.06x (base 0.06x, +2.6%) ok
arraySequential 4 1.64x (base 1.75x, -6.1%) 0.38x (base 0.38x, +0.1%) ok
arrayRandom 4 0.99x (base 1.00x, -1.4%) 0.23x (base 0.23x, +0.6%) ok
objectAllocation 4 1.56x (base 1.65x, -5.3%) 0.26x (base 0.21x, +23.4%) ok
valueEscape 4 0.10x (base 0.10x, +0.5%) 0.05x (base 0.04x, +3.7%) ok
hashMapChurn 4 1.60x (base 1.59x, +0.3%) 0.08x (base 0.09x, -12.5%) ok
stringBuilding 4 1.43x (base 1.39x, +3.0%) 0.27x (base 0.27x, -0.1%) ok
recursion 4 1.42x (base 1.44x, -1.2%) 0.06x (base 0.06x, +3.2%) ok
quicksort 4 1.06x (base 1.06x, +0.2%) 0.12x (base 0.12x, +1.6%) ok

Result: no regression

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 210 screenshots: 210 matched.
Native Windows port (arm64 / Apple Silicon - Arm): full hellocodenameone screenshot suite rendered offscreen with Direct2D/DirectWrite, plus the real benchmarks (base64 native/CN1/SIMD, image createMask/applyMask/modifyAlpha/PNG/JPEG, NEON SIMD kernels). Compared against the in-repo baseline in scripts/windows/screenshots.

Benchmark Results

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 54ms / native 4ms = 13.5x speedup
SIMD float-mul (64K x300) java 52ms / native 3ms = 17.3x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 native bridge unavailable (CN1 + SIMD + image benchmarks only)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path gated to scalar (CPU autovectorizes scalar; explicit SIMD not beneficial here)
Base64 CN1 encode 63.000 ms
Base64 CN1 decode 68.000 ms
Base64 SIMD encode 78.000 ms
Base64 encode ratio (SIMD/CN1) 1.238x (23.8% slower)
Base64 SIMD decode 68.000 ms
Base64 decode ratio (SIMD/CN1) 1.000x (0.0% slower)
Image encode benchmark iterations 100
Image createMask (SIMD off) 6.000 ms
Image createMask (SIMD on) 1.000 ms
Image createMask ratio (SIMD on/off) 0.167x (83.3% faster)
Image applyMask (SIMD off) 16.000 ms
Image applyMask (SIMD on) 16.000 ms
Image applyMask ratio (SIMD on/off) 1.000x (0.0% slower)
Image modifyAlpha (SIMD off) 15.000 ms
Image modifyAlpha (SIMD on) 10.000 ms
Image modifyAlpha ratio (SIMD on/off) 0.667x (33.3% faster)
Image modifyAlpha removeColor (SIMD off) 22.000 ms
Image modifyAlpha removeColor (SIMD on) 11.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 0.500x (50.0% faster)

ParparVM vs HotSpot (JDK 25): Windows arm64

Runner CPU: ARMv8 (64-bit) Family 8 Model D49 Revision 0, MICROSOFT CORPORATION (baseline windows-arm64@armv8-64-bit-family-8-model-d49-microsoft-corporation)

Ratios are ParparVM / JDK 25: below 1.00x ParparVM is faster (time) or smaller (RAM). Median of 5 interleaved, paired rounds; every run's output was verified. A ratio more than 20% (time) / 25% (RAM) above its baseline in vm/selfhost/perf-baseline/ is a regression and fails; one that far below is an improvement, reported but not failed (a row whose calibration runs were noisier carries a wider tolerance), and for RAM the change must also exceed 0.05x in absolute terms. Both run unpinned on all of the runner's CPUs, with their own default thread counts.

Benchmark Cores Time RAM Status
hello (WinHelloMain) 4 1.45x (base 1.45x, +0.3%) 0.77x (base 0.75x, +2.1%) ok
translator (self) 4 0.92x (base 0.84x, +9.2%) 0.49x (base 0.51x, -2.6%) ok
intArithmetic 4 1.04x (base 1.04x, -0.0%) 0.06x (base 0.06x, +3.2%) ok
longArithmetic 4 0.80x (base 0.80x, +0.0%) 0.06x (base 0.06x, +3.5%) ok
mathTranscendental 4 0.73x (base 0.72x, +1.8%) 0.07x (base 0.06x, +1.9%) ok
arraySequential 4 0.40x (base 0.46x, -11.5%) 0.39x (base 0.39x, +0.0%) ok
arrayRandom 4 0.94x (base 0.94x, +0.4%) 0.23x (base 0.23x, +0.3%) ok
objectAllocation 4 1.75x (base 1.82x, -4.0%) 0.41x (base 0.43x, -5.5%) ok
valueEscape 4 0.76x (base 0.76x, +0.0%) 0.05x (base 0.05x, +2.4%) ok
hashMapChurn 4 0.99x (base 0.98x, +0.6%) 0.12x (base 0.12x, -1.5%) ok
stringBuilding 4 1.54x (base 1.43x, +8.0%) 0.27x (base 0.27x, -0.1%) ok
recursion 4 1.42x (base 1.46x, -2.9%) 0.07x (base 0.07x, +1.3%) ok
quicksort 4 1.00x (base 1.00x, +0.2%) 0.12x (base 0.12x, +0.6%) ok

Result: no regression

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 193 screenshots: 193 matched.
✅ Native iOS Metal screenshot tests passed.

Benchmark Results

  • VM Translation Time: 0 seconds
  • Compilation Time: 2902 seconds

Build and Run Timing

Metric Duration
Simulator Boot 75000 ms
Simulator Boot (Run) 2000 ms
App Install 25000 ms
App Launch 1000 ms
Test Execution 611000 ms

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 118ms / native 4ms = 29.5x speedup
SIMD float-mul (64K x300) java 169ms / native 3ms = 56.3x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path active (NEON-accelerated)
Base64 CN1 encode 119.000 ms
Base64 CN1 decode 65.000 ms
Base64 native encode 2472.000 ms
Base64 encode ratio (CN1/native) 0.048x (95.2% faster)
Base64 native decode 733.000 ms
Base64 decode ratio (CN1/native) 0.089x (91.1% faster)
Base64 SIMD encode 130.000 ms
Base64 encode ratio (SIMD/CN1) 1.092x (9.2% slower)
Base64 SIMD decode 167.000 ms
Base64 decode ratio (SIMD/CN1) 2.569x (156.9% slower)
Base64 encode ratio (SIMD/native) 0.053x (94.7% faster)
Base64 decode ratio (SIMD/native) 0.228x (77.2% faster)
Image encode benchmark iterations 100
Image createMask (SIMD off) 5.000 ms
Image createMask (SIMD on) 1.000 ms
Image createMask ratio (SIMD on/off) 0.200x (80.0% faster)
Image applyMask (SIMD off) 105.000 ms
Image applyMask (SIMD on) 755.000 ms
Image applyMask ratio (SIMD on/off) 7.190x (619.0% slower)
Image modifyAlpha (SIMD off) 104.000 ms
Image modifyAlpha (SIMD on) 166.000 ms
Image modifyAlpha ratio (SIMD on/off) 1.596x (59.6% slower)
Image modifyAlpha removeColor (SIMD off) 97.000 ms
Image modifyAlpha removeColor (SIMD on) 248.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 2.557x (155.7% slower)

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 193 screenshots: 193 matched.
✅ Native iOS Metal screenshot tests passed.

Benchmark Results

  • VM Translation Time: 0 seconds
  • Compilation Time: 2851 seconds

Build and Run Timing

Metric Duration
Simulator Boot 115000 ms
Simulator Boot (Run) 1000 ms
App Install 14000 ms
App Launch 1000 ms
Test Execution 450000 ms

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 304ms / native 23ms = 13.2x speedup
SIMD float-mul (64K x300) java 123ms / native 4ms = 30.7x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path active (NEON-accelerated)
Base64 CN1 encode 91.000 ms
Base64 CN1 decode 91.000 ms
Base64 native encode 1184.000 ms
Base64 encode ratio (CN1/native) 0.077x (92.3% faster)
Base64 native decode 1418.000 ms
Base64 decode ratio (CN1/native) 0.064x (93.6% faster)
Base64 SIMD encode 86.000 ms
Base64 encode ratio (SIMD/CN1) 0.945x (5.5% faster)
Base64 SIMD decode 71.000 ms
Base64 decode ratio (SIMD/CN1) 0.780x (22.0% faster)
Base64 encode ratio (SIMD/native) 0.073x (92.7% faster)
Base64 decode ratio (SIMD/native) 0.050x (95.0% faster)
Image encode benchmark iterations 100
Image createMask (SIMD off) 7.000 ms
Image createMask (SIMD on) 4.000 ms
Image createMask ratio (SIMD on/off) 0.571x (42.9% faster)
Image applyMask (SIMD off) 298.000 ms
Image applyMask (SIMD on) 42.000 ms
Image applyMask ratio (SIMD on/off) 0.141x (85.9% faster)
Image modifyAlpha (SIMD off) 69.000 ms
Image modifyAlpha (SIMD on) 45.000 ms
Image modifyAlpha ratio (SIMD on/off) 0.652x (34.8% faster)
Image modifyAlpha removeColor (SIMD off) 83.000 ms
Image modifyAlpha removeColor (SIMD on) 168.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 2.024x (102.4% slower)

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 210 screenshots: 210 matched.
Native Linux port (x64), GTK3/Cairo/Pango, ParparVM bytecode-to-C (no JVM): the hellocodenameone screenshot suite rendered by a native ELF built + run on the GitHub x64 runner. Baseline: scripts/linux/screenshots.

ParparVM vs HotSpot (JDK 25): Linux x64

Runner CPU: INTEL(R) XEON(R) PLATINUM 8573C (baseline linux-x64@intel-xeon-platinum-8573c)

Ratios are ParparVM / JDK 25: below 1.00x ParparVM is faster (time) or smaller (RAM). Median of 5 interleaved, paired rounds; every run's output was verified. A ratio more than 20% (time) / 25% (RAM) above its baseline in vm/selfhost/perf-baseline/ is a regression and fails; one that far below is an improvement, reported but not failed (a row whose calibration runs were noisier carries a wider tolerance), and for RAM the change must also exceed 0.05x in absolute terms. Both run unpinned on all of the runner's CPUs, with their own default thread counts.

Benchmark Cores Time RAM Status
hello (LinuxHelloMain) 4 1.10x (base 0.94x, +17.0%) 0.77x (base 0.88x, -11.9%) ok
translator (self) 4 0.68x (base 0.67x, +2.5%) 0.44x (base 0.44x, +0.9%) ok
intArithmetic 4 1.03x (base 1.03x, +0.0%) 0.04x (base 0.07x, -47.2%) ok
longArithmetic 4 0.96x (base 0.96x, -0.1%) 0.04x (base 0.05x, -29.4%) ok
mathTranscendental 4 1.03x (base 1.03x, -0.0%) 0.04x (base 0.06x, -28.2%) ok
arraySequential 4 2.43x (base 2.69x, -9.6%) 0.36x (base 0.35x, +1.1%) ok
arrayRandom 4 1.01x (base 1.05x, -3.7%) 0.20x (base 0.21x, -3.1%) ok
objectAllocation 4 0.87x (base 1.03x, -15.7%) 0.36x (base 0.51x, -28.8%) ok
valueEscape 4 0.13x (base 0.13x, -0.3%) 0.03x (base 0.04x, -30.1%) ok
hashMapChurn 4 0.82x (base 1.00x, -18.1%) 0.12x (base 0.10x, +26.3%) ok
stringBuilding 4 1.55x (base 1.53x, +1.3%) 0.26x (base 0.22x, +18.7%) ok
recursion 4 1.24x (base 1.05x, +17.2%) 0.04x (base 0.06x, -31.7%) ok
quicksort 4 1.03x (base 1.08x, -4.2%) 0.10x (base 0.11x, -10.4%) ok

Result: no regression

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 210 screenshots: 210 matched.
Native Linux port (arm64), GTK3/Cairo/Pango, ParparVM bytecode-to-C (no JVM): the hellocodenameone screenshot suite rendered by a native ELF built + run on the GitHub arm64 runner. Baseline: scripts/linux/screenshots-arm.

ParparVM vs HotSpot (JDK 25): Linux arm64

Runner CPU: Neoverse-N2 (baseline linux-arm64@neoverse-n2)

Ratios are ParparVM / JDK 25: below 1.00x ParparVM is faster (time) or smaller (RAM). Median of 5 interleaved, paired rounds; every run's output was verified. A ratio more than 20% (time) / 25% (RAM) above its baseline in vm/selfhost/perf-baseline/ is a regression and fails; one that far below is an improvement, reported but not failed (a row whose calibration runs were noisier carries a wider tolerance), and for RAM the change must also exceed 0.05x in absolute terms. Both run unpinned on all of the runner's CPUs, with their own default thread counts.

Benchmark Cores Time RAM Status
hello (LinuxHelloMain) 4 1.14x (base 1.02x, +12.1%) 0.78x (base 0.66x, +18.9%) ok
translator (self) 4 0.79x (base 0.70x, +13.8%) 0.47x (base 0.43x, +8.6%) ok
intArithmetic 4 1.04x (base 1.04x, -0.0%) 0.03x (base 0.03x, +4.7%) ok
longArithmetic 4 0.79x (base 0.79x, -0.3%) 0.03x (base 0.03x, +5.0%) ok
mathTranscendental 4 1.09x (base 1.10x, -0.3%) 0.04x (base 0.03x, +9.7%) ok
arraySequential 4 0.38x (base 0.36x, +5.1%) 0.37x (base 0.37x, +0.3%) ok
arrayRandom 4 0.95x (base 0.94x, +1.4%) 0.20x (base 0.20x, +0.5%) ok
objectAllocation 4 1.96x (base 1.77x, +11.1%) 0.35x (base 0.37x, -4.5%) ok
valueEscape 4 0.51x (base 0.51x, -0.1%) 0.03x (base 0.03x, +6.0%) ok
hashMapChurn 4 0.86x (base 0.84x, +2.4%) 0.12x (base 0.12x, -1.6%) ok
stringBuilding 4 1.39x (base 1.57x, -11.2%) 0.32x (base 0.26x, +20.5%) ok
recursion 4 1.56x (base 1.43x, +8.5%) 0.04x (base 0.04x, +3.4%) ok
quicksort 4 0.99x (base 0.99x, +0.2%) 0.10x (base 0.09x, +2.7%) ok

Result: no regression

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 253 screenshots: 253 matched.
✅ Native Apple Watch (watchOS, Core Graphics) screenshot tests passed.

@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 188 screenshots: 188 matched.
✅ Native Apple TV (tvOS, Metal) screenshot tests passed.

@shai-almog
shai-almog merged commit 78ac218 into master Nov 10, 2026
97 of 98 checks passed
@shai-almog
shai-almog deleted the steam-machine-support branch November 10, 2026 17:29
@shai-almog

shai-almog commented Nov 10, 2026 •

Copy link
Copy Markdown
Collaborator Author

Compared 192 screenshots: 192 matched.
✅ Native Mac screenshot tests passed.

Benchmark Results

  • VM Translation Time: 0 seconds
  • Compilation Time: 803 seconds

Detailed Performance Metrics

Metric Duration
SIMD kernel backend SSE2 (x64) / NEON (arm64) native kernels
SIMD int-add (64K x300) java 65ms / native 9ms = 7.2x speedup
SIMD float-mul (64K x300) java 52ms / native 2ms = 26.0x speedup
SIMD kernel correctness PASS (native result == scalar reference)
Base64 payload size 8192 bytes
Base64 benchmark iterations 6000
Base64 SIMD byte path active (NEON-accelerated)
Base64 CN1 encode 70.000 ms
Base64 CN1 decode 82.000 ms
Base64 native encode 892.000 ms
Base64 encode ratio (CN1/native) 0.078x (92.2% faster)
Base64 native decode 400.000 ms
Base64 decode ratio (CN1/native) 0.205x (79.5% faster)
Base64 SIMD encode 59.000 ms
Base64 encode ratio (SIMD/CN1) 0.843x (15.7% faster)
Base64 SIMD decode 63.000 ms
Base64 decode ratio (SIMD/CN1) 0.768x (23.2% faster)
Base64 encode ratio (SIMD/native) 0.066x (93.4% faster)
Base64 decode ratio (SIMD/native) 0.158x (84.3% faster)
Image encode benchmark iterations 100
Image createMask (SIMD off) 12.000 ms
Image createMask (SIMD on) 7.000 ms
Image createMask ratio (SIMD on/off) 0.583x (41.7% faster)
Image applyMask (SIMD off) 101.000 ms
Image applyMask (SIMD on) 76.000 ms
Image applyMask ratio (SIMD on/off) 0.752x (24.8% faster)
Image modifyAlpha (SIMD off) 69.000 ms
Image modifyAlpha (SIMD on) 68.000 ms
Image modifyAlpha ratio (SIMD on/off) 0.986x (1.4% faster)
Image modifyAlpha removeColor (SIMD off) 95.000 ms
Image modifyAlpha removeColor (SIMD on) 113.000 ms
Image modifyAlpha removeColor ratio (SIMD on/off) 1.189x (18.9% slower)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant