SiliconScope v1.4.0

Menu-bar cockpit + chip-agnostic accuracy.

• Live 6-bar menu-bar glyph — CPU / GPU / ANE / Media / memory-usage / memory-bandwidth as colored mini bars with a stacked "SS" label, drawn as a bitmap for reliable rendering. The whole glyph blinks red on an alert (swap, memory-pressure critical, or GPU throttle).
• Revamped dropdown — six color-matched trend graphs mirroring the glyph, each on a fixed Y axis matched to its bar (no auto-scale exaggeration); memory usage is now a line graph; top processes; and an Open Dashboard button that brings the main window forward from the background. Tighter, denser layout.
• Honest AI attribution — distinguishes a loaded runtime from an idle daemon (Ollama (idle)) and an unmanaged in-app / MLX-Swift workload (in-app / unmanaged); the dashboard no longer credits an idle daemon for GPU work done by another app.
• Chip-agnostic bottleneck verdict — bandwidth-bound is judged against the machine's own observed achievable bandwidth peak (self-calibrating across M1…M5+), not a fixed fraction of the theoretical spec. Observed peaks decay slowly so a one-off spike no longer pins the scale.
• Compact dashboard — smaller card padding / heights / spacing and a narrower default window.

Notarized DMG below. Apple Silicon only, sudoless.

SiliconScope v1.3.0

Local-AI monitoring — a dedicated cockpit for running LLMs on Apple Silicon.

• AI runtime detection — recognizes Ollama, llama.cpp, LM Studio, MLX, Jan, GPT4All, vLLM by process (bundle-first, sudoless) with RAM/CPU.
• Model memory budget — "largest model that fits" + "if you unload " per quant, with a rate-based swap-risk signal that warns before tokens/sec collapse.
• Runtime API (opt-in, off by default) — reads the loaded model, the authoritative GPU/CPU offload split (Ollama size_vram/size), and tokens/sec (llama.cpp /metrics) from 127.0.0.1. Nothing leaves your Mac.
• AI Workload classifier calibrated against real M1 Max LLM runs (MoE / dense / memory-pressured) and stabilized with a rolling average.
• Honest attribution — GPU activity is not assumed to be AI; "LLM" is asserted only with evidence (a runtime serving a model), ANE ⇒ CoreML, Media engine ⇒ video, otherwise "type unknown."

Designed via a multi-stage review, then validated and tuned on-device. Thanks to @durul for the earlier AI-workload groundwork.

Install: download the DMG, open it, drag SiliconScope to Applications. Developer-ID signed + notarized. macOS 14+, Apple Silicon.

SiliconScope v1.2.0

AI-workload monitoring — the next-version hero feature.

• AI Workload view — a bottleneck classifier with a single verdict: bandwidth-bound / compute-bound / thermal-throttled / memory-pressured (plus idle / GPU-active). Front-and-center hero card on the dashboard; mirrored as a Workload: line in the menu bar.
• Per-chip memory-bandwidth ceiling table + a "% of ceiling" gauge (M1–M4; Max bins split by P-core count; self-corrects to the observed peak for chips outside the table).
• GPU throttle detector — flags the GPU clock held below its slowly-decaying rolling peak while thermal pressure has risen (warning banner + menu-bar flame).
• Compact GPU menu-bar mode — single line: GPU% / GPU W / GPU GB/s / die °C.

Thanks to @durul for contributing this feature (#2).

Install: download the DMG below, open it, drag SiliconScope to Applications. Signed with Developer ID + notarized by Apple. Requires macOS 14+ on Apple Silicon.

SiliconScope v1.1.0

Renamed WhisPlayInfo → SiliconScope.

The project outgrew its origin as a companion utility; the name now reflects what it is — a general Apple Silicon / SoC inspector. No functional changes to the metrics in this release.

• App / product name: SiliconScope (was WhisPlayInfo)
• Bundle identifier: ai.calidalab.SiliconScope
• SwiftPM targets: SiliconScope (app), SiliconScopeCore (data library), sscope-cli (CLI)
• Repository: github.com/kennss/SiliconScope (the old URL redirects)

Because the bundle identifier changed, this installs alongside any existing WhisPlayInfo rather than upgrading it in place — delete the old app if you have it.

Install: download the DMG below, open it, drag SiliconScope to Applications. Signed with Developer ID + notarized by Apple. Requires macOS 14+ on Apple Silicon.

Thanks to @durul for the naming proposal (#4).

WhisPlayInfo v1.0.2

Crash fix: launch failure on macOS 27.

• Fixed an immediate crash on launch under macOS 26/27 (EXC_BREAKPOINT in Bundle.module). The SwiftPM resource bundle ships as a flat folder with no Info.plist; macOS 27's stricter bundle validation rejects it, so SwiftPM's generated Bundle.module accessor hit its fatalError.
• The app icon is now resolved via the main bundle (with a dev-run fallback), removing all dependence on Bundle.module.

Thanks to @colaH16 for the report (#1).

Install: download the DMG below, open it, drag WhisPlayInfo to Applications. Signed with Developer ID + notarized by Apple.

WhisPlayInfo v1.0.1

⚠️ The original build here crashed on launch under macOS 27. The download below has been replaced with the fixed v1.0.2 binary. Please use v1.0.2.

---

Bug fix: memory-bandwidth Media Engine reporting.

• Fixed Media Engine bandwidth reading 0 while a media-engine app (e.g. video transcoding) was active — now classifies the real channels (VENC / VDEC / ISP / JPEG / STRM CODEC / ProRes), matching NeoAsitop.
• MSR is no longer miscounted as Media; it now falls into Other.
• Total bandwidth now uses the chip-wide DCS aggregate, with Other derived as total − CPU − GPU − Media, so the parts sum to the real total (previously double-counted, ~104 vs ~50 GB/s).

Install: download the DMG below, open it, drag WhisPlayInfo to Applications. Signed with Developer ID + notarized by Apple.

WhisPlayInfo v1.0.0

⚠️ The original build here crashed on launch under macOS 27. The download below has been replaced with the fixed v1.0.2 binary. Please use v1.0.2.

---

First public release — a sudoless Apple Silicon system monitor (native SwiftUI GUI) with first-class ANE / Media Engine / memory-bandwidth tracking.

Install

Download WhisPlayInfo-1.0.0.dmg, open it, and drag WhisPlayInfo to Applications.
Signed with Developer ID and notarized by Apple — opens with no Gatekeeper warnings. Requires macOS 14+ on Apple Silicon.

Highlights

• E-core / P-core usage + frequency, GPU usage/power/freq, ANE power, Media Engine bandwidth
• Memory (Wired / Active / Compressed / Free) + bandwidth + macOS pressure alerts
• Network ↑/↓, Disk read/write + capacity, with live graphs
• Temperatures (CPU / GPU / Memory / Battery), fans, thermal pressure
• Per-domain power, battery; processes (sort / filter / kill)
• Menu bar + full dashboard

No sudo. Not on the App Store (uses private but un-entitled APIs).