github ur-grue/toptop v1.0.2
on GitHub

latest releases: v1.1.1, v1.1.0
one month ago

What's Changed

  • Fix signal-menu defects: outcomes, labels, and platform numbers by @ur-grue in #20
  • Derive signal name/number from one authoritative SIGNALS table by @ur-grue in #21
  • Update tests badge to 68 by @ur-grue in #22
  • Document signal-menu permission-denied feedback by @ur-grue in #23
  • Recolor AI-view hero to match the default gruvbox theme by @ur-grue in #24
  • Add macOS to the platform badge by @ur-grue in #25
  • Add CHANGELOG documenting shipped signal-menu work by @ur-grue in #26
  • Add ROADMAP (themes, tracked issues + new proposals, horizons) by @ur-grue in #27
  • Link roadmap proposals to their tracking issues by @ur-grue in #47
  • Add --export csv by @ur-grue in #48
  • Add renice (change process priority) from a menu by @ur-grue in #50
  • GPU VRAM column + sort in the process table by @ur-grue in #49
  • AI view adapts honestly to the hardware environment by @ur-grue in #51
  • Apple Silicon GPU utilization (IOKit, no root) by @ur-grue in #53
  • Refocus ROADMAP on the 1.1 "inference everywhere" bet by @ur-grue in #52
  • Add Claude Code GitHub Workflow by @ur-grue in #55
  • ai: detect MLC LLM serve by @vaishnavidesai09 in #54
  • Make alert thresholds configurable (config.conf + CLI) by @ur-grue in #56
  • Config polish: --config <path>, --no-save, don't swallow save errors by @ur-grue in #57
  • Unit tests for app.rs state logic by @ur-grue in #60
  • AI view: tokens/sec and KV-cache history sparklines by @ur-grue in #61
  • Extend the TGI parser: tokens/sec and TTFT by @ur-grue in #58
  • Detect more inference runtimes: TensorRT-LLM, MLC LLM, LM Studio by @ur-grue in #62
  • Extract arg parsing into pure parse_args + per-flag tests by @ur-grue in #63
  • Completions for all flags, tests badge, changelog dedupe by @ur-grue in #64
  • docs: correct the launch-kit checklist against reality by @ur-grue in #67
  • Config-driven customization: user themes, keybindings, columns by @ur-grue in #65
  • Alert transitions: actions, built-in sinks, history timeline by @ur-grue in #75
  • Inference telemetry: SLO triad, prefill/decode split, SGLang, manual targets by @ur-grue in #69
  • KV-cache preemption and GPU compute-vs-bandwidth history by @ur-grue in #76
  • Process detail: open files, sockets and environment by @ur-grue in #70
  • Flight recorder: --record and --replay by @ur-grue in #71
  • UI polish: honest GPU readings, deduplicated disks, content-sized AI view by @ur-grue in #74
  • Basic Windows support, verified by a windows-latest CI job by @ur-grue in #66
  • CI hardening and a hot-path regression guard by @ur-grue in #72
  • AI view: wrap instead of clipping, and no trend over blank rows by @ur-grue in #77
  • Responsive process table: drop columns instead of squeezing them by @ur-grue in #78
  • Diagnosis panel: name the bottleneck, not just the numbers by @ur-grue in #79
  • Container awareness: cgroup v2 limits and grouping by pod by @ur-grue in #80
  • OpenTelemetry (OTLP/HTTP) metrics export by @ur-grue in #81
  • Refresh the Grafana dashboard, and stop it rotting again by @ur-grue in #82
  • Make the look coherent: cyberpunk by default, and legible dim text by @ur-grue in #83
  • CI: use a prebuilt cargo-audit instead of building it every run by @ur-grue in #85
  • --demo: plot the simulated GPU, and show it accurately in the README by @ur-grue in #84
  • Make installing it a one-liner on every platform we claim to support by @ur-grue in #86
  • Point the empty AI view at --demo by @ur-grue in #87
  • Cross-compile the Intel macOS build instead of queueing forever by @ur-grue in #88
  • Refuse to publish a release whose tag disagrees with Cargo.toml by @ur-grue in #89

New Contributors

Full Changelog: v1.0.1...v1.0.2

What's Changed

  • Fix signal-menu defects: outcomes, labels, and platform numbers by @ur-grue in #20
  • Derive signal name/number from one authoritative SIGNALS table by @ur-grue in #21
  • Update tests badge to 68 by @ur-grue in #22
  • Document signal-menu permission-denied feedback by @ur-grue in #23
  • Recolor AI-view hero to match the default gruvbox theme by @ur-grue in #24
  • Add macOS to the platform badge by @ur-grue in #25
  • Add CHANGELOG documenting shipped signal-menu work by @ur-grue in #26
  • Add ROADMAP (themes, tracked issues + new proposals, horizons) by @ur-grue in #27
  • Link roadmap proposals to their tracking issues by @ur-grue in #47
  • Add --export csv by @ur-grue in #48
  • Add renice (change process priority) from a menu by @ur-grue in #50
  • GPU VRAM column + sort in the process table by @ur-grue in #49
  • AI view adapts honestly to the hardware environment by @ur-grue in #51
  • Apple Silicon GPU utilization (IOKit, no root) by @ur-grue in #53
  • Refocus ROADMAP on the 1.1 "inference everywhere" bet by @ur-grue in #52
  • Add Claude Code GitHub Workflow by @ur-grue in #55
  • ai: detect MLC LLM serve by @vaishnavidesai09 in #54
  • Make alert thresholds configurable (config.conf + CLI) by @ur-grue in #56
  • Config polish: --config <path>, --no-save, don't swallow save errors by @ur-grue in #57
  • Unit tests for app.rs state logic by @ur-grue in #60
  • AI view: tokens/sec and KV-cache history sparklines by @ur-grue in #61
  • Extend the TGI parser: tokens/sec and TTFT by @ur-grue in #58
  • Detect more inference runtimes: TensorRT-LLM, MLC LLM, LM Studio by @ur-grue in #62
  • Extract arg parsing into pure parse_args + per-flag tests by @ur-grue in #63
  • Completions for all flags, tests badge, changelog dedupe by @ur-grue in #64
  • docs: correct the launch-kit checklist against reality by @ur-grue in #67
  • Config-driven customization: user themes, keybindings, columns by @ur-grue in #65
  • Alert transitions: actions, built-in sinks, history timeline by @ur-grue in #75
  • Inference telemetry: SLO triad, prefill/decode split, SGLang, manual targets by @ur-grue in #69
  • KV-cache preemption and GPU compute-vs-bandwidth history by @ur-grue in #76
  • Process detail: open files, sockets and environment by @ur-grue in #70
  • Flight recorder: --record and --replay by @ur-grue in #71
  • UI polish: honest GPU readings, deduplicated disks, content-sized AI view by @ur-grue in #74
  • Basic Windows support, verified by a windows-latest CI job by @ur-grue in #66
  • CI hardening and a hot-path regression guard by @ur-grue in #72
  • AI view: wrap instead of clipping, and no trend over blank rows by @ur-grue in #77
  • Responsive process table: drop columns instead of squeezing them by @ur-grue in #78
  • Diagnosis panel: name the bottleneck, not just the numbers by @ur-grue in #79
  • Container awareness: cgroup v2 limits and grouping by pod by @ur-grue in #80
  • OpenTelemetry (OTLP/HTTP) metrics export by @ur-grue in #81
  • Refresh the Grafana dashboard, and stop it rotting again by @ur-grue in #82
  • Make the look coherent: cyberpunk by default, and legible dim text by @ur-grue in #83
  • CI: use a prebuilt cargo-audit instead of building it every run by @ur-grue in #85
  • --demo: plot the simulated GPU, and show it accurately in the README by @ur-grue in #84
  • Make installing it a one-liner on every platform we claim to support by @ur-grue in #86
  • Point the empty AI view at --demo by @ur-grue in #87
  • Cross-compile the Intel macOS build instead of queueing forever by @ur-grue in #88
  • Refuse to publish a release whose tag disagrees with Cargo.toml by @ur-grue in #89

New Contributors

Full Changelog: v1.0.1...v1.0.2

Check out latest releases or
releases around ur-grue/toptop v1.0.2

Don't miss a new toptop release

NewReleases is sending notifications on new releases.

Get notifications