Phoronix has published a benchmark roundup putting upstream Arch Linux head to head with CachyOS and the freshly released Omarchy 4.5 — a timely comparison that puts numbers behind a question anyone picking a daily-driver distribution keeps asking: does the tuning a distro ships actually show up in real workloads?
The timing is what makes the comparison worth running, according to Phoronix: CachyOS has recently rolled out a Linux 7.3-based kernel alongside a set of broader improvements, while Omarchy 4.5 landed this week. The review positions the three distributions as a spectrum rather than a ranking.
Three distributions, three design goals. Stock Arch Linux remains the baseline — close to upstream, with the kernel and scheduler configuration largely left as shipped. CachyOS, by contrast, is built explicitly around tuning: its value proposition is kernel-level and package-level optimisation done for the user, rather than left to a post-install configure step. Omarchy 4.5 takes a different angle again, shipping as a curated desktop and workflow — an opinionated arrangement of tools and defaults rather than a raw performance play. All three draw on the same Arch heritage, which means shared package tooling and access to the AUR for anyone extending the setup.
Because the kernel version is central to how the results should be read, the per-test charts in the original review — not this article — remain the authoritative record of the numbers.
Four things not to read into the results.
- The kernel is the variable. Any distribution shipping its own tuned kernel is testing a kernel difference as much as a distribution difference. Reproducing the comparison requires pinning the kernel, CPU governor, and power profile across all systems — otherwise the experiment measures configuration drift, not distribution quality.
- Test configuration matters. Compiler flags, background services, and filesystem choices can move results by more than the gap being measured. A result from one machine and one setup does not generalise automatically.
- Microbenchmarks are not perceived performance. A synthetic throughput win rarely translates into a visibly faster desktop session; input latency, application launch, and idle power draw tell a different story.
- Defaults are not a ceiling. A distribution's out-of-the-box score reflects its shipping configuration, not its achievable one. A stock Arch user who applies the same scheduler and governor settings is competing on different terms.
Practically, treat spreads in the 1–2% range as noise unless they repeat across multiple runs and multiple test suites.
What this means for a rollout decision. The benchmark spread is real but narrow; the decision that actually shapes daily work is the workflow each project commits you to. Arch gives maximum control and the smallest maintenance abstraction. CachyOS front-loads tuning work that a performance-minded user would otherwise do manually. Omarchy 4.5 optimises for a coherent working environment from first boot. For a home lab or a workstation that must be maintained alongside everything else, that difference in maintenance burden tends to matter more than a chart position.
Whichever route is chosen, note that all three sit on a rolling-release model: there is no annual upgrade window to plan around, but there is continuous incremental change to validate — a consideration for anyone managing systems over a long horizon. Users downloading packages over the network can also point their package manager at a geographically closer mirror to cut download latency, a generic optimisation available regardless of which of the three is selected.
The full set of per-test results, including the specific deltas behind each workload, is available in the original Phoronix review.
Phoronix 發布了一項基準測試比較，將原版 Arch Linux 與 CachyOS 及剛推出的 Omarchy 4.5 正面對比——這是一次適逢其時的比較，為任何挑選日常主力發行版的用戶一直提出的問題提供數據支撐：發行版出廠自帶的調校，真的會在實際 workload 中顯現出來嗎？
Phoronix 認為，正是時機令這次比較值得進行：CachyOS 最近推出基於 Linux 7.3 的 kernel，並附帶一系列更廣泛的改進；與此同時，Omarchy 4.5 於本周登場。是次評測將三個發行版定位為一個光譜，而非高下排名。
三個發行版，三套設計目標。 原版 Arch Linux 維持基準地位——貼近 upstream，kernel 與 scheduler 配置大致維持出廠設定。相比之下，CachyOS 明確以調校為核心打造：其價值主張在於為用戶完成 kernel 層面及 package 層面的優化，而非留待安裝後的配置步驟。Omarchy 4.5 則採取另一角度，以精選桌面及工作流程的形式推出——是一套有明確立場的工具與預設配置安排，而非單純追求效能。三者同樣承襲 Arch 血統，意味著共用 package 工具鏈，任何擴展配置的用戶均可使用 AUR。
由於 kernel 版本是解讀結果的關鍵，原始評測中的各項測試圖表——而非本文——才是數據的權威來源。
四點不應從結果中過度解讀之處。
- kernel 才是變數。 任何自帶調校 kernel 的發行版，測試的既是 kernel 差異，也是發行版差異。要重現該比較，必須在所有系統上固定 kernel、CPU governor 及 power profile——否則實驗衡量的是配置漂移，而非發行版質素。
- 測試配置至關重要。 compiler flags、背景服務及檔案系統的選擇，其對結果的影響可能大於所要量度的差距。單一機器、單一配置得出的結果，不能自動推而廣之。
- 微基準測試不等於實際體驗效能。 合成吞吐量上的勝利很少轉化為肉眼可見的桌面流暢度提升；輸入延遲、應用啟動速度及閒置耗電量則講出另一個故事。
- 預設配置不代表上限。 發行版開箱即用的分數反映其出廠配置，而非其可達上限。原版 Arch 用戶若套用相同的 scheduler 及 governor 設定，便是在不同條件下競爭。
實務上，1–2% 範圍內的差距應視為噪音，除非它在多次運行及多套測試中反覆出現。
這對部署決策有何意義。 基準測試的差距確實存在但幅度窄；真正塑造日常工作的是每個項目為你預設的工作流程。Arch 提供最大程度的控制權與最少的維護抽象層。CachyOS 將注重效能的用戶原本要手動進行的調校工作提前完成。Omarchy 4.5 則針對首次開機即擁有連貫工作環境作優化。對於要與其他事務一併維護的 home lab 或工作站而言，維護負擔的差異往往比圖表排名更重要。
無論選擇哪條路線，須注意三者均採用 rolling-release 模型：沒有年度升級窗口需要規劃，但有持續的增量變更需要驗證——這對任何長期管理系統的人而言都是考量因素。透過網絡下載 package 的用戶，亦可將 package manager 指向地理位置更近的 mirror，以縮短下載延遲，這是一項不論選擇三者中哪一個均可採用的一般性優化。
完整的各項測試結果，包括每個 workload 背後的具體差值，可在 Phoronix 原始評測 查閱。