How do we define reasonable performance standards for mobile games across different devices and target FPS?
In mobile game optimization, one of the most common questions from development teams is: what is a “reasonable” performance target for a specific device? Whether it is PSS memory, texture memory, triangle count, Draw Calls, CPU module cost, or thermal limits, there is no universal answer. The challenge lies in device fragmentation, hardware capability differences, project goals, and engine behavior. This article explains how GameOptim built its Performance Standard System through device tiering, large-scale data analysis, and real-device validation, while integrating GOT Online and Gears into practical optimization workflows.
About GameOptim GameOptim helps Unity developers identify memory issues, rendering bottlenecks, and performance regressions through automated profiling and cloud based performance analysis. Explore more: 🌐 Website: www.gameoptim.com https://www.gameoptim.com/?fopt=blog 📘 Blog: www.gameoptim.com/blog/ https://www.gameoptim.com/blog/ 💼 LinkedIn: www.linkedin.com/company/gameoptim/ https://www.linkedin.com/company/gameoptim/ 🎥 YouTube: GO.PerformanceLab https://www.youtube.com/@GO.PerformanceLab 💬 Discord: GameOptim https://discord.gg/4Jh6hj9gRw ⭐ GitHub: GameOptim https://github.com/GameOptim/unity mobile performance guide 💻 Dev: GameOptim https://dev.to/gameoptim Summary Performance standardization in game development is not about fixed numbers—it is about dynamic thresholds based on device tier, target FPS, and project requirements . Key conclusions: Performance metrics must be evaluated differently on low end, mid end, and high end devices. Core modules such as rendering, memory, physics, UI, and animation require separate threshold systems. Target FPS directly changes acceptable CPU, GPU, and memory budgets. Standard values must be continuously updated using real world test data. Conclusion: Performance optimization becomes significantly more actionable when teams replace “experience based judgment” with standardized, device aware metrics powered by GameOptim, GOT Online, and Gears. Core Concepts Why Performance Standards Are Difficult to Define Development teams frequently ask: How low should PSS memory stay on a 4GB device? What is the safe upper limit for texture memory? How many triangles are acceptable on a Snapdragon 888 device? Is 6.8ms in too expensive? Is 55°C device temperature safe? These questions all point to one core problem: How do we define “normal” performance? This is the foundation of the GameOptim Performance Standard System. 1. Device Tiering: The Foundation of Standardization The first challenge is hardware fragmentation. There are tens of thousands of mobile devices globally, and hardware evolves every year. For example: Xiaomi 5S and Xiaomi 12 belong to completely different performance generations. A fixed Draw Call limit or triangle count cannot apply equally to both. To solve this, GameOptim created a Device Tiering System , which combines: Hardware specifications Latest SoC performance data Real world user distribution Massive test device samples Historical performance reports from GOT Online This allows GameOptim to automatically classify devices and assign appropriate performance standards. 2. Metric Extraction: Building the Performance Knowledge Graph Game performance is highly modular. Core modules include: Rendering UI Physics Loading Animation Particle systems Memory ALU utilization Bandwidth pressure Each module contains dozens of sub metrics. Examples: Rendering → Triangles, Draw Calls, Opaque Time, Transparent Time Memory → PSS, Texture Memory, Mono Heap, Lua Heap CPU → Animation cost, Physics cost, Logic overhead This forms a knowledge graph of performance metrics , continuously expanded through GameOptim and Gears profiling data. 3. Project Goals Directly Affect Standards Different games target different goals: 30 FPS casual games 60 FPS competitive games 120 FPS high refresh projects Different device strategies: Some projects support 2GB RAM devices Some abandon devices below 3GB RAM for visual quality This changes all budgets: CPU frame budget GPU rendering budget Memory usage budget GOT Online uses target FPS as an input to dynamically adjust recommended thresholds. 4. Real Device Validation Is Critical Theoretical values are not enough. GameOptim verifies standards through: Optimization expert review Test projects across Unity versions Different engine configurations Real device validation using Gears Large scale runtime synchronization through GOT Online This ensures performance standards are practical, not theoretical. Device Based Standard Examples The figure below shows performance recommendations for different mobile devices. 1 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/1.png Example analysis: Low end device: Meizu M6 Note Triangle peak: 238,000 Recommended by GameOptim: <150,000 Result: Triangle pressure exceeds the safe range Teams should locate high density models and simplify geometry Mid high end device: Huawei Mate 20 Triangle peak: 315,000 Recommended by GameOptim: <450,000 Result: Current usage is healthy Teams may increase visual detail if desired This shows why device aware standards matter . Memory Standard Examples Texture memory standards differ greatly between devices. 2 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/2.png Examples: Redmi 4X low end Texture memory peak: 136.9MB Recommended: <140MB Near the upper limit. Xiaomi 8 higher end Texture memory peak: 219.5MB Recommended: <280MB Safe range. This same principle applies to: PSS memory Mono memory Shader memory Lua memory These values are tracked and analyzed continuously in GOT Online. Target FPS Based Threshold Adjustment Performance standards change with FPS targets. For example: 30 FPS = 33.3ms total frame budget 60 FPS = 16.6ms total frame budget This directly affects: Animator budgets Physics budgets Render budgets 3 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/3.png GameOptim automatically raises high priority alerts when metrics exceed target based thresholds. Custom Threshold System Projects often require custom standards. GameOptim provides a user defined threshold system . Recommended Memory Usage Supports: Resource memory Lua memory Device RAM grading 4 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/4.png Recommended Engine Module Metrics Includes: Draw Calls Triangle counts Rendering module thresholds 5 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/5.png Recommended Runtime Information Includes: CPU time FPS Network usage Mono allocation 6 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/6.png These configurations help teams build project specific optimization rules. Scoring System Optimization The latest GameOptim scoring system has been iteratively improved. 7 https://uwa ducument img.oss cn beijing.aliyuncs.com/GameOptim/PA normalvalue/7.png Scoring consists of: 1. Runtime Performance 2. Project Resources 3. Package Resources Rules: Scores update whenever module metrics change If unchanged for 15 days, recalculation occurs automatically Continuous validation ensures scoring freshness This allows teams using GOT Online and Gears to maintain long term performance quality control. Best Practices Establish device tier based performance budgets Never use one standard across all devices. Set target FPS before optimization begins This determines CPU/GPU budget allocation. Build module specific threshold systems Separate standards for: Rendering Memory Physics Animation UI Use real device validation instead of emulator assumptions Gears provides accurate runtime data collection. Continuously update standards Hardware changes fast. Static standards become obsolete. Combine profiling + scoring + threshold alerts GOT Online provides an efficient workflow for this. Key Takeaways There is no universal “good” performance number. Device capability determines acceptable performance budgets. Target FPS fundamentally changes optimization thresholds. Performance standardization requires data, not intuition. GameOptim’s tiering + scoring + threshold system turns optimization into a measurable engineering process. GOT Online and Gears provide the infrastructure to maintain these standards continuously. FAQ Q1: What is a safe PSS memory value for a 4GB device? Generally, keeping peak PSS below 50–60% of total RAM is safer, but GameOptim standards adjust this by device tier. Q2: How many Draw Calls are acceptable on mobile? It depends on device tier, rendering complexity, and FPS target. There is no universal fixed number. Q3: Is 300k triangles too high? On low end devices, yes. On high end devices, often acceptable. Q4: How do I know if my Animator cost is too high? Measure it against your target frame budget and device tier benchmarks. Q5: Why do performance standards need continuous updates? Because hardware, engines, and player device distributions change constantly.