**Introduction**
Modern workstations demand greater memory capacity due to intensifying content creation workloads like video editing, 3D rendering, motion graphics, and game development. Enough RAM is necessary to manage large datasets, caches, and calculations. However, rising memory prices and limited supply have challenged the PC industry, largely fuelled by the expansion of AI infrastructure requiring vast memory in data centers. As professionals set up new workstations, the crucial question becomes: **how much RAM is really necessary**? While reducing memory slightly may cut costs without performance loss, insufficient RAM can hamper workflows.
This guide will examine RAM usage and emphasize specific workflows requiring substantial memory, equipping you to determine current and future RAM needs.
**Why Have RAM Prices Increased?**
AI’s rapid growth in training and inference infrastructure heavily drives recent memory shortages. Training AI models involves handling enormous data batches while storing model parameters and caching data to retain a steady workflow, requiring large memory pools in AI servers to stage datasets efficiently.
As AI servers are equipped with hundreds of gigabytes or terabytes of system memory, and when multiplied across numerous global data centers, RAM demand surges dramatically. This has strained the global memory supply chain as manufacturers focus on server-class memory, escalating prices and reducing traditional desktop and workstation memory availability.
Despite higher costs, substantial memory configurations for workstations remain possible, but capacity planning becomes crucial to avoid overspending.
**What Happens When You Don’t Have Enough RAM?**
Understanding the consequences of lacking memory is essential when selecting a system’s memory capacity. When active applications collectively use more memory than available, the operating system resorts to virtual memory, or a pagefile, on a storage drive. This swap space frees up RAM by moving inactive data to the drive, but it introduces delays since storage is significantly slower than RAM. Heavy paging may result in:
– Sluggish applications
– Playback stuttering
– Prolonged render times
– Frequent pauses during data swaps
Excessive RAM shortages can render systems unresponsive. Prioritizing sufficient RAM to preclude frequent paging during routine tasks is ideal rather than maximizing RAM capacity.
**How to Monitor Your RAM Usage**
Observing your memory usage during typical workflows offers insights into how much RAM you really need. Rather than relying on general advice, this approach allows you to see your project’s specific needs.
**Windows Task Manager**
Task Manager, embedded in Windows since NT 4.0, provides a simple way to monitor memory:
– Right-click the taskbar and select **Task Manager**
– Go to the **Performance** tab
– Select **Memory**
Here, you can see:
– Total installed RAM
– Current memory usage
– Memory speed and configuration
– Committed and cached memory
Operating the Task Manager during your usual workload delivers data on whether your memory usage is near its limit.
**Understanding “In Use” vs. “Committed Memory”**
“In Use” means memory actively used by applications and the operating system. “Committed Memory” is the total requested by applications, which might be more than used to reserve resources. If the peak usage is 31.2 GB on a system with 32GB RAM, but committed memory is 44GB, there’s a risk of exceeding RAM and relying on pagefile.
**Ubuntu System Monitor**
For Linux users, Ubuntu’s System Monitor provides a graphical option akin to Task Manager. It displays memory usage as a number and percentage, with a 1-minute usage graph. Access it using the Super key and “System Monitor” or the command `gnome-system-monitor`.
Leave System Monitor running during demanding workflows to accurately gauge RAM needs. It shows Swap space usage, unlike Task Manager.
**Monitoring RAM Capacity for Your Workflow**
**Motion Graphics**
Memory needed for motion graphics varies with application, settings, and frame complexity. Adobe After Effects and Blackmagic Fusion use RAM differently for video previews. When RAM is full, After Effects caches rendered frames onto disk, while Fusion reprocesses frames not in RAM.
Resolution, frame rate, and bit depth influence data generated per frame, consequently impacting RAM usage for caching frames. Higher specs increase data per frame, limiting the number of storable frames, while lower specs allow more frames in RAM. Rendering multiple frames speeds playback but uses more memory.
**Photography**
RAM requirements for photo editing fluctuate based on tools, image types, and editing techniques. In Adobe Lightroom Classic and Photoshop, RAM acts as a cache for reading, writing, and storing edit data, but cache size varies. RAW files, containing unprocessed sensor data, demand more RAM due to the data-intensive demosaicing process.
An editor’s methods and tools impact cached RAM size. RAM-intensive tools include:
– **Camera Raw Details** for finer detail in demosaicing
– **Denoise** reduces noise by processing every pixel
– **Super Resolution** creates higher-resolution images
– Merge processes like panorama, HDR, and stacking
Layer compositing stacks additional data into RAM