Using the System Information Window in Process Explorer

Learn how to open and interpret Process Explorer's System Information window, including CPU, memory, I/O, network, disk, and GPU performance graphs.

Process Explorer's System Information window is a system-wide performance-monitoring dialog. It presents current values, recent-history charts, and category-specific tabs for investigating CPU, memory, I/O, network, disk, and GPU activity. Its purpose is similar to the performance-focused views in Windows Task Manager, while keeping the information alongside Process Explorer's detailed process list.

This guide assumes that you know how to navigate the running-process list and understand the basic meanings of CPU use, RAM, disk activity, and network traffic.

Open the System Information window

Open Process Explorer.
Select View > System Information.
Alternatively, press Ctrl+I.

You can leave this window open while examining the process list. This makes it possible to compare a system-wide spike with individual processes and then open the suspected process's properties for more investigation.

Tab	Charts or metrics	What the data represents	Best use during troubleshooting
Summary	Compact current and history views for CPU, memory, and I/O	A quick overview of the main resource categories	Find the resource category that needs closer investigation
CPU	Total CPU history, kernel-mode activity, and optional per-CPU graphs	Processor capacity used by applications, Windows, and drivers	Locate CPU spikes and detect uneven logical-CPU usage
Memory	System Commit, Physical Memory, and lower-section memory measurements	Committed memory and RAM usage	Separate commit pressure from physical-RAM pressure
I/O	I/O Bytes, Network Bytes, Disk Bytes, and operation counters	File, device, network, and local-disk activity	Find storage, device, or network-heavy workloads
GPU	GPU-utilization graphs	Use of graphics-processing resources	Relate elevated GPU activity to a supported workload

Read the Summary tab

The Summary tab is a compact overview of the principal system measures. Its paired-chart design places a current activity or value display beside a recent-activity history chart. The summarized categories are CPU, memory, and I/O-related activity.

Move the pointer over a history chart to reveal the time represented by the selected point. On CPU and I/O charts, the tooltip can also identify the process that was the largest consumer at that point in history. This is useful when a short spike has already ended by the time you begin investigating it.

Indicator	Meaning	Where shown
Red CPU area	CPU time spent in kernel mode	CPU activity graphs
Green CPU line	Total processor utilization	CPU activity graphs
Yellow commit line	Current commit-charge level	System Commit charts on the Memory tab

Interpret the CPU tab

Total CPU utilization and kernel mode

CPU utilization is the percentage of available processor capacity currently in use. The total value includes work performed by applications and work performed by Windows.

Kernel mode is the privileged execution mode used by the Windows kernel and device drivers. In the CPU graph, the red portion represents kernel-mode CPU time. The area beneath the green CPU line represents total processor use. A large red portion means that much of the processor activity is occurring in operating-system or driver work rather than only in application code.

Use CPU history to find a spike

Inspect the historical CPU chart and hover over the high point. The tooltip reports the corresponding time and can identify the process that led CPU consumption at that point. Return to the process list and inspect that process's CPU columns, threads, or properties.

Combined versus per-logical-CPU graphs

A logical CPU is a processor execution unit presented to Windows. It may be provided by a physical processor core or by simultaneous multithreading. The normal combined graph summarizes activity across all logical processors.

On systems with multiple logical processors, enable Show one graph per CPU on the CPU tab to display separate CPU Usage History charts. The combined graph answers “How busy is the processor system overall?” Individual graphs answer “How busy is each logical processor?”

For example, one logical CPU may be saturated while the others are mostly idle. Overall utilization can then appear moderate even though a single-threaded workload is constrained. Per-logical-CPU graphs expose this imbalance.

Interpret the Memory tab

System Commit

Commit charge is committed private bytes across processes plus paged-pool memory. Private bytes are memory committed for a process that is not shareable with other processes. Paged pool is kernel-managed memory that can be paged to disk when it is not needed in physical RAM.

The System Commit charts show committed-memory activity. The yellow commit line indicates the commit-charge level. Review this chart with the memory measurements in the lower section of the tab rather than treating the chart as a direct measurement of occupied RAM.

Physical Memory

Physical memory is the installed RAM and the amount currently being used by the system. The Physical Memory charts show how much of that RAM is in use. The lower memory metrics provide supporting detail that helps distinguish committed-memory growth from physical-RAM consumption.

High commit usage matters even when some RAM remains available. Commit represents memory that Windows has promised to back with RAM or suitable paging storage. If committed memory approaches the system's commit limit, applications may fail allocations or become unstable before every byte of installed RAM appears occupied.

To diagnose memory pressure, compare the System Commit chart, including its yellow line, with the Physical Memory chart and the lower measurements. A rising commit value points toward committed-memory growth; high physical-memory use points toward RAM pressure; both can occur together.

Interpret the I/O tab

Graphs

I/O Bytes measures throughput for file and device input/output. Network Bytes represents network-transfer activity. Disk Bytes represents throughput directed to local disks. Comparing these graphs helps distinguish broad device I/O from network traffic and local-storage traffic.

Hover over an I/O history spike to see its time and, when available, the process that was the leading consumer. Use that process as the starting point for process-level investigation.

Operation counters

The lower portion of the I/O tab contains operation counters. I/O and disk activity include Read, Write, and Other operations. Network activity includes Receive, Send, and Other operations.

These counters describe activity accumulated since the most recent refresh, not a permanent lifetime total. Interpret them as a recent activity sample and compare successive refreshes when looking for a sustained pattern.

Measurement or counter	Resource category	Interpretation
I/O Bytes	File and device I/O	Total throughput for input/output activity
Disk Bytes	Local disk	Throughput sent to local disk storage
Network Bytes	Network	Inbound and outbound transfer activity
Read operations	I/O or disk	Operations that retrieve data
Write operations	I/O or disk	Operations that store data
Other operations	I/O or disk	Operations that are neither reads nor writes
Network receive operations	Network	Operations receiving network data
Network send operations	Network	Operations sending network data

Use the GPU tab

A GPU is the graphics-processing hardware used for rendering and other parallel workloads. On Windows Vista and later supported systems, the System Information window can provide a GPU tab with GPU-utilization graphs.

Availability and displayed data depend on Windows support and the graphics environment, including the available driver and hardware monitoring information. When the tab is present, observe GPU utilization during the workload under investigation and correlate an elevated period with the process or application being tested. The tab may be absent or may expose less information on unsupported configurations.

Performance-investigation workflow

Start on the Summary tab and identify whether CPU, memory, or I/O is the constrained category.
Move to the matching detailed tab: CPU, Memory, or I/O. Use GPU when a supported graphics workload is involved.
Hover over the history chart at the time of the spike to obtain its timestamp and, for applicable CPU or I/O data, the leading process.
On multi-core systems, enable Show one graph per CPU when overall CPU usage does not explain poor responsiveness.
Use the process list and process properties alongside the system charts to confirm what the identified process is doing.

Example: find a short CPU spike

Open System Information and select the CPU tab.
Locate the high point in CPU history.
Hover over that point to read its timestamp and the reported largest CPU consumer.
Return to the process list and inspect that process.

Example: investigate uneven multi-core performance

Open the CPU tab.
Enable Show one graph per CPU.
Compare the charts for each logical processor.
Determine whether activity is concentrated on one logical CPU or distributed across the system.

Example: separate commit pressure from RAM pressure

Select the Memory tab.
Compare the System Commit chart and yellow commit line with the Physical Memory chart.
Review the lower memory measurements.
Decide whether committed-memory growth, physical-RAM consumption, or both explain the problem.

Example: identify storage or device activity

Select the I/O tab.
Compare I/O Bytes with Disk Bytes to determine whether activity is specifically local-disk related.
Inspect read, write, and other-operation counts.
Hover over an I/O spike to identify the leading process when that information is available.

Example: distinguish network transfer from disk transfer

Open the I/O tab during the slowdown.
Compare Network Bytes, Disk Bytes, and I/O Bytes.
Review network receive and send counts alongside disk read and write counts.
Focus the next investigation on the network-heavy or disk-heavy workload indicated by the charts.

Troubleshooting patterns

High CPU use with no obvious cause: use CPU history tooltips to identify the leading process, then compare total CPU activity with the red kernel-mode portion. High kernel-mode activity can indicate operating-system or driver work.
Moderate overall CPU use but a slow system: enable individual logical-CPU graphs and look for one heavily loaded processor. A single thread or uneven scheduling may be limiting the workload.
Insufficient-memory errors or unstable applications: compare commit charge and physical-memory charts and inspect the lower memory metrics. The system may be approaching its commit limit, consuming substantial RAM, or experiencing both conditions.
Pauses during file access: compare I/O Bytes with Disk Bytes and inspect disk read, write, and other operations. High disk throughput or operation volume can indicate storage pressure.
Slow downloads, synchronization, or remote applications: compare Network Bytes with Disk Bytes and inspect receive and send operations. Network transfer may be the bottleneck rather than local disk access.
Missing GPU tab: verify that Windows and the graphics environment provide supported GPU monitoring data. GPU monitoring is platform-dependent.