HWiNFO64 Guide: How to Monitor Your PC Like a Pro (2026)

If you only install one piece of software on your PC this year, make it HWiNFO64. I have been repairing computers professionally for 16 years at Cincinnati PC Repair LLC, and there is one tool that lives permanently on my USB drive regardless of what job I am heading to: HWiNFO64. Not a paid suite, not a proprietary OEM utility — a free, constantly updated monitoring application that tells me everything about a machine within two minutes of plugging in my drive.

HWiNFO64 reads CPU temperatures, GPU temps, SSD health data, fan RPM, voltages, and real-time power consumption — all from a single window. It has diagnosed more failing coolers, dying SSDs, and unstable overclocks for me than I can count. The problem is that the first time most people open it, they see hundreds of sensor rows and immediately close it. That is exactly why I wrote this guide. By the end, you will know which sensors matter, how to set up proper logging, how to run it alongside a stress test, and how to connect it to other tools for an in-game overlay. Let’s get into it.

What Is HWiNFO64?

HWiNFO64 is a free hardware monitoring and diagnostic tool developed by REALiX. It supports virtually every modern Windows system and is updated regularly to support new hardware — which matters because a tool that doesn’t recognize your CPU’s power reporting is essentially useless for deep diagnostics. You can download the latest version from the official HWiNFO website at hwinfo.com. There is both an installer version and a portable version; I keep the portable build on my USB drive because it runs without installation on any customer machine.

When you first launch HWiNFO64, you will see a startup dialog with three options:

• Summary Only: Opens a compact hardware overview — processor name, core count, memory speed, GPU model, storage devices. Good for a quick hardware inventory but not for live monitoring.
• Sensors Only: Opens the real-time sensor panel only. This is what you will use 90% of the time. Every temperature, fan speed, voltage, and power reading updates live.
• Run (neither box checked): Opens both the System Information window and the Sensors window simultaneously. Useful for a first look, but I generally go straight to Sensors Only for diagnostic work.

The System Information window is essentially a detailed hardware report. It breaks your system down by component — CPU, motherboard, memory, storage, GPU — and gives you specifics like stepping, revision, PCI bus topology, and memory subtimings. It is excellent for verifying what hardware is actually installed, especially when a customer tells me they “upgraded the RAM” and I need to confirm what slots are populated and whether the sticks are running in dual-channel mode.

For the purposes of this HWiNFO64 guide, we are focusing on the Sensors window. That is where the diagnostic power lives.

First Launch: Summary vs Sensors

The first time you open HWiNFO64, I recommend running the full launch (both windows) just to orient yourself. Let the System Summary window load completely — it takes 10 to 30 seconds while the tool scans all hardware. If you see a prompt asking about an embedded EC (Embedded Controller) sensor, click “Do not monitor.” On most consumer systems, the EC sensor causes a hang during startup and provides data that is either redundant or unreadable. Skip it.

Once the System Summary finishes loading, take a look through it. Confirm your CPU model, RAM capacity and speed, and your storage drives are all showing correctly. If something is missing or wrong here — say your SSD is not appearing — that is a hardware or driver issue worth investigating before you go any further. This window is also where I check whether a laptop’s RAM is soldered to the board or in a removable slot, since HWiNFO will show the memory type and slot configuration.

Now close the Summary window and focus on the Sensors window. You will see a long scrolling list organized by hardware group: CPU at the top, then GPU, then individual drives, motherboard voltage rails, and so on. It looks overwhelming — and it is until you know what to ignore. The rule I follow in my shop: for any diagnostic, you are looking at maybe 10 to 15 rows out of hundreds. Everything else is noise for most use cases.

Before you do anything else with a new install, go to Settings inside the Sensors window and set the global polling frequency to 2000 ms for general use or 500 ms when you need precision during a stress test. The default is usually 2 seconds, which is fine for casual monitoring. Dropping it below 900 ms for extended periods can add slight CPU overhead, so reserve sub-second polling for active testing sessions rather than leaving it running all day.

The Essential Sensors to Watch

This is the core of any HWiNFO64 tutorial. Here are the sensors I look at on every machine I work on, and what each one tells me.

CPU Package Temperature

This is the single most important temperature on the board. The CPU Package reading represents the hottest die temperature across the entire processor. HWiNFO shows it in four columns: Current, Minimum, Maximum, and Average. Under normal desktop load, a well-cooled processor should stay under 80°C. During a sustained stress test, anything under 85°C is acceptable for most desktop CPUs. Once you start seeing sustained Package temps above 90°C, throttling becomes likely — and above 95°C on most Intel chips or above 90°C on AMD Ryzen, you have a cooling problem that needs to be addressed immediately. If you are dealing with an overheating laptop, check out my guide on the best thermal paste for laptops — a repaste is often the fix for systems running hot.

The Min and Max columns are particularly useful. After you reset them (right-click the sensor window → Reset Min/Max) and then run a workload, the Max column captures the true peak the CPU hit during that session. That is the number I care about, not the current reading.

GPU Temperature

For discrete graphics cards, look for the GPU Temperature row in the GPU section. This is the core die temp. Most desktop GPUs are comfortable running up to 80–85°C under full gaming load. Some cards — particularly NVIDIA RTX 40-series — also report a GPU Hot Spot temperature, which reads the hottest point on the die rather than the average. Hot Spot temps running 10–15°C above the core temp is normal. If you see Hot Spot consistently exceeding 100°C, that card needs better airflow or a thermal pad replacement on the VRAM.

SSD and HDD Temperature

Scroll down to your storage devices in the sensor list. NVMe SSDs in particular run hot, especially in compact systems without heatsinks. I want to see NVMe drives staying under 70°C during sustained reads and writes. Above 80°C and many drives will thermal throttle, cutting transfer speeds in half. SATA SSDs and spinning hard drives should stay under 50°C during normal operation. If a hard drive is pushing 55°C or higher, airflow in that case is insufficient.

CPU VCore Voltage

The CPU VCore reading tells you what voltage the processor is actually operating at. On a stock system, this will fluctuate dynamically — Intel processors running in Turbo Boost will often hit 1.35–1.45V under short-duration loads, dropping back down under light loads. If you see VCore sitting abnormally high (above 1.5V on Intel, above 1.45V on AMD Ryzen under sustained load) without a deliberate overclock applied, check your BIOS settings. Some boards auto-apply aggressive voltage offsets that degrade CPU longevity over time.

Fan RPM Readings

HWiNFO reads fan speeds from your motherboard’s fan headers. You will see each header labeled (CPU Fan, Chassis Fan 1, etc.) with the current RPM. This is invaluable for confirming that fans are actually spinning and responding to temperature increases. I have caught more than a few “my PC is overheating” calls where one of the chassis fans had failed — it was running at 0 RPM while every other reading looked fine. Without HWiNFO, that is nearly impossible to spot without opening the case.

Power Consumption

Modern CPUs report actual power consumption in watts. Look for CPU Package Power or CPU Total Power in the CPU section. This tells you what your processor is actually drawing from the board. Paired with the CPU temperature, it gives you a complete picture of thermal performance. A CPU drawing 125W and sitting at 75°C has excellent cooling. A CPU drawing 95W and hitting 95°C has a serious cooling problem.

Setting Up for Stress Testing

One of the most valuable uses of HWiNFO64 is running it alongside a stress test to see exactly how your system behaves under sustained full load. This is the procedure I run on every PC I finish building or servicing before it leaves my shop.

Step 1: Reset Min/Max values. Right-click anywhere in the Sensors window and select “Reset All Min/Max/Avg.” This clears all historical data so you are working with a clean slate. If you do not do this, any peak temperatures from a previous session will pollute your max readings.

Step 2: Let the system idle for 2–3 minutes. Let HWiNFO collect baseline idle data first. Your idle CPU Package temps should be in the 30–50°C range depending on ambient temperature and cooling quality. Note these values — they are your baseline.

Step 3: Launch your stress tool of choice. For CPU testing, I use Cinebench R23 or R24 for a 10-minute sustained multi-core run, or Prime95 (Small FFT test) for maximum thermal stress. For GPU testing, FurMark is the classic choice — it pushes GPU temperature harder than almost any real-world game. If you want to test both simultaneously, running Prime95 and FurMark together will show you the absolute worst-case scenario for your cooling system.

Step 4: Watch these values in real time. During a CPU stress test, I am watching for:

• CPU Package temp climbing and stabilizing (good) vs. still rising after 5 minutes (bad — insufficient cooling)
• CPU Package Power — confirms the processor is actually under full load
• Any indication of thermal throttling, which HWiNFO exposes via the “CPU Throttling” or “Power Limit Throttling” sensors (these show True/False or 1/0)
• Fan RPMs ramping up in response to heat — if fans are not increasing under a 90°C CPU, your fan curve needs adjustment in BIOS

Step 5: Note the Max column after the test completes. The Maximum values in that column after a 10-minute stress run tell you the definitive thermal ceiling of the system. These are the numbers you compare against manufacturer specifications and safe operating limits. If a PSU-related issue is causing instability during your tests, that is a separate diagnosis — the voltage rails in HWiNFO can point toward power delivery problems worth investigating further.

Logging Sensor Data

Real-time monitoring is useful, but logging is where HWiNFO64 becomes genuinely powerful for diagnosing intermittent problems. Logging records every sensor value at every polling interval to a CSV file that you can open in Excel, Google Sheets, or a dedicated viewer like the free Generic Log Viewer.

How to Start Logging

1. Open the Sensors window in HWiNFO64.
2. Click the green icon with a “+” symbol in the toolbar (it looks like a paper with a plus sign). This is the logging button.
3. Choose a save location and file name for your CSV. I use descriptive names like “stress_test_2026-01-15.csv” so I can find them later.
4. Click Save. Logging starts immediately.
5. Run your workload — gaming session, benchmark, stress test, or just normal usage if you are hunting an intermittent issue.
6. When done, click the same icon to stop logging.

Analyzing the Log

Open the CSV in Excel or Google Sheets and look at the columns for your key sensors — CPU Package Temp, GPU Temp, and CPU Package Power. Sort or graph these over time. What you are looking for is temperature spikes that correlate with crashes or performance drops. A machine that hits 98°C for 3 seconds before freezing tells you exactly what the problem is. A machine that crashes with all temps looking normal points you toward something else — driver stability, memory errors, or a power delivery issue.

Logging is also the only reliable way to catch thermal issues in machines that only misbehave during specific workloads. I have had customers bring in systems that “randomly freeze” while gaming — by logging a gaming session and reviewing the CSV afterward, I found the pattern: every freeze happened within 30 seconds of the CPU Package hitting 94°C. The fix was a dried-out thermal pad on the VRM heatsink, something impossible to diagnose without the data.

Catching Intermittent Problems

For machines with problems that do not reproduce consistently, I set up logging and send the machine home with the customer with HWiNFO running in the background. A few hours of logged normal usage gives me a complete picture. This approach has saved me countless hours of bench time trying to reproduce conditions that only happen “sometimes.”

Customizing the Interface

Out of the box, the Sensors window shows every sensor the tool can read — which on a modern system can be 200+ rows. Learning to customize the view is what separates a productive monitoring setup from an overwhelming wall of data.

Hiding Irrelevant Sensors

Right-click any sensor row and select “Disable/Remove Sensor.” This hides it from the display without permanently deleting it — you can always re-enable everything from the Configure Sensors menu. I typically hide all individual CPU core temperatures except Package (I already know what they are), all SMART attributes for drives except the temperature and reallocated sector count, and any chassis intrusion or legacy sensor rows that are not relevant to the machine at hand. Getting from 200+ rows down to 30–40 focused sensors makes the window actually readable.

If you ever make a mess of your layout, go to Configure Sensors → Layout → Restore Original Order to reset everything back to default.

Color-Coding Thresholds

HWiNFO64 lets you set warning and critical thresholds that change the color of a sensor’s row — green for normal, yellow for warning, red for critical. To set these, right-click a sensor row and choose “Customize.” I set CPU Package temp to yellow at 85°C and red at 95°C. GPU Temperature goes yellow at 85°C and red at 95°C. This means I can glance at the sensors window during a stress test and immediately see if anything is in the danger zone without reading individual numbers.

These visual thresholds are also how I hand off a monitoring session to a customer who is not technically fluent. “If you see any red rows while gaming, stop and call me” is a much clearer instruction than “let me know if the CPU Package goes above 95°C.”

System Tray Icons

For ongoing monitoring without keeping the full Sensors window open, HWiNFO64 can push any sensor value into your Windows system tray as a small icon with a live readout. Right-click any sensor row → “Add to Tray.” I keep CPU Package temp and GPU temp in my tray at all times on my own workstation. Two small numbers in the corner of my screen tell me everything I need to know at a glance without switching windows. This is particularly useful during gaming when you want a temperature check without alt-tabbing.

HWiNFO64 + Other Tools

HWiNFO64 is genuinely more powerful when paired with other tools. Here are the three integrations I use regularly in my shop and on my own machines.

HWiNFO64 + RivaTuner Statistics Server (OSD Overlay in Games)

This is the integration I recommend most often to gamers. RivaTuner Statistics Server (RTSS) is the software engine that powers on-screen overlays in games — it is the same engine behind MSI Afterburner’s overlay. When you install RTSS and configure it to read from HWiNFO’s shared memory, you can display any sensor value as a live overlay inside any game. CPU temp, GPU temp, frame time, fan speed — all visible without leaving the game.

To set this up: Install HWiNFO64 using the installer version (not portable — shared memory requires the installer). Install MSI Afterburner and RTSS. In HWiNFO’s settings, enable “Shared Memory Support.” In MSI Afterburner’s monitoring settings, set each metric source to HWiNFO. Then configure your OSD in RTSS. The result is a fully custom in-game overlay pulling real data from HWiNFO’s sensor engine — which is more accurate than Afterburner’s native readings for many CPU metrics.

HWiNFO64 + MSI Afterburner

Even without RTSS, MSI Afterburner and HWiNFO64 complement each other well. Afterburner handles GPU overclocking, fan curves, and power limits. HWiNFO handles the monitoring and logging side with greater depth. I use Afterburner to set a custom fan curve for the GPU and then HWiNFO to verify that the fan curve is actually being applied correctly under load — the RPM readings in HWiNFO confirm the fan is ramping up the way the profile specifies.

HWiNFO64 + ThrottleStop

For laptop users, pairing HWiNFO64 with ThrottleStop is the go-to combination for diagnosing and addressing CPU throttling. ThrottleStop allows you to undervolt Intel mobile processors and adjust power limits, while HWiNFO provides the monitoring data to verify the effect. The workflow is: baseline test in HWiNFO, apply undervolt in ThrottleStop, run the same stress test again, compare the Max temperature and CPU Package Power readings. When I can drop a laptop’s sustained CPU temp by 12°C with a -0.1V undervolt, HWiNFO is what proves it.

Richard’s Quick Setup Guide for New Users

If you have made it this far, you have everything you need to use HWiNFO64 like a professional. Here is the quick-start checklist I go through every time I set up HWiNFO on a new machine:

1. Download: Grab the latest portable or installer version from hwinfo.com.
2. Launch in Sensors Only mode. Skip the EC sensor warning if it appears.
3. Set polling to 2000 ms for general use, 500 ms for stress testing.
4. Hide irrelevant sensors by right-clicking rows you do not need.
5. Set color thresholds: CPU Package and GPU temp yellow at 85°C, red at 95°C.
6. Add CPU Package temp and GPU temp to the system tray for quick-glance monitoring.
7. Reset Min/Max, then run a 10-minute stress test — Cinebench for CPU, FurMark for GPU.
8. Enable logging to CSV for any session where you need a record of what happened.
9. If you game: Enable Shared Memory Support and hook up RTSS via MSI Afterburner for an in-game overlay.

HWiNFO64 has been part of my toolkit since the early days of my repair business, and it has never let me down. Whether you are troubleshooting a thermal problem, validating an overclock, or just want to keep an eye on your system’s health over time, there is no better free tool available. No bloat, no paywalls, no subscriptions — just accurate hardware data, available any time you need it.

If you are dealing with unusual temperature spikes after reapplying thermal paste, be sure to check my guide on choosing the best thermal paste for your laptop — the compound itself makes more of a difference than most people expect. And if your voltage readings in HWiNFO are showing irregularities during load, your power supply may be struggling under demand — a topic worth investigating separately as part of a thorough system diagnosis.

Download HWiNFO64, spend 20 minutes with it, and you will understand your PC better than most people who have owned it for years.