HWiNFO64 Guide: Read Sensors, Detect Bottlenecks and Monitor Your PC
HWiNFO64. You open it for the first time and something unexpected happens.
Not the download. Not the install. Those are easy.
The unexpected thing is the moment the Sensors window opens and you see hundreds of rows of numbers, all updating at once, all with names you have never heard before. Things like CPU IA Cores Power. VRM MOS. CPU Ring Ratio. VDDIO2. CPU Core Voltages 1 through 16.
You expected a simple temperature and FPS readout. What you got looks like the cockpit of a fighter jet.
I have seen this reaction play out in forums dozens of times. A user on Tom’s Hardware wrote: “I opened HWiNFO and now I have a headache. There are literally 400 sensors and I have no idea what any of them mean.”
Another on Reddit posted: “Why does HWiNFO show 8 different CPU temperatures? Which one is the real one?”
And one on the HWiNFO forums: “I installed this to check if my GPU was overheating. Now I am more confused than before I started.”
Three different people. Same feeling. Completely understandable.
Here is the truth nobody tells you upfront: you are never supposed to read all of those sensors. The 400 numbers are not all meant for you. HWiNFO64 is a professional diagnostic tool that was built to give technicians every single piece of data from every piece of hardware. For gaming and bottleneck diagnosis, you only ever need about 10 of those numbers.
This guide tells you exactly which 10 they are. What each one means. How to use them to find out if your CPU is bottlenecking your GPU. How to detect thermal throttling before it destroys your FPS. And how to fix the three most common problems people run into with this tool.
By the end of this article, HWiNFO64 will feel less like a fighter jet cockpit and more like a dashboard with five clearly labeled dials.
Let us get into it.
What Is HWiNFO64 and Why Gamers Use It
HWiNFO64 is a free hardware monitoring tool built by a developer named Martin Malík. It reads live data from every sensor on your PC — your CPU, GPU, motherboard, RAM, and storage — and shows you every number, updated in real time.
It has been around since 1997. Over time it became the tool that serious overclockers, system builders, and hardware enthusiasts trust more than any other monitoring software. Not because it is easy to use — it is not. But because it is the most accurate and the most detailed.
For gaming specifically, HWiNFO64 does four things that other tools cannot match:
It shows per-core CPU temperatures and usage. This is critical for bottleneck detection. Task Manager shows you an average across all cores. HWiNFO64 shows you every core individually. One maxed out core at 100% while others sit at 20% is a CPU bottleneck that Task Manager would completely hide from you.
It shows throttling reasons. When your CPU or GPU slows itself down due to heat or power limits, HWiNFO64 tells you exactly why. Not just “performance dropped” but “GPU throttled due to thermal limit” or “CPU throttled due to power limit.” That is a level of diagnosis no other free tool provides.
It logs everything to a file. You can run a gaming session, come back, and review exactly what happened at every second. Temperature spikes, usage drops, clock speed crashes. All timestamped and saved.
It shows hot spot temperatures. For AMD RDNA 2, 3, and 4 cards (RX 6000, 7000, 9000 series), the hot spot temperature is what actually drives throttling, not the average GPU temperature. MSI Afterburner sometimes does not show this accurately. HWiNFO64 always does.
That is why people who are serious about understanding their hardware use HWiNFO64 even if they also run MSI Afterburner.
How to Download and Set It Up Correctly
HWiNFO64 is only safe to download from one official source:
Official download: hwinfo.com/download
There is also a trusted mirror at guru3d.com where it appears as a regular entry in their tools section.
Do not download it from any other site. There are third party pages that bundle it with adware.
Installer vs Portable — which one should you choose:
The installer version adds HWiNFO64 to your programs list and can start with Windows automatically. This is the better choice if you plan to use it regularly.
The portable version is a single .exe you can run from anywhere, including a USB drive, with no installation needed. This is better for diagnosing a friend’s PC or keeping on a USB for emergency use.
For regular gaming monitoring, download the installer version.
How to set it up for the first time:
Step 1: Run the installer and complete the installation.
Step 2: Open HWiNFO64. A welcome screen will appear with two checkboxes: Sensors Only and Summary Only. For your first time, leave both unchecked and click Run. This opens all three windows so you can see what everything looks like.
Step 3: After this first look, close HWiNFO64. Reopen it and this time tick Sensors Only before clicking Run. This skips the hardware summary window which you rarely need and goes straight to the live data that matters.
Step 4: In the main Settings menu, tick Start with Windows and Minimize to Tray on Startup. This means HWiNFO64 loads automatically every time you start your PC, sitting quietly in the background collecting data without getting in your way.
Step 5: Still in Settings, go to the Monitoring tab. Find any sensors related to disk SMART data (hard drive health readings) and uncheck them. These make HWiNFO64 spin up sleeping drives every time it reads sensors, which creates unnecessary noise and activity. Disabling them makes the tool faster and less intrusive.
You are now set up. HWiNFO64 is running in the background, collecting data, ready to show you exactly what your PC is doing.
The Two Windows You Need to Know
HWiNFO64 has multiple windows but there are only two that matter for gaming diagnosis:
Window 1: The System Summary
This is the overview screen. It shows your hardware components in a visual layout with basic readings. Think of it as a quick health check dashboard. You can see at a glance whether temperatures look normal, whether clocks are where they should be, and whether anything is obviously wrong.
Most of the time you will glance at this and close it.
Window 2: The Sensors Window
This is where everything important happens. The Sensors window is the long scrolling list of hundreds of readings. It is overwhelming at first glance but it is organized into sections:
- CPU section at the top (temperatures, voltages, power, usage per core)
- GPU section below it (temperature, hot spot, usage, clock speed, power draw)
- Motherboard section (VRM temperatures, fan speeds)
- Memory section (RAM temperatures and speeds)
- Storage section (drive temperatures and health data)
Each section has a header row you can click to collapse or expand. This is how you tame the chaos. Collapse every section except CPU and GPU and suddenly you have a manageable, focused view.
To collapse a section: click the small triangle or arrow icon on the left side of the section header.
After collapsing everything except CPU and GPU, your Sensors window shrinks from 400 rows to about 30. Those 30 rows contain everything you need.
The 10 Sensors That Actually Matter for Gaming
This is the section nobody else writes clearly. Out of hundreds of readings, these are the only ones that tell you something actionable about your gaming performance.
1. CPU Package Temperature
This is your CPU’s overall temperature. It is the average across the entire chip. You will see multiple CPU temperature readings in HWiNFO64 with names like CPU Die, Tctl, CPU Core Average, and CPU Package. The one labeled CPU Package is the most representative of your CPU’s overall thermal state.
Safe ranges: Under 80°C during gaming is excellent. 80 to 90°C is normal for modern Intel and AMD Ryzen 7000/9000 series chips. Consistently above 95°C means thermal throttling is happening or about to start.
Read our complete CPU Temperature Range guide for the full breakdown by CPU generation.
2. CPU Core Temperatures (Individual)
These are the temperatures of each individual CPU core. They matter because gaming almost always maxes out specific cores rather than the whole chip. If Core 3 is at 95°C while Cores 0, 1, 2, and 4 are at 65°C, your cooling paste may have dried and is no longer spreading evenly across the CPU die.
3. CPU Core Usage Per Core (THE most important CPU reading)
This is what Task Manager hides from you. When you look at overall CPU usage in Task Manager, you see an average. If your CPU has 8 cores and one of them is at 100% while the others are at 10%, Task Manager shows you approximately 22% average usage. That looks fine. It is not fine.
In HWiNFO64 you can see every core individually. One core pinned at 100% while others are idle is a single-thread CPU bottleneck. This causes stuttering even when average FPS looks acceptable and even when overall CPU usage looks low. This is the number one hidden bottleneck cause in 2026.
If you see this pattern, use our PC Bottleneck Calculator to confirm whether your specific CPU and GPU combination is mismatched and which upgrade resolves it.
4. CPU Package Power (Watts)
How much total power your CPU is consuming right now. This reading helps you confirm whether your CPU is running at full performance or being limited by power constraints. If your CPU’s rated TDP is 65W and it is only drawing 40W during a demanding game, a power limit may be cutting its performance.
5. GPU Temperature
The average temperature across your GPU chip. For most gaming sessions this should stay below 85°C. For safe gaming temperature ranges by GPU brand and generation, read our GPU Temperature Range guide.
6. GPU Hot Spot Temperature
This is the most misunderstood reading in HWiNFO64, especially for AMD card owners. The hot spot is the highest temperature recorded anywhere on the GPU die. It is always higher than the average GPU temperature. On AMD RDNA 2, 3, and 4 cards, hot spot temperatures of 90 to 105°C are completely normal even when the average GPU temperature shows 75°C.
The hot spot matters because it is what actually triggers thermal throttling on AMD cards, not the average temperature. If your game is stuttering every 20 to 30 seconds and you are only watching average GPU temperature, you may be missing the real cause.
A safe hot spot for most AMD cards: under 100°C. Above 105°C consistently means your thermal paste may need replacing or your case needs better airflow.
For NVIDIA cards, watch the junction temperature if shown. Same principle applies.
7. GPU Core Clock (MHz)
How fast your GPU is running right now. Compare this to your GPU’s rated boost clock. If the clock speed is dropping significantly during gaming while temperatures look normal, you likely have a power limit throttling issue, not a thermal one.
The tell-tale sign: GPU clock drops from 2400 MHz to 1500 MHz suddenly, then recovers. Temperature is fine. This means your GPU hit its power limit, not its thermal limit.
8. GPU Load (%)
How hard your GPU is working. Should be 95 to 100% during gaming for maximum performance. If this is consistently 60 to 70% while FPS is lower than expected, your CPU is the bottleneck. It cannot prepare frames fast enough to keep the GPU busy.
This is the exact reading that tells you whether to upgrade your CPU or your GPU. Low GPU Load with low FPS means CPU upgrade. High GPU Load with low FPS means GPU upgrade.
9. GPU Memory Used (MB)
How much VRAM your GPU is using right now. If this reading is consistently at or above your GPU’s total VRAM capacity, your game is swapping textures to system RAM. System RAM is much slower than VRAM. The result is stuttering and FPS drops that come and go as textures load and unload.
The fix for VRAM overflow is simple: lower texture quality settings in your game.
10. GPU Board Power Draw (Watts)
Total power your GPU is consuming. This combined with GPU Load confirms the full picture. A GPU at 99% load drawing near its rated TDP (for example 300W on an RTX 5080) is a healthy, well-utilized system. A GPU at 60% load drawing only 120W while FPS is poor confirms a CPU bottleneck is starving the GPU of work.
How to Use HWiNFO64 to Detect a Bottleneck
This is the practical process for using HWiNFO64 as a diagnostic tool during gaming.
Step 1: Set up your view
Open HWiNFO64 in Sensors Only mode. Collapse all sections except CPU and GPU. You now have a focused view of the 10 sensors above.
Step 2: Position HWiNFO64 on a second monitor or use logging
If you have two monitors, put HWiNFO64 on the second screen while gaming on the main one. Watch the readings live.
If you only have one monitor, use the log to file feature (covered in the next section) and review the data after your gaming session.
Step 3: Play your game for 15 to 20 minutes
Focus on the most demanding sections. An open-world area, a multiplayer match, a busy city environment. These are where bottlenecks reveal themselves.
Step 4: Match your readings to these scenarios:
🔍 HWiNFO64 Bottleneck Diagnosis Guide
Match your sensor readings to the scenario below. Each one has an exact cause and an exact fix.
Your system is healthy and GPU-limited. You are getting every frame your card can produce. To get more FPS you need a faster GPU.
Classic single-thread CPU bottleneck. Your GPU is waiting for one CPU core that cannot keep up. Use our PC Bottleneck Calculator to find the best CPU upgrade for your specific GPU.
Thermal throttling. Watch GPU Temperature and GPU Core Clock simultaneously. If temperature spikes and clock drops at the same moment, heat is slowing your GPU down. Clean the GPU fans. Improve case airflow. Consider repasting.
Power limit throttling. Your GPU has hit its TDP limit. Check Board Power Draw. If it is capped well below the GPU’s rated TDP, your PSU may be struggling or your GPU’s power limit needs adjusting in MSI Afterburner.
VRAM overflow. Textures are being swapped to system RAM which is much slower. Lower texture quality in your game settings. This specific stutter cannot be fixed any other way without buying a GPU with more VRAM.
Hot spot thermal throttling on AMD RDNA cards. Remember the average GPU temperature is not the issue here. The hot spot is. Clean the GPU thoroughly, check that thermal pads on VRAM chips are intact, and improve case airflow.
How to Log Data and Find Problems You Missed
The logging feature is the most underused and most powerful part of HWiNFO64. Most people run it as a live monitor only. That means they miss problems that happen briefly — a 2 second temperature spike, a 5 second clock speed crash, a momentary VRAM overflow because they were not watching at exactly the right moment.
Logging captures everything continuously and saves it to a file for review afterward.
How to set up logging:
Step 1: Open HWiNFO64 and the Sensors window.
Step 2: In the Sensors window, click the small logging button at the bottom of the window. It looks like a small graph icon with a play symbol.
Step 3: Choose a location to save the log file. Your Desktop or Documents folder is fine. The file will be saved as a CSV (comma-separated values) format that opens in Excel, Google Sheets, or any spreadsheet application.
Step 4: Launch your game and play your normal session for 20 to 30 minutes.
Step 5: Close the game and return to HWiNFO64. Click the logging button again to stop recording.
Step 6: Open the log file in Excel or Google Sheets.
What to look for in the log:
Create a simple chart with time on the X axis and GPU Temperature on the Y axis. Any sudden spikes above your normal range stand out immediately.
Create a second chart with GPU Load % and CPU Core 0 Usage % on the same axis. When GPU Load drops at exactly the same time that a CPU core spikes to 100%, you have a timestamped proof of a CPU bottleneck event.
Create a third chart with GPU Core Clock. Any drops from the rated boost clock tell you when and how often thermal or power throttling occurred.
This kind of analysis is what separates genuine diagnosis from guessing. One 30-minute gaming log can tell you more about your system’s behavior than weeks of live monitoring.
Common Problems and Exact Fixes
These are the issues that appear constantly across the HWiNFO64 official forums, Reddit, and tech support threads.
Problem 1: The Sensors window will not open or appears blank
This is one of the most reported issues on the HWiNFO64 forum. The window exists — you can see it in Alt-Tab — but it is invisible or blank.
The cause is almost always that the Sensors window was resized or moved off-screen on a previous session. Windows remembers the last position of every window. If you moved HWiNFO64 to a second monitor that is now disconnected, the Sensors window opens off-screen where you cannot see it.
Fix: Open the HWiNFO64 settings and click Reset Preferences. This resets all window positions to default. If that does not work, a user on the official forums found a reliable fix using the Registry Editor:
Open Run (Windows + R), type regedit, navigate to HKEY_CURRENT_USER, Software, HWiNFO64, Sensors, and delete the WndTopX and WndTopY values. This forces the window back to position 0,0 (top left of your main screen) on the next launch.
Problem 2: HWiNFO64 is causing FPS drops while gaming
This was reported on the official forum by multiple users. The PC performs normally until HWiNFO64 is open, then FPS becomes unstable.
The cause is almost always one of these sensor types: ASUS or MSI EC (Embedded Controller) sensors, disk SMART sensors, or GPU VRM sensors on certain cards. These sensor reads require the tool to poll hardware registers that can briefly interrupt the GPU driver.
Fix: In the Sensors window, right-click on the header of any disk SMART section and select “Disable Monitoring.” Do the same for any EC sensors from your motherboard brand. Restart HWiNFO64 and test again. Most users see the FPS drops stop completely after disabling disk SMART sensors specifically.
Problem 3: There are 8 different CPU temperature readings. Which one is real?
This confuses almost every new HWiNFO64 user. On modern Intel and AMD CPUs you will see readings like:
CPU Package, CPU Die (average), CPU Core 1, CPU Core 2 (and so on for each core), Tctl (AMD specific), and sometimes CPU CCD1, CPU CCD2 on multi-chiplet AMD processors.
The answer is simple. Use CPU Package for the overall CPU temperature. Use individual Core readings to check if specific cores are running hotter than others. Ignore Tctl because it adds an offset for motherboard compatibility and does not represent the actual chip temperature. Ignore CCD readings unless you have a multi-chiplet Ryzen CPU (like the 9950X) and want to compare the temperatures between chiplets.
For gaming diagnosis: CPU Package tells you if your CPU is running too hot overall. Individual core temps tell you if cooling is uneven. Per-core usage tells you if a single core is bottlenecking your game.
Problem 4: HWiNFO64 shows sensors I did not have before after a driver update
This is normal and expected. When you update your GPU or motherboard drivers, new sensors often become available because the updated driver exposes additional hardware monitoring registers that HWiNFO64 can now read. New sensors appearing after a driver update is a feature, not a problem.
If the new sensors are showing readings that seem wrong or are flagged in red, right-click on that sensor row and select Configure. You can set custom warning thresholds and rename sensors so they make more sense to you.
Problem 5: HWiNFO64 shows my GPU running at a low PCIe slot speed
For example, it shows PCIe x8 instead of PCIe x16. This can look alarming but is sometimes completely normal.
Cause 1: Some motherboards reduce the primary PCIe slot to x8 bandwidth when a second card, M.2 drive, or other device is installed in a secondary slot. Check your motherboard manual.
Cause 2: Some GPUs reduce their PCIe link width at idle or low load to save power. Check HWiNFO64 while actually running a demanding game. The slot speed should return to full x16 under load.
Cause 3: Your GPU may genuinely be in a secondary slot that only provides x8 lanes. Move it to the primary slot (usually the topmost PCIe x16 slot on the motherboard).
HWiNFO64 vs MSI Afterburner vs GPU-Z
You have probably seen all three recommended in various places. Here is when to use each one and what none of them can do alone.
| Feature | HWiNFO64 | MSI Afterburner | GPU-Z |
|---|---|---|---|
| In-game overlay | No (needs RTSS integration) | Yes, via RivaTuner | No |
| Per-core CPU temps | Yes, detailed | Yes, basic | No |
| GPU throttling reasons | Yes, detailed | No | No |
| Hot spot temperature | Yes, accurate | Partial | Partial |
| Fan curve control | No | Yes | No |
| Overclocking | No | Yes | No |
| Log to CSV file | Yes, comprehensive | Yes, basic | Yes |
| GPU specs detail | Moderate | Basic | Very detailed |
| BIOS backup | No | No | Yes |
| Ease of use | Complex | Medium | Easy |
| System resource use | Low | Low | Very low |
The honest verdict:
Use HWiNFO64 when you need the deepest level of system data. Per-core CPU temperatures, throttling reason codes, hot spot accuracy, VRAM overflow detection, and long-term CSV logging for post-session analysis. This is the tool for serious diagnosis.
Use MSI Afterburner when you want an in-game overlay showing real-time data while playing. It is also the only tool for fan curve control and GPU overclocking. For daily gaming monitoring, Afterburner with RivaTuner is more practical.
Use GPU-Z when you want detailed GPU specifications, want to save a BIOS backup before overclocking, or want to run a quick sensor check without opening a complex tool.
The combination that covers everything:
Run HWiNFO64 in the background for deep data collection and logging. Run MSI Afterburner for the in-game overlay and fan curve. Open GPU-Z only when you need to check specific GPU specs or save a BIOS backup.
All three running together use minimal system resources and do not conflict with each other.
When it comes to confirming whether your CPU or GPU is actually mismatched for each other, none of these three tools gives you a clear answer. They tell you what is happening at the hardware level. Our PC Bottleneck Calculator tells you why it is happening and exactly what to do about it.
Summary: Everything in One Place
HWiNFO64 has hundreds of sensors. You need about ten of them.
The 10 sensors for gaming:
- CPU Package Temperature — overall CPU heat
- Individual Core Temperatures — uneven cooling detection
- Per-Core CPU Usage — the real bottleneck detector
- CPU Package Power — confirms CPU is running at full performance
- GPU Temperature — average chip heat
- GPU Hot Spot — the real throttling trigger on AMD cards
- GPU Core Clock — confirms GPU is boosting properly
- GPU Load % — tells you whether CPU or GPU is the limit
- GPU Memory Used — detects VRAM overflow
- GPU Board Power Draw — confirms GPU is fully utilized
The bottleneck shortcut: Low GPU Load (50 to 75%) plus one maxed CPU core plus low Power Draw equals CPU bottleneck. Take those readings straight to our PC Bottleneck Calculator and find the exact upgrade that fixes it.
The throttling shortcut: GPU Load and GPU Core Clock both drop at the same moment temperatures spike equals thermal throttling. Clean your GPU. Improve airflow.
The logging shortcut: If problems are intermittent and you cannot catch them live, turn on CSV logging before a gaming session and review it in Excel afterward. Every spike is timestamped.
Frequently Asked Questions
Is HWiNFO64 free?
Yes, completely free for personal use. There is a Pro version with remote monitoring features for IT professionals but everything a gamer needs is available in the free version. Download only from hwinfo.com.
Does HWiNFO64 affect gaming performance?
It can, but only if certain sensors are enabled. Specifically, disk SMART sensors and some Embedded Controller sensors on certain motherboard brands can cause brief interruptions that show as FPS drops. Disable disk SMART monitoring in the Sensors window settings and the problem stops. For most users with standard gaming hardware, HWiNFO64 runs in the background with zero impact.
What is the difference between CPU Package and CPU Temperature?
CPU Package is the overall average temperature of the entire CPU chip. Individual CPU Core readings show the temperature of each specific processing core. For gaming, watch CPU Package for overall health and individual cores to detect uneven cooling or hotspots on specific cores.
Why does HWiNFO64 show 8 different GPU temperature readings?
Modern GPUs have multiple temperature sensors across different components of the card. GPU Temperature is the average chip temperature. GPU Hot Spot is the highest point on the die. GPU Memory Temperature is the VRAM temperature. GPU VRM Temperature is the voltage regulator modules. For gaming, watch GPU Temperature and GPU Hot Spot. Everything else is secondary.
What does it mean if HWiNFO64 shows GPU throttling?
HWiNFO64 has a sensor labeled GPU Performance Cap Reason or similar (varies by GPU brand and driver). When this shows Thermal, your GPU is slowing down due to heat. When it shows Power, your GPU hit its power limit. When it shows Utilization, the GPU is not being given enough work which means you have a CPU bottleneck. This is the most direct bottleneck indicator in the entire tool.
Can I use HWiNFO64 on a laptop?
Yes. HWiNFO64 works on laptops. Laptop sensor readings are sometimes less accurate than desktop readings due to manufacturer-restricted sensor access on some OEM firmware, but core readings like CPU temperature, GPU temperature, and usage percentages are reliable on most laptops.
Why are my CPU core temperatures very different from each other?
If some cores are consistently 10 to 15°C hotter than others, your thermal paste may not be spreading evenly across the CPU heat spreader. This is more common after repasting or on older CPUs where paste has dried unevenly. Replacing the thermal paste with a fresh application should even out the core temperatures.
HWiNFO64 is showing sensors in red. Is my PC about to die?
Probably not. Red highlighting in HWiNFO64 simply means a reading has exceeded a warning threshold that was set by default. Many of these default thresholds are conservative. For example, AMD Ryzen 7000 and 9000 chips running at 90°C will show red in HWiNFO64 by default even though 90°C is within AMD’s normal operating range for those chips. Right-click on any red sensor and select Configure to adjust the warning threshold to something more appropriate for your specific hardware.
One Last Thing Before You Go
HWiNFO64 rewards patience. The first time you open it, close it and come back a few days later after you have read something like this guide. The second time it looks less chaotic. By the third time, you will recognize your CPU section, your GPU section, and the 10 numbers you care about, and everything else will fade into background noise.
The most useful thing I can tell you is this: do not try to understand every sensor. Understand the 10 in this guide and you will have more useful information about your PC than 99% of people who game on the same hardware.
Watch GPU Load. Watch per-core CPU usage. Watch temperatures. Watch Power Draw. Those four tell you almost everything that matters.
And when those four readings point to a CPU and GPU that are mismatched for each other, that is exactly what our free PC Bottleneck Calculator was built to diagnose.
Now open HWiNFO64 and find out what your PC has been hiding.
Found something in HWiNFO64 you do not recognize? Drop a comment below with the sensor name and what it is showing. If it comes up enough we will add it to this guide.
