I updated my NVIDIA driver to 387.92 and now API Overhead Test hangs at the start of the Vulkan test. Why?
This appears to be a problem with the test. It occurs only if you are running with the default full screen mode and the resolution of the test is different than the resolution of your desktop. As a workaround, you can temporarily switch your desktop resolution to match the test resolution (720p) and the test will work. Alternatively you can revert to a NVIDIA driver that is older than 387.92. We are working to figure out the root cause, but most likely this will require a benchmark patch. It is also possible that other GPUs can run into this same issue as their Vulkan drivers receive updates.
I updated from Windows 8 to Windows 8.1 or from previous Windows version to Windows 10 and now 3DMark refuses to start. Incompatible?
3DMark is compatible with Windows 8.1 and Windows 10. The problem is that when updating to a new version of Windows, the update removes the required Futuremark SystemInfo service from Windows and 3DMark cannot start the required service on startup. Please uninstall Futuremark SystemInfo, then either reinstall 3DMark or install the latest stand-alone Futuremark SystemInfo (available here) and it will correct the issue.
This issue can also happen when installing major Windows 10 updates (for example, Windows 10 November 2015 update) that are effectively performing an in-place OS upgrade.
I'm experiencing total system hang during a test or test exits with "process exited unexpectedly" error.
Most common cause for this is that either you system does not support High Precision Event Timer (HPET) or it is disabled in BIOS. You can test this by going to help tab and unticking "SystemInfo Hardware Monitoring". If the problem goes away, check if your motherboard BIOS has the option to enable HPET and enable it. If problem persists even with HPET enabled when you re-enable Hardware Monitoring, contact Futuremark Support and include full details about your hardware configuration.
If you are getting "process exited unexpectedly" immediately at the start of the first test and you meet the minimum requirements, this may be due to monitor scaling settings. You can try switching between "stretched" and "letterboxed" on the help tab, try overriding a different output resolution in the help tab, or go to your video driver settings and set display scaling to be done on the GPU.
If neither of these apply and you are getting "process exited unexpectedly", then it may be an indication of a video driver issue. Ensure your video drivers are up to date and if still experiencing problems, contact Futuremark Support and include details about your hardware configuration and background applications you have running on the system.
I'm having problems starting the Steam version of 3DMark
Verify that all pre-requisites installed properly. 3DMark requires that Futuremark SystemInfo and NET Framework 4.5 are properly installed and normally Steam version of 3DMark installs these. However, if installation fails, 3DMark can then fail to start. If problem persists, contact Futuremark Support and include details about your hardware configuration and background applications you have running on the system.
I'm having problems viewing results online ("View result online" doesn't do anything or an error comes up)
Check if you have antivirus software that blocks 3DMark from opening a browser. Reportedly at least Avira AV will prevent 3DMark from opening the default browser to display results. Temporarily disable such software and/or contact your security software support and ask for assistance. If this is not the cause, make sure you actually have a default browser set in Windows.
Also note that 3DMark automatically uses the proxy settings of Windows (which are configured in Internet Options) - if you are using a web proxy server that requires authentication, 3DMark cannot contact www.3dmark.com and online features of 3DMark cannot work with such configuration.
I'm having problems getting Hardware Monitoring to show temperatures and clock speeds (Advanced Edition)
This was a known issue with 3DMark v1.1 and is corrected in the current version - install latest update to 3DMark. Note that if you are seeing some but not all of the data, this is normal - not all hardware configurations provide all different forms of monitoring data.
I'm having problems with SLI/Crossfire performance and I have more than two video cards
Note that Multi-GPU configurations with more than two cards use up additional video memory due to resources that have to be duplicated on every card. For Fire Strike we recommend at least 1.5GB per card for 3-4x SLI/Crossfire and Fire Strike Extreme we recommend 2GB or more per card. The tests should run even when using cards with less memory but there will be a performance penalty as video driver has to use main RAM to extend the video memory.
I'm having stability issues when running loop testing and get error about display resolution changing
While the loop tickbox in Custom tab does disable most of the anti-tampering checks, when loop testing for hours in full screen mode, the run can be interrupted if some background application draws a popup or task bar notification as it causes a switch back to the desktop view. 3DMark cannot recover from such display mode change.
For very long stability stress tests we recommend testing using the windowed mode. Windowed mode is not interrupted by task bar popups or other interrupts from background applications, only actual driver crashes or general system instability will stop the test. When looping all parts of a test, there are loading screens with light load between each part. For maximum load set the stress test loop to loop a single part of a test - Fire Strike Game Test 1, for example - and the loop will be continous without loading screens. Note that when looping, the run will never produce a score.
Also we recommend disabling Scan Systeminfo from the help tab for loop testing - you are not going to get a result file anyway, so there is little point in gathering hardware data and in very long loops the hardware monitoring dataset (temperatures, clock frequencies etc.) can be so large that it could in theory cause stability problems.
Sky Diver Demo hangs with black screen (or very dark screen with faint outline of the main character) just as the scene switches from outdoors to the cave.
This issue was (mostly) fixed in 3DMark v1.5.884, but unfortunately it seems that it can still (rarely) occur. If you are still experiencing this issue, contact Futuremark Support with full details about your hardware configuration, including motherboard model and BIOS version.
Is 3DMark compatible with Windows 10?
Yes, as long as the latest SystemInfo module is installed (4.39 or later). Steam version automatically includes this update and standalone version will notify if your SystemInfo is out of date.
Is 3DMark compatible with Windows 8 / 8.1?
Yes, 3DMark is compatible with Windows 8 and 8.1
What version of DirectX does 3DMark use?
Every test in 3DMark except Time Spy uses a DirectX 11 engine. We then use a DirectX 11 property called Direct3D feature levels to create tests for specific versions of DirectX compatible hardware. Direct3D feature levels allow developers to use a DirectX 11 engine while still supporting existing hardware, from DirectX 9 to DirectX 11, from lightweight tablets to high-end desktops.
Ice Storm uses a DirectX 11 engine limited to Direct3D feature level 9_1, while Cloud Gate uses a DirectX 11 engine limited to Direct3D feature level 10_0. Sling Shot Extreme, Sky Diver and Fire Strike use a multi-threaded DirectX 11 engine.
Time Spy test uses a DirectX 12 engine.
Does 3DMark use DirectX 11.1?
Yes, but only in a minor way and with a fallback for DirectX 11 to ensure compatibility with the widest range of hardware and to ensure that all tests work with Windows Vista and Windows 7 as well as Windows 8.
DirectX 11.1 API features were evaluated and those that could be utilized to accelerate the rendering techniques in the tests designed to run on DirectX 11.0 were used.
Discard resources and resource views
In cases where subsequent Direct3D draw calls will overwrite the entire resource or resource view and the application knows this, but it is not possible for the display driver to deduce it, a discard call is made to help the driver in optimizing resource usage. If DirectX 11.1 is not supported, a clear call or no call at all is made instead, depending on the exact situation. This DX11.1 optimization may have a performance effect with multi-GPU setups or with hardware featuring tile based rendering (sometimes found in tablets and entry level notebooks).
16 bpp texture formats
The 16 bpp texture formats supported by DirectX 11.1 are used on Ice Storm game tests to store intermediate rendering results during post processing steps. If support for those formats is not found, 32 bpp formats are used instead. This optimization gives a noticeable performance effect on hardware for which the Ice Storm tests provide a suitable benchmark workload (tablets, entry level notebooks).
There are no visual differences between the tests when using DX11 or DX11.1 in 3DMark and the practical performance difference from these optimizations is limited to Ice Storm on very low end Windows hardware.
Does 3DMark use DirectX 11.2?
If Ice Storm is a DirectX 9 test, why doesn't 3DMark work on Windows XP?
Ice Storm uses a DirectX 11 engine limited to Direct3D feature level 9_1. 3DMark will not run on Windows XP PCs, since DirectX 11 is not available on Windows XP.
What is the difference between 3DMark Fire Strike and 3DMark 11?
3DMark 11 is our first DirectX 11 benchmark, released in December 2010. 3DMark Fire Strike, released in February 2013, is our second generation DirectX 11 benchmark and is the more modern and demanding of the two.
Default rendering resolution
- 3DMark 11 - 1280x720 (720p)
- 3DMark Fire Strike - 1920x1080 (1080p)
Extreme preset rendering resolution
- 3DMark 11 -1920x1080 (1080p)
- 3DMark Fire Strike - 2560x1440 (1440p)
While 3DMark 11's Entry preset can prove useful for testing entry-level hardware, in most cases we recommend using 3DMark Fire Strike for testing DirectX 11 based systems. Scores from 3DMark 11 and 3DMark Fire Strike cannot be directly compared.
What is the difference between 3DMark Sky Diver and 3DMark Fire Strike?
Together, Sky Diver and Fire Strike let you test the full range of DirectX 11 graphics hardware. Fire Strike is equivalent to testing a system with a modern DirectX 11 game on ultra-high settings. Sky Diver is more like running a game on normal settings. As a general guide:
- If a system scores less than 2800 in Fire Strike you should run Sky Diver.
- If a system scores more than 12000 in Sky Diver, you should run Fire Strike.
Please note that Sky Diver and Fire Strike scores are not directly comparable.
What is the difference between 3DMark Fire Strike and 3DMark Fire Strike Extreme
3DMark Fire Strike Extreme is a slightly modified version of 3DMark Fire Strike and is designed for high end multi-GPU (SLI/Crossfire) systems and future hardware generations. In addition to the higher rendering resolution, additional visual quality improvements increase the rendering load to ensure accurate performance measurements for truly extreme hardware setups.
Scores from 3DMark Fire Strike and Fire Strike Extreme should not be compared to each other - they are separate tests with their own scores, even if they share the same benchmark scenes.
What is the difference between 3DMark Fire Strike Extreme and 3DMark Fire Strike Ultra
3DMark Fire Strike Ultra is the exactly same test as Fire Strike Extreme, but with the rendering resolution increased to UHD (2160p). It is designed for testing high end multi-GPU (SLI/Crossfire) systems and future hardware generations for their performance at UHD (2160p) resolution.
Scores from 3DMark Fire Strike Ultra should not be directly compared to Fire Strike or Fire Strike Extreme scores - they are separate tests with their own scores, even if they share the same benchmark scenes.
What is the difference between 3DMark Cloud Gate and 3DMark Vantage?
3DMark Vantage and 3DMark Cloud Gate are both benchmarks for DirectX 10 capable hardware. The difference is in the engine powering each benchmark. 3DMark Vantage, released in April 2008, uses a DirectX 10 engine. In comparison, 3DMark Cloud Gate uses a DirectX 11 engine limited to Direct3D feature level 10_0. Using Direct3D feature levels is the modern approach to game engine design as it allows developers to use a DirectX 11 engine and still support older generation hardware all the way down to DirectX 9 level models.
We recommend using 3DMark Cloud Gate for testing DirectX 10 based systems. Scores from 3DMark Vantage and 3DMark Cloud Gate cannot be directly compared.
What is the difference between 3DMark Cloud Gate and 3DMark Sling Shot Extreme - they appear to be almost identical?
3DMark Cloud Gate is a test for DirectX 10 capable hardware. 3DMark Sling Shot Extreme (and the lighter ES3.0 variant, 3DMark Sling Shot) are tests designed for mobile devices. When porting Cloud Gate to mobile devices, several limitations of the mobile hardware required us to make changes to the test content to ensure reasonable frame rates. In short, 3DMark Sling Shot and 3DMark Sling Shot Extreme are more lightweight versions of the Cloud Gate scene, optimized for mobile devices. Cloud Gate and Sling Shot scores are not comparable to each other.
What is the difference between 3DMark Ice Storm and 3DMark06?
3DMark06 and 3DMark Ice Storm are both benchmarks for DirectX 9 compatible hardware, however, they use different engines and target different types of hardware.
- 3DMark06 - DirectX 9
- 3DMark Ice Storm - DirectX 11 limited to Direct3D feature level 9_1
- 3DMark06 - powerful desktop gaming PCs (in 2006)
- 3DMark Ice Storm - powerful mobile devices (in 2013)
For testing modern gaming PCs we recommend using 3DMark Sky Diver or 3DMark Fire Strike rather than 3DMark06, since such systems will almost certainly support DirectX 11. 3DMark Ice Storm is the ideal modern benchmark for DirectX 9 compatible mobile devices such as tablets, netbooks, ultra-portable notebooks and entry-level PCs. Scores from 3DMark06 and 3DMark Ice Storm cannot be directly compared.
My benchmark score from 3DMark Ice Storm fluctuates wildly between runs. Is something wrong?
3DMark Ice Storm was designed for mobile devices such as tablets and netbooks, so is a very lightweight test for any modern high end gaming PC. On high end systems the framerate can be over 1000 frames per second. On such systems 3DMark Ice Storm really isn't the correct test to run as even a tiny variation in the time it takes to render each frame can cause major changes in the framerate, leading to major fluctuation in the final score. As a rule of thumb, if you score more than 40 000 points in 3DMark Ice Storm, your system is "too good" for the test and you should instead use 3DMark Sky Diver or 3DMark Fire Strike to benchmark your system.
Also note that some power saving features may trigger when running Ice Storm. The benchmark load is so light that the video card may decide that lower operating frequency is fine and this may cause the score to fluctuate. If you are seeking the highest possible score in this test on a high end system, you may need to explicitly force your video card to operate in "high power" mode through video driver settings.
On very fast setups and on most multi-GPU (SLI/Crossfire) configurations, 3DMark Cloud Gate and 3DMark Sling Shot Extreme will also end up being limited by your processor and you will see similar (if smaller) effect. You should always use the benchmark most appropriate for the hardware you are benchmarking - the performance range of hardware that can be benchmarked using 3DMark is simply far too wide to cover it all with just a single test and that is the reason why 3DMark includes four different tests and the extreme preset for Fire Strike to cover all cases - even situations involving heavily overclocked processors and multi-GPU SLI/Crossfire setups cooled using liquid nitrogen or liquid helium.
I changed my desktop resolution and my scores changed slightly. Why?
3DMark works differently from 3DMark 11 to tackle the issue of different display sizes and resolutions. 3DMark always renders the tests internally at the resolution specified for each test ("offscreen rendering") and then scales the image to your desktop resolution before displaying it, preserving the aspect ratio (applying letterboxing if the desktop aspect ratio is not 16:9).
Internal rendering resolutions of 3DMark tests
- 3DMark Ice Storm - 1280 × 720 (720p)
- 3DMark Cloud Gate - 1280 × 720 (720p)
- 3DMark Sling Shot Extreme - 2560 × 1440 (1440p)
- 3DMark Sky Diver - 1920 × 1080 (1080p)
- 3DMark Fire Strike - 1920 × 1080 (1080p)
- 3DMark Fire Strike Extreme - 2560 × 1440 (1440p)
- 3DMark Fire Strike Ultra - 3840 × 2160 (4K UHD, also know as 2160p)
- 3DMark Time Spy - 2560 × 1440 (1440p)
If you alter your desktop resolution, you will also alter the target resolution for the final image scaling. On modern hardware, display scaling is almost "free" and in general the scores are comparable no matter what desktop resolution you use. However, when a test becomes limited by the CPU and framerates get very high (Ice Storm on almost all desktop hardware, Cloud Gate and Sling Shot Extreme on high end hardware), there can be a small impact to the score from image scaling. If you seek to obtain the absolutely highest score from a test (for overclocking contest purposes, for example), setting the desktop resolution to match the internal rendering resolution of the test you are running may provide a very slight increase to your score.
Note that the introduction of 4K UHD ("4K") displays have slightly changed the situation regarding image scaling. If you are running Windows desktop at a very high resolution but your GPU is not very powerful, the scaling may actually require considerable amount of GPU resources and video memory. Running with integrated graphics or low end discrete GPUs at 4K and above desktop resolution with 3DMark is not recommended and your score may be degraded by the performance penalty of image scaling.
As a workaround, you can override the output resolution of 3DMark from the help tab and instead let your monitor handle the lower resolution image as-is. For most tests under these circumstances, an output resolution of 1920 × 1080 (1080p) is recommended.
Help tab also includes an option to change between letterboxed and stretched scaling. This can be used, for example, to scale the image to fill a triple monitor Eyefinity setup. Aspect ratio is not preserved when stretched is selected. Note that this does not change the workload - it is still rendered in the internal rendering resolution, just scaled differently.
Custom tab (Professional and Advanced Edition) allows you to alter the internal rendering resolution for a custom run. The usual caveat applies that custom runs do not produce an overall score and sub-scores cannot be directly compared to scores obtained with default settings.
Why does the benchmark run so slowly? Why are my framerates so low?
Benchmarks are designed to place a heavy load on your PC to identify the weak points and bottlenecks in your hardware. It is sometimes the case that only small differences in visual quality can be seen from very large increases in the amount of processing required.
Fire Strike runs fine on my system, yet Fire Strike Ultra is a complete slideshow, why?
UHD (2160p) resolution adds an immense amount of work for the GPU - the number of pixels being rendered is quadrupled over the 1080p rendering resolution of Fire Strike and this factor alone will cut the performance with Fire Strike Ultra to around 25-30% of what you'd see with Fire Strike. Also note that due to the high resolution, Fire Strike Ultra requires 3GB or more dedicated video RAM. If your system does not meet this requirement, it is up to the video driver to juggle resources around by using much slower main RAM as extension of the dedicated video RAM. This means the performance is greatly hampered on systems with too little dedicated video memory. You may also experience video driver crashes that ultimately boil down to the test running out of video memory. UHD (2160p) resolution is a tough challenge, even for the best GPUs out there.
Why are there MSI and Galaxy/KFA2 logos in the benchmark demos?
The logo placements are a special sponsorship arrangement between those companies and Futuremark. Advertising placements ensure that we can offer the free 3DMark Basic Edition alongside the paid editions.
I have issues with the SystemInfo scan at the start with 3DMark - how do I fix this?
First make sure you have the latest update to SystemInfo installed.
Alternatively you can try disabling SystemInfo scan by unticking "Scan SystemInfo" in the Help tab but do note that any results from such a run cannot validated as they are missing information on the hardware configuration.
If you are getting a complete system hang during SystemInfo scan, ensure that HPET is enabled in your BIOS settings. SystemInfo Hardware Monitoring requires HPET (High Precision Event Timer) to function. Also ensure that your motherboard BIOS is up to date.
If you are running into problems with installing the latest SystemInfo update, you may have corrupted files in Windows installer cache. Easiest way of cleaning these up would be to try to uninstall Futuremark SystemInfo using a third party uninstallation tool that can clean up any corrupt or conflicting "junk" left over in Windows caches. There are many tools out there that can perform this, one example would be Revo Uninstaller but there are also others.
I'm using Windows Vista and I can't get 3DMark to start?
This is most likely due to a missing Windows update that is required for DirectX 11 applications to run on Windows Vista. Download and install the Platform Update for Windows Vista from Microsoft.
If this update is present, do also note that AMD has ceased providing driver updates to Windows Vista (Catalyst 13.12 is currently the latest driver for Vista). Without a more recent AMD video driver available for Windows Vista, 3DMark is not guaranteed to work on such system configurations in all situations. We strongly recommend upgrading to a more recent version of Windows if you have AMD video hardware.
3DMark fails with a black screen. Help?
There appears to be a rare issue when updating from older 3DMark to latest version that can cause this. Workaround is to uninstall, then reinstall 3DMark. We've made sure that this should not happen with future 3DMark updates.
3DMark fails to load on my phone - it just stops during loading or loads the first test and then exits. Why?
The most likely cause is lack of memory. First, verify that your device has 1 GB (1024 MB) memory to meet the minimum requirements. Second, note that any applications that are running in the background also consume memory. We recommend closing all background apps before running the tests as such programs always consume some resources and may influence your benchmark score negatively.
3DMark for Android v1.0.1 update includes changes that should resolve this issue for any 1 GB device models. Verify that you have the updated version. Devices with less than 1 GB are officially unsupported but may be able to run the tests if you untick "run demo" in the settings.
3DMark crashes on my Android device. Help!
We've tried to flag all the incompatible devices on the Google Play store but as there are so many devices and variants, we may have missed some. Please help us improve the accuracy of our list by telling us about your incompatible device.
The most common compatibility problem is with devices, such as the Samsung Galaxy Nexus or Samsung Galaxy Tab 2, that have the TI OMAP 4430 or TI OMAP 4460 chipset. We don't currently have a workaround, unfortunately, but now that 3DMark is publicly available we hope that the manufacturers and operators offering devices with these chips will notice the compatibility issues and work with us to fix them. Note that unticking "run demo" in the settings may allow you to complete a run on OMAP-based devices but there will be graphics corruption in the physics test.
For all other devices and any other crash issues, please post on our support pages including as much detail about the problem and your device as possible.
The benchmark stopped running in the middle of the test. Why?
If you are using a compatible device, make sure that all other apps running in the background have been closed. You may also want to temporarily disable notifications as the benchmark run will end prematurely if it is interrupted by an external application.
What are the minimum hardware requirements for 3DMark Ice Storm on Android?
OS: Android 4.0 or laterMemory: 1 GBGraphics: OpenGL ES 2.0 compatibleStorage: 235 MB
What are the minimum hardware requirements for 3DMark Sling Shot on Android?
OS: Android 5.0 or laterMemory: 1 GBGraphics: OpenGL ES 3.0 compatibleStorage: 203 MB
Where can I find more technical detail about the tests?
Our 3DMark technical guide contains detailed descriptions of the tests and an explanation of how scores are calculated. The guide covers all platforms, including Android.
3DMark says I have a Mystery Machine. What does that mean?
This message simply means that we haven't mapped your device details in our database yet. There are more than a thousand compatible Android devices and we are working hard to map and verify them all.
Even if your device is not identified, your benchmark score is still perfectly valid for making comparisons. The model and hardware details won't be visible in the Device Channel until we update our device database. You can get the latest list by completely closing 3DMark on your device, then restarting 3DMark. We refresh the database every hour.
If you have a brand new device model, feel free to report the problem and include a link to a website detailing the hardware specs of your device.
Some of the hardware detail data shown on 3DMark for my device is incorrect.
We do our best to verify all data. There is a button at the bottom of the "My Device" page where you can report errors you have spotted to us for fixing. Please describe what you believe is incorrect.
Also note that some popular phone models can have very different hardware depending on where and when they were bought, even if they share the same model name. You may want to first verify your exact model so that you are absolutely certain the data is in error.
If you notice that we have mis-mapped a device (i.e. the data fits to another model of the same device but not yours), do mention that when reporting the issue.
My 3DMark scores seem to vary wildly between runs. Can I trust these results?
Some high-performance devices use aggressive thermal throttling to control the temperatures of the processors inside the device. As the device heats up it intentionally slows the processor to avoid overheating, which in turn leads to lower performance scores on successive runs.
To reduce this effect, we recommended waiting 15 minutes between runs to allow the device to cool down. You should also unplug the device during testing as the battery charging circuitry also generates heat.
You can also try to keep the device cool by running the test in a cold environment. We tested several devices in a freezer during our own testing!
Measuring the effects of thermal throttling is another interesting way of using 3DMark to compare the performance of different devices.
What's going on with the Device Channel score lists? I ran the test and scored better but my score is not there.
Device Channel is not a "Hall of Fame". Android version of 3DMark has far too little information about the hardware and settings of the phone to make any determination if the run is actually legitimate so we actively chose not to do any kind of high score lists. "My Device" shows your personal top scores and you can share results using social networks but there is no functionality to "compete" with high benchmark scores.
The Device Channel scores are averages of all recorded runs from users of that specific phone model, with unusually high or low results filtered out and with a minimum threshold of results before a device can appear on the list. Most popular models have thousands of runs so even if you do a super-overclocked run, the average most likely won't change - it's just one run among thousands and most likely isn't even counted in the average due to being an outlier result. The goal is to show how devices typically perform with stock settings.
I'm running a custom ROM and my device isn't recognized (or it is recognized incorrectly). Why?
Most results with custom ROMs simply do not have required model and manufacturer information present so we cannot tell what kind of hardware you have. 3DMark on mobile devices currently does not do any kind of hardware detection - it just reads the manufacturer and model information. This is also the reason why any overclocks do not show in "My device" and processor information only shows the possible maximum known clock speed for the processor model found in the device. Sometimes manufacturers limit the default clock speed to improve battery life and reduce heat.
3DMark is simply not designed to extract this information from the hardware. This also means that if you are using a custom ROM that has a "fake" device string, 3DMark can show incorrect information. Most common case seems to be that some Chinese Android tablets have ROMs available that claim that they are "gt-i9300" which is Samsung Galaxy S III (Exynos 4 Quad). Currently there is no workaround for this - if your device manufacturer / model information is incomplete or wrong, 3DMark can't identify it. Same is also true for various "knockoff" phones that may physically appear to look like a well-known model (Samsung Galaxy S III and S IV seems to be the most common target at the moment) but the actual internal hardware of the device is completely different. If it's not a genuine device, 3DMark most likely won't identify it at all.
In all these cases the benchmark score is still valid - benchmark workloads do not know which device you are using and the performance measured is still accurate, so the issue is mostly cosmetic.
"Ouch! Something has gone wrong.", "An unknown error occurred." or "An error occurred while running benchmark."
These are generic errors that comes up when something unexpected happens. These errors should include a full technical error message about the issue. If you have no idea what could have caused the problem, contact Futuremark Support with the full error message for further assistance.
"It seems that you're running Windows 7"
3DMark requires Service Pack 1 on Windows 7 to run. Please install Windows 7 SP1. We also recommend installing all recommended Windows Updates.
"It seems that you're running Windows Vista" or "It seems that you're running Windows XP or an older version of Windows."
On Windows Vista 3DMark requires Service Pack 2 and DirectX 11 Platform Update for Windows Vista from Microsoft. We also recommend installing all recommended Windows Updates.
There is no way to run 3DMark on Windows XP. If you are still using Windows XP (or even older version of Windows), you will need to upgrade your operating system to correct this issue.
"Window focus lost during benchmark run."
This error occurs if some background application interferes with 3DMark and tries to draw a popup window, an overlay or otherwise switches the window focus from full screen 3DMark view during a benchmark run. In such cases the score would no longer be accurate so the test is aborted.
Most common culprits causing this error include various instant messaging and email applications that give popups when a new message arrives. Turn off such applications and try again. In general it is recommended to close any unneeded background applications before benchmarking a system to ensure that all system resources are available during the test.
"Connecting to www.3DMark.com failed. Are you connected to the Internet?"
Online result validation and result viewing online both require internet connection. This error usually means that you are not connected to the internet or you cannot reach 3DMark.com server at all. 3DMark.com itself is almost certainly still online - you just can't reach it.
"Could not connect online." or "Service temporarily unavailable."
A subtly different problem - you have an internet connection and 3DMark.com can be reached but for some reason the server is down. This could be a temporary downtime due to deployment of of new updates to 3DMark.com result service - in such cases it should clear itself within a few minutes. Alternatively this could be an unplanned downtime and our administrators are already busy restoring service. If the problem persists for longer, contact Futuremark Support to ensure that we're aware of the problem.
"Online result validation failed."
Something is wrong with the 3DMark.com result validation service. If there is nothing out of the ordinary with your benchmark run and the problem persists, contact Futuremark Support with details about your hardware and save a copy of your result to a file (if using Professiona or Advanced Edition) in case it is needed to investigate the problem.
"Checking DirectX compatibility failed."
This is a highly unusual error message that would imply that DirectX is not properly installed or it simply refuses to provide information on the hardware available. Contact Futuremark Support with full details about your hardware configuration and your operating system for further assistance.
"DirectX 9/10/11/12 compatible hardware required."
A DirectX 9 compatible video card is the minimum for starting 3DMark and it is required for the Ice Storm test. There is a further requirement of DirectX 10 capable card for Cloud Gate, DirectX 11 capable card for Sling Shot Extreme, Sky Diver and Fire Strike tests and DirectX 11 capable card with DirectX 12 compatible drivers (and Windows 10) for Time Spy. If you know for a fact that you have the proper hardware and are still getting this error, it may be caused by corrupt video driver installation. Try completely removing your video drivers with a third party driver cleaning utility, then reinstall the latest available drivers and try again. Contact Futuremark Support with details about your hardware (including DXDIAG output text file) if you cannot resolve the issue and think you are getting this message in error.
"Loading result failed."
There is something wrong with the result file you tried to load. If you are certain that you have not tampered with the file, contact Futuremark Support and ensure you can provide the result file for analysis.
"Saving result failed." or "Exporting results to XML failed." (Professional Edition only)
These errors usually indicate that your disk is full or you do not have permission to write to the folder where you are trying to save or export a result. Correct the problem and try again.
"Test run was cancelled by user."
Yes, pressing ESC cancels the test. This is not really an error, more of a notification why the test ended.
"Invalid upgrade key"
First ensure that the key you are inputting is in this form:
YYYYY sections are alphanumeric strings that contain numbers (0-9) and letters. Note that letters I and O will never show up in this part of the key. Anything resembling an I is number 1 (one) and O is always 0 (zero).
If you still can't get your key to work, please send us an email and include a photo or scan of the key if it was printed on paper and tell us where you got it so we can assist further.
"Registering a professional licence requires an internet connection."
Connect your system to internet to resolve this error. If you are using 3DMark Professional Edition in an environment where no network connection is available, contact Futuremark Support for further assistance.
"This program requires Steam. Please restart Steam client and try again."
This error can come up only with the Steam version of 3DMark. Something is wrong with Steam on your system. For whatever reason, 3DMark cannot start Steam API and communicate with Steam client. If verifying local files does not solve this problem and it happens on all Steam editions (also 32bit), contact Futuremark Support or use a stand-alone version of 3DMark.
Updated December 7, 2016