Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Or if you are logged into your account, then you can see the result as you can always see your own results - even hidden ones. Just can't share them if the hardware is unknown.
I tried to search and fix any 6650XT models that I could find so it is possible the result is now visible if you load up a saved result from Documents/3DMark/ and view it online again. If not, post a link or email it to us.
Steps on this page: https://support.benchmarks.ul.com/support/solutions/articles/44001954929-enumdisplaydevicesa-failed
not help. Restart also not always help. What interesting all other software (games, benchmarks) work without any issues. Tried to clean install Windows 11 (may release with latest updates). Tried 2 different Nvidia drivers version 516.40 and previous one. Tried replace display port cable with other ones. Tried to use different display port on videocard. The problem keeps repeating. It may work without any issues until restart. After restart it may show that error again (not always). Turn off. Turn on. Working. Restart. Working. Restart again. Error. Moreover, the pattern is not clear. No issues (visual) with monitor. No issues with other software. Looks like 3DMark issue itself.
What can you recommend? Could this be some sort of hardware issue with the monitor? I've never seen this problem on my previous monitor. On the other hand, the rest of the software works. Strange.
I tried some another things. I changed monitor again to another model Acer XV272M. I even tried different videocard. Did clean install Windows 11. I can reproduce this bug again and again. I think I exclude everything I can from my configuration (monitor, videocard, OS, drivers, DP cable). I think this is pure 3DMark issue perhaps after some update. I never have issue like this before after 120+ hours of this software usage. I literally made test on different independent machine with different monitor. I didn't tried reproduce it on Windows 10.
Let me clearly explain how to reproduce it. Restart PC. Right after Steam icon shown in tray start 3DMark and immediately switch to "TimeSpy" from "Recommended benchmarks" zone. You'll see "Display: NoDeviceString (\\.\DISPLAY1)". After trying start test you'll get: "EnumDisplayDevicesA() failed". Please note that this may require 3 - 5 restarts to catch this bug.
Now interesting part. Let say we catch this issue. If after that close 3DMark, run any game, exit game and run 3DMark again this bug will gone. It will show "Display: Generic PnP monitor (\\.\DISPLAY1)".
Perhaps issue related to some obsolete windows api function to get monitors list. Because games and other software doesn't have such issues. Please investigate situation.
Also, this is a completely normal Windows APi call we are using
https://docs.microsoft.com/en-us/windows/win32/api/winuser/nf-winuser-enumdisplaydevicesa
Will try and report. Thank you.
Is that possible to use it in terms of some `test.cmd`? I can run it in console and report results when issue with 3DMark happens.
Monitor details (there maybe more than one if you have multiple monitors) looks like this
"deviceId" : "\\?\DISPLAY#ENC1887#5&762f02f&0&UID4352#{e6f07b5f-ee97-4a90-b076-33f57bf4eaa7}",
"deviceKey" : "\Registry\Machine\System\CurrentControlSet\Control\Class\{4d36e96e-e325-11ce-bfc1-08002be10318}\0001",
"deviceString" : "S2431W Digital",
"attachedToDesktop" : true,
"primaryDisplay" : true,
"deviceName" : "\\.\DISPLAY1",
"width" : 1200,
"height" : 1920
And to clarify, this runs before 3DMark UI opens and data from that is then used for the rest of the time. The data is not re-read unless you restart 3DMark.
Seems there are no connection between time and bug. I tried to wait for 5 minutes and 1 hour. Re-run 3DMark after several periods of time not lead to different results.
Thank you. This help a lot for debugging issue. I run this when catch bug and it show next results for "monitors" json section:
Numbers of run doesn't matter it show same result always (tried to run under admin and user credentials). But if I run some game or click on Desktop with RMB and select Display Settings and close it (without changing anything there) this tool result change to:
What interesting. In current user session bug never happen again until restart PC and as I say earlier this happened not after each restart (3 - 5 needed to catch bug with this console tool).
Perhaps some Windows 11 bug after some recent release. Never saw this before. Will try on fresh installed Windows 10 at weekend.
It is sad that from hardware point of view nothing can be explained there. My last experiments was new monitor and connecting to integrated processor video (i5 12400). Same situation. Anyway nothing critical, just strange behavior.
But I have personally been testing power performance and there are ways measuring how much power the CPU and GPU uses during a benchmark. So why not include that as an option? Either as just a common statistic for benchmarks or maybe as its own benchmark.
What I was testing was how well different GPUs would perform under different settings, when overclocking and underclocking to see how different systems compare at different power levels. How much power efficiency you get from a new generation.
Especially now that Valve has released the Steam Deck and with other gaming handheld devices like the ones from GPD... its obviously something that is here to stay and for these devices it seems it would be a very nice feature to be able to compare power efficiency between such devices and on the device itself using different settings.
As far as I can tell it should not be that complicated to add such a measurement to the benchmarks. I ran hwinfo64 to measure it between my RTX2080ti and my RTX3080ti, and now I am thinking of running some similar benchmarks on my GPD Win Max with and without an external GPU. I found it interesting that when limiting the 2080ti and 3080ti to 100W the 2080ti would perform better than the newer GPU, but the 3080ti would quickly gain the upper hand at higher watts.