Maintenance for the week of November 25:
• PC/Mac: NA and EU megaservers for maintenance – November 25, 4:00AM EST (9:00 UTC) - 7:00AM EST (12:00 UTC)
• Xbox: NA and EU megaservers for maintenance – November 27, 6:00AM EST (11:00 UTC) - 9:00AM EST (14:00 UTC)
• PlayStation®: NA and EU megaservers for maintenance – November 27, 6:00AM EST (11:00 UTC) - 9:00AM EST (14:00 UTC)
• PC/Mac: NA and EU megaservers for maintenance – November 25, 4:00AM EST (9:00 UTC) - 7:00AM EST (12:00 UTC)
• Xbox: NA and EU megaservers for maintenance – November 27, 6:00AM EST (11:00 UTC) - 9:00AM EST (14:00 UTC)
• PlayStation®: NA and EU megaservers for maintenance – November 27, 6:00AM EST (11:00 UTC) - 9:00AM EST (14:00 UTC)
Freezes where game sound keeps running but game is hung, graphics driver and display errors
UnseenCat
✭✭✭✭
For some time now I've been trying to solve a problem where the game simply freezes but sounds -- including UI sounds when you click or activate controls -- go on working. The game has to be forcibly quit form the Task Manager. The Event log will show an "Event ID 0" error from the graphics driver and immediately followed by a "Display" Warning with Event ID 4101 stating the display driver "stopped responding and has successfully recovered."
The graphics card is running on its stock settings (Nvidia, in my case).
It happens whether addons are present or not.
It happens whether the CPU is running on overclocked settings or not (and overclock is tested stable).
It happens whether memory is set to its XMP profile or not.
Sound familiar?
After some searching regarding fixes for crashes in another game (Control under DX12) I ran into another older thread here on the ESO forums where similar crashing had been extensively examined and the results posted -- and the resolution was to underclock the GPU slightly. Interesting. I tried it and it did indeed stop the kinds of freezes/crashes I was experiencing in both Control and ESO.
Underclocking the GPU by 50MHz worked on my system (the original posted solution was underclocking by 90MHz), but it got me thinking -- there's something else not quite right going on, because the crashes happen even when the graphics driver's power setting is set to "maximum performance" mode. And the computer's power supply is over-specced for the load being put on it; monitoring it with other software never showed any power issues.
So what's happening? It seems that the default power limits for the GPU are actually set too low for some games, and will cause the GPU to under-perform at even its stock speed -- leading to the freezes and graphics driver/display errors.
The solution is to use a tool like MSI Afterburner, EVGA Precision X1, or your GPU's included tweaking tool to increase just the GPU power and voltage limits. In the case of my GPU, a 2080 Super, they can safely be increased to their maximum limits because there's a had cap set in VBIOS that prevents them from ever going above a safe level. The GPU temperature limit was left alone at its default, and the cards memory and GPU speeds were left at the card's defaults. And the freezing/hanging completely vanished. There's no increase in heat on the card either -- it's simply allowed to run at the GPU manufacturer's full range of power and voltage as designed.
So why would a graphics card fail to perform adequately under its default settings for power and voltage limits? My best guess is that defaults were set up to be less likely to incur thermal throttling in pre-built and popular cases which may have limited airflow, and therefor GPU temperatures might climb very quickly under load. You won't have that problem if you PC's case is well-ventilated.
So if you're experiencing these kinds of freezes/hangs and related errors, it may be worthwhile to try increasing the GPU voltage and power (wattage) limits. You may not even need to run them wide-open if you're concerned with saving power. Just monitor your GPU temperature and make sure it stays below 84 degrees or so. Mine never leaves the mid-60s no matter what.
This isn't so much an ESO problem as it is a design/power settings issue with GPU manufacturers, since it's affecting a number of games.
Your mileage may vary; proceed with caution, of course. But these are considered "safe" settings to work with particularly on modern graphics cards, and their effect is to improve stability rather than introduce potential problems like GPU overclocking can.
The graphics card is running on its stock settings (Nvidia, in my case).
It happens whether addons are present or not.
It happens whether the CPU is running on overclocked settings or not (and overclock is tested stable).
It happens whether memory is set to its XMP profile or not.
Sound familiar?
After some searching regarding fixes for crashes in another game (Control under DX12) I ran into another older thread here on the ESO forums where similar crashing had been extensively examined and the results posted -- and the resolution was to underclock the GPU slightly. Interesting. I tried it and it did indeed stop the kinds of freezes/crashes I was experiencing in both Control and ESO.
Underclocking the GPU by 50MHz worked on my system (the original posted solution was underclocking by 90MHz), but it got me thinking -- there's something else not quite right going on, because the crashes happen even when the graphics driver's power setting is set to "maximum performance" mode. And the computer's power supply is over-specced for the load being put on it; monitoring it with other software never showed any power issues.
So what's happening? It seems that the default power limits for the GPU are actually set too low for some games, and will cause the GPU to under-perform at even its stock speed -- leading to the freezes and graphics driver/display errors.
The solution is to use a tool like MSI Afterburner, EVGA Precision X1, or your GPU's included tweaking tool to increase just the GPU power and voltage limits. In the case of my GPU, a 2080 Super, they can safely be increased to their maximum limits because there's a had cap set in VBIOS that prevents them from ever going above a safe level. The GPU temperature limit was left alone at its default, and the cards memory and GPU speeds were left at the card's defaults. And the freezing/hanging completely vanished. There's no increase in heat on the card either -- it's simply allowed to run at the GPU manufacturer's full range of power and voltage as designed.
So why would a graphics card fail to perform adequately under its default settings for power and voltage limits? My best guess is that defaults were set up to be less likely to incur thermal throttling in pre-built and popular cases which may have limited airflow, and therefor GPU temperatures might climb very quickly under load. You won't have that problem if you PC's case is well-ventilated.
So if you're experiencing these kinds of freezes/hangs and related errors, it may be worthwhile to try increasing the GPU voltage and power (wattage) limits. You may not even need to run them wide-open if you're concerned with saving power. Just monitor your GPU temperature and make sure it stays below 84 degrees or so. Mine never leaves the mid-60s no matter what.
This isn't so much an ESO problem as it is a design/power settings issue with GPU manufacturers, since it's affecting a number of games.
Your mileage may vary; proceed with caution, of course. But these are considered "safe" settings to work with particularly on modern graphics cards, and their effect is to improve stability rather than introduce potential problems like GPU overclocking can.
1