Crashes in ucrtbase.dll as of v3.2.0 and v3.2.1

[josersleal]

I know nothing else other than crashes for no reasons and on restart all the chats and related content are gone.

[manyoso]

You have to fill out more information otherwise this issue is useless. Please try harder or we're not going to be able to help.

[josersleal]

What kind of information you need? a log? where do I get it from? Also after that crash, its not possible to add more collections for localdocs. pS. the previous version was unusable, crashed constantly not even starting the app. So I decided to give it another try and this happened. let me know what you need

found the logs: this type of message is repeated ad-nauseam: "[Critical] (Thu Aug 1 18:02:55 2024): QFileSystemWatcher: FindNextChangeNotification failed for "path to collection folder" (Insufficient system resources exist to complete the requested service.) "

Also a windows event: Faulting application name: chat.exe, version: 0.0.0.0, time stamp: 0x66a54678 Faulting module name: ucrtbase.dll, version: 10.0.22621.3593, time stamp: 0x10c46e71 Exception code: 0xc0000409 Fault offset: 0x000000000007f6fe Faulting process id: 0x0x4990 Faulting application start time: 0x0x1DAE3F268220F17 Faulting application path: (path)\gpt4all2\bin\chat.exe Faulting module path: C:\Windows\System32\ucrtbase.dll Report Id: bb43b81e-c56f-42c8-8b6e-21f36a48f8b0 Faulting package full name: Faulting package-relative application ID:

[manyoso]

Start with specifying whether you're talking about the chat client or bindings. You've tagged this as 'bindings' but your references above seem to imply this is a crash with the chat client.

Next, give me detailed information about your system: operating system, hardware, etc.

Next, give detailed information on what version of GPT4All you're logging this against and how you installed it.

Next, give detailed information on how you're using the app:

1) Models you're using 2) Features you're using (like localdocs) 3) If localdocs then what kinds of collections and sizes

Don't assume that just because this is the experience you're having - which I'm sorry it isn't a good one - that others are also having that experience.

When writing a bug report the best way to help us help you is to provide enough detailed information so that we have a reasonable shot at finding the difference between what you're doing and what others are doing that explains the problem. The holy grail is for us to be able to actually reproduce what you're experiencing. Generally speaking, once we've been able to reproduce a problem the fix becomes easy. The hard part is usually reproducing the problem.

Thanks!

[josersleal]

Start with specifying whether you're talking about the chat client or bindings. You've tagged this as 'bindings' but your references above seem to imply this is a crash with the chat client.

Next, give me detailed information about your system: operating system, hardware, etc.

version 3.1.1 windows

Next, give detailed information on how you're using the app:

Models you're using ollama3.1 8B instruct 128k

Features you're using (like localdocs) using localdocs all settings at default normal collection with pdfs, about 300-500 in doifferent collections. the one I was using had about 99 pdfs

If localdocs then what kinds of collections and sizesStart with specifying whether you're talking about the chat client or bindings. You've tagged this as 'bindings' but your references above seem to imply this is a crash with the chat client.

I guess its the chat client

Next, give me detailed information about your system: operating system, hardware, etc.

OS Name Microsoft Windows 11 Enterprise Version 10.0.22631 Build 22631 Other OS Description Not Available OS Manufacturer Microsoft Corporation System Name ASUSTUF System Manufacturer ASUSTeK COMPUTER INC. System Model ASUS TUF Gaming F15 FX507ZM_FX507ZM System Type x64-based PC System SKU
Processor 12th Gen Intel(R) Core(TM) i7-12700H, 2300 Mhz, 14 Core(s), 20 Logical Processor(s) BIOS Version/Date American Megatrends International, LLC. FX507ZM.318, 14/12/2023 SMBIOS Version 3.5 Embedded Controller Version 3.09 BIOS Mode UEFI BaseBoard Manufacturer ASUSTeK COMPUTER INC. BaseBoard Product FX507ZM Installed Physical Memory (RAM) 32,0 GB Total Physical Memory 31,6 GB Available Physical Memory 5,93 GB Total Virtual Memory 41,2 GB Available Virtual Memory 2,99 GB Page File Space 9,56 GB

I am not assuming anything. maybe its my system but since I cant pinpoint it I am reporting. I am your best friend

[cosmic-snow]

ollama3.1 8B instruct 128k

You mean the one from the download list, right? Because 'ollama' with an 'o' is something else.

Can you go to Settings and set Device to CPU, then try again?

Also, looks like you're using a lot of your memory already, you might want to close some other things down first, because these things can be very memory hungry themselves. Or try a smaller one which uses a bit less, like 'Phi-3 Mini Instruct' or 'Qwen2-1.5B-Instruct'.

Finally, can you give some info on your graphics card? (Important if you run with Device set to something other than CPU.)

[josersleal]

[josersleal]

this morning after loosing all chats again, I started two new ones

then I restarted the app and could see the spinning "saving chats" thingy. I thought "cool! this works!"

2 minutes laterrr:

PS: the memory info I showed you is with the gpt4all already running with a model

[josersleal]

I can't find your unit or other tests... coudl you point me in the right direction please?

[cosmic-snow]

Thanks for the info.

Have you tried that with device set to CPU and did it still crash? Or not?

The chats are saved alongside the models in what is specified as 'Download Path' in your settings. Does that folder have special permissions on your machine, that is, can the application not write there for some reason? Maybe try a different folder.

Also, have a look at the log file, which should typically be in: C:\Users\<name>\AppData\Local\nomic.ai\GPT4All\. What does it say regarding chats?

[SINAPSA-IC]

Wasn't there an Issue with Intel Iris graphics cards? /nototalrecall

[kbvtjkjay]

You have to fill out more information otherwise this issue is useless. Please try harder or we're not going to be able to help.

Hi, Facing the same issue.

OS Name Microsoft Windows 11 Enterprise Version 10.0.22621 Build 22621 Other OS Description Not Available OS Manufacturer Microsoft Corporation System Name DESKTOP-5KNK6K7 System Manufacturer LENOVO System Model 20EQS0VV0U System Type x64-based PC System SKU LENOVO_MT_20EQ_BU_Think_FM_ThinkPad P50 Processor Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz, 2712 Mhz, 4 Core(s), 8 Logical Processor(s) BIOS Version/Date LENOVO N1EETA2W (1.75 ), 18-03-2024 SMBIOS Version 2.8 Embedded Controller Version 1.18 BIOS Mode UEFI BaseBoard Manufacturer LENOVO BaseBoard Product 20EQS0VV0U BaseBoard Version SDK0J40697 WIN Platform Role Mobile Secure Boot State On PCR7 Configuration Elevation Required to View Windows Directory C:\Windows System Directory C:\Windows\system32 Boot Device \Device\HarddiskVolume1 Locale United States Hardware Abstraction Layer Version = "10.0.22621.2506" User Name DESKTOP-5KNK6K7\jd Time Zone India Standard Time Installed Physical Memory (RAM) 32.0 GB Total Physical Memory 31.9 GB Available Physical Memory 22.5 GB Total Virtual Memory 45.9 GB Available Virtual Memory 34.7 GB Page File Space 14.0 GB Page File E:\pagefile.sys Kernel DMA Protection Off Virtualization-based security Not enabled Windows Defender Application Control policy Enforced Windows Defender Application Control user mode policy Off Device Encryption Support Elevation Required to View Hyper-V - VM Monitor Mode Extensions Yes Hyper-V - Second Level Address Translation Extensions Yes Hyper-V - Virtualization Enabled in Firmware Yes Hyper-V - Data Execution Protection Yes

GPT4All v3.2.0 Currently qwen2-1_5b-instruct-q4_0.gguf working properly. But Nous-Hermes-2-Mistral-7B-DPO.Q4_0.gguf Meta-Llama-3-8B-Instruct.Q4_0.gguf Meta-Llama-3.1-8B-Instruct-128k-Q4_0.gguf crashes chat.exe

Application Error ID:1000 Application Crashing Event Faulting application name: chat.exe, version: 0.0.0.0, time stamp: 0x66ba2bdb Faulting module name: ucrtbase.dll, version: 10.0.22621.3593, time stamp: 0x10c46e71 Exception code: 0xc0000409 Fault offset: 0x000000000007e008 Faulting process id: 0x0x293C Faulting application start time: 0x0x1DAED8E7F1CFB41 Faulting application path: C:\Users\jd\gpt4all\bin\chat.exe Faulting module path: C:\Windows\System32\ucrtbase.dll Report Id: b798edb0-c074-40c2-8d11-0690f06cd46f Faulting package full name: Faulting package-relative application ID:

Tried sfc to repair system files. didn't work. Reinstalled gpt4all with the latest installer. didn't work. Reinstalled gpt4all with the latest as well older installer at different empty folder. didn't work.

[kbvtjkjay]

You have to fill out more information otherwise this issue is useless. Please try harder or we're not going to be able to help.

Hi, Facing the same issue.

OS Name Microsoft Windows 11 Enterprise Version 10.0.22621 Build 22621 Other OS Description Not Available OS Manufacturer Microsoft Corporation System Name DESKTOP-5KNK6K7 System Manufacturer LENOVO System Model 20EQS0VV0U System Type x64-based PC System SKU LENOVO_MT_20EQ_BU_Think_FM_ThinkPad P50 Processor Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz, 2712 Mhz, 4 Core(s), 8 Logical Processor(s) BIOS Version/Date LENOVO N1EETA2W (1.75 ), 18-03-2024 SMBIOS Version 2.8 Embedded Controller Version 1.18 BIOS Mode UEFI BaseBoard Manufacturer LENOVO BaseBoard Product 20EQS0VV0U BaseBoard Version SDK0J40697 WIN Platform Role Mobile Secure Boot State On PCR7 Configuration Elevation Required to View Windows Directory C:\Windows System Directory C:\Windows\system32 Boot Device \Device\HarddiskVolume1 Locale United States Hardware Abstraction Layer Version = "10.0.22621.2506" User Name DESKTOP-5KNK6K7\jd Time Zone India Standard Time Installed Physical Memory (RAM) 32.0 GB Total Physical Memory 31.9 GB Available Physical Memory 22.5 GB Total Virtual Memory 45.9 GB Available Virtual Memory 34.7 GB Page File Space 14.0 GB Page File E:\pagefile.sys Kernel DMA Protection Off Virtualization-based security Not enabled Windows Defender Application Control policy Enforced Windows Defender Application Control user mode policy Off Device Encryption Support Elevation Required to View Hyper-V - VM Monitor Mode Extensions Yes Hyper-V - Second Level Address Translation Extensions Yes Hyper-V - Virtualization Enabled in Firmware Yes Hyper-V - Data Execution Protection Yes

GPT4All v3.2.0 Currently qwen2-1_5b-instruct-q4_0.gguf working properly. But Nous-Hermes-2-Mistral-7B-DPO.Q4_0.gguf Meta-Llama-3-8B-Instruct.Q4_0.gguf Meta-Llama-3.1-8B-Instruct-128k-Q4_0.gguf crashes chat.exe

Application Error ID:1000 Application Crashing Event Faulting application name: chat.exe, version: 0.0.0.0, time stamp: 0x66ba2bdb Faulting module name: ucrtbase.dll, version: 10.0.22621.3593, time stamp: 0x10c46e71 Exception code: 0xc0000409 Fault offset: 0x000000000007e008 Faulting process id: 0x0x293C Faulting application start time: 0x0x1DAED8E7F1CFB41 Faulting application path: C:\Users\jd\gpt4all\bin\chat.exe Faulting module path: C:\Windows\System32\ucrtbase.dll Report Id: b798edb0-c074-40c2-8d11-0690f06cd46f Faulting package full name: Faulting package-relative application ID:

Tried sfc to repair system files. didn't work. Reinstalled gpt4all with the latest installer. didn't work. Reinstalled gpt4all with the latest as well older installer at different empty folder. didn't work.

GPU Details if required.

Name NVIDIA Quadro M1000M PNP Device ID PCI\VEN_10DE&DEV_13B1&SUBSYS_223017AA&REV_A2\4&26BA905&0&0008 Adapter Type Quadro M1000M, NVIDIA compatible Adapter Description NVIDIA Quadro M1000M Adapter RAM (2,147,483,648) bytes Installed Drivers C:\Windows\System32\DriverStore\FileRepository\nvlt.inf_amd64_b72b32629cc698ac\nvldumdx.dll,C:\Windows\System32\DriverStore\FileRepository\nvlt.inf_amd64_b72b32629cc698ac\nvldumdx.dll,C:\Windows\System32\DriverStore\FileRepository\nvlt.inf_amd64_b72b32629cc698ac\nvldumdx.dll,C:\Windows\System32\DriverStore\FileRepository\nvlt.inf_amd64_b72b32629cc698ac\nvldumdx.dll Driver Version 31.0.15.3818 INF File oem28.inf (Section060 section) Color Planes Not Available Color Table Entries 4294967296 Resolution 1920 x 1080 x 59 hertz Bits/Pixel 32 Memory Address 0xB2000000-0xB2FFFFFF Memory Address 0xA0000000-0xAFFFFFFF Memory Address 0xB0000000-0xB1FFFFFF IRQ Channel IRQ 4294967283 Driver C:\WINDOWS\SYSTEM32\DRIVERSTORE\FILEREPOSITORY\NVLT.INF_AMD64_B72B32629CC698AC\NVLDDMKM.SYS (31.0.15.3818, 56.48 MB (59,221,016 bytes), 16-05-2024 15:46)

[cosmic-snow]

Thanks for all the info, but I have to ask: Which device setting were you guys using? Was it always with GPU? (Vulkan or CUDA, if so which?)

Can you try:

• Set Device to 'CPU'
• Restart to make sure it's just that
• Try a model. Does it crash? Does it save a chat if you close it now after a conversation?

[kbvtjkjay]

have given all the details at

#2867

Thanks for all the info, but I have to ask: Which device setting were you guys using? Was it always with GPU? (Vulkan or CUDA, if so which?)

Can you try:

• Set Device to 'CPU'
• Restart to make sure it's just that
• Try a model. Does it crash? Does it save a chat if you close it now after a conversation?

[cosmic-snow]

@kbvtjkjay in the other issue you wrote:

So the test results.

1. Explicit CPU: works fine with all models. I copied my old chats and models to ~\AppData\Local\nomic.ai\ ALL GOOD.
2. Explicit CUDA: Nous-Hermes-2 let's me give prompt but crashes. Other models: just crashes at loading.
3. Explicit VULCAN: Crashes: Any model

You have an Nvidia Quadro M1000M with 2GB VRAM. First of all, it obviously shouldn't crash, no doubt. But you're not going to do much with that card, either, if you're trying to set it to one of the GPU options.

What you can try is: go to Settings > Model > GPU Layers and set it to 1. Test that and if it works, try to slowly increase it while monitoring your VRAM use. You probably won't get much out of it, though. I'm wondering, did it not just always use the CPU fallback in your specific case? 🤔

[cosmic-snow]

Looks like this kind of crash has happened in even earlier versions. #2674 might be the same thing.

Another report is here on Discord.