Closed pallebone closed 3 weeks ago
sudo RUST_BACKTRACE=1 ./nvidia_oc set --index 0 --freq-offset 272 --mem-offset 1000
thread 'main' panicked at src/main.rs:73:26:
Failed to set GPU frequency offset: "Error code: 6"
stack backtrace:
0: rust_begin_unwind
1: core::panicking::panic_fmt
2: core::result::unwrap_failed
3: nvidia_oc::main
note: Some details are omitted, run with RUST_BACKTRACE=full
for a verbose backtrace.
sudo RUST_BACKTRACE=full ./nvidia_oc set --index 0 --freq-offset 272 --mem-offset 1000
thread 'main' panicked at src/main.rs:73:26:
Failed to set GPU frequency offset: "Error code: 6"
stack backtrace:
0: 0x55fd14612d05 -
Hi @pallebone,
According to the NVML documentation, the error code you encountered is described as follows:
NVML_ERROR_NOT_FOUND = 6 A query to find an object was unsuccessful.
This indicates that the issue originates from NVIDIA's side, not from my application. Therefore, my ability to resolve it is limited. However, if you can provide the exact model of your GPU, I can conduct further research to identify any potential workarounds.
Geforce GTX 1050 Strange, under xorg and greenwithenvy it overclocks fine. Thought I would try wayland and your script.
Geforce GTX 1050 Strange, under xorg and greenwithenvy it overclocks fine. Thought I would try wayland and your script.
I'm not certain, but I will check the NVML API for alternative methods. There seem to be some legacy methods that might work with your GPU.
Please note that GreenWithEnvy uses a completely different method, which is only possible under Xorg due to the special integration with the NVIDIA driver.
@pallebone are you using the open source nvidia drivers?
By the way, the error you're encountering is quite unusual. The NVIDIA NVML documentation does not list that error as a possible result.
@pallebone are you using the open source nvidia drivers?
No I am using nvidas driver 535.183.01 installed by apt via apt-get install nvidia-driver
@pallebone are you using the open source nvidia drivers?
No I am using nvidas driver 535.183.01 installed by apt via apt-get install nvidia-driver
That's quite an old version. On my system, I have 555.58.02. That could be the issue. Unfortunately, I can't test that version since I'm using Arch, and I don't want to deal with the dependency hell to install such an outdated version.
Ok I will look to update then and let you know
I checked and the current rivers on Nvidias website are:
Linux x86_64/AMD64/EM64T Latest Production Branch Version: 550.100 Latest New Feature Branch Version: 555.58.02 Latest Beta Version: 560.28.03
The version you are using is the new feature branch. I may have to wait a little while for these drivers to be included into Debian Trixie. The next driver coming down in 2 weeks is 545.23.06-1.
I imagine after this then the next one will be a 555 or > version. However that is a month away. The current driver I am using is only 1 month old. However I will wait for the drivers to come down and retry.
This guide used 545.29.02 so in 1 month I should have the same or newer version to test with. https://www.reddit.com/r/linux_gaming/comments/17kx3vu/how_to_crudely_overclock_your_nvidia_gpu_on/
Just FYI I tried with the next version to come down and have a similar issue, so I will wait for one more later version and see after this:
aragorn@Aragorn:~$ sudo RUST_BACKTRACE=full ./nvidia_oc set --index 0 --freq-offset 272 --mem-offset 1000
[sudo] password for aragorn:
thread 'main' panicked at src/main.rs:73:26:
Failed to set GPU frequency offset: "Error code: 6"
stack backtrace:
0: 0x558c003e3d05 - <std::sys_common::backtrace::_print::DisplayBacktrace as core::fmt::Display>::fmt::h1e1a1972118942ad
1: 0x558c0040534b - core::fmt::write::hc090a2ffd6b28c4a
2: 0x558c003e1cff - std::io::Write::write_fmt::h8898bac6ff039a23
3: 0x558c003e3ade - std::sys_common::backtrace::print::ha96650907276675e
4: 0x558c003e4d99 - std::panicking::default_hook::{{closure}}::h215c2a0a8346e0e0
5: 0x558c003e4add - std::panicking::default_hook::h207342be97478370
6: 0x558c003e5233 - std::panicking::rust_panic_with_hook::hac8bdceee1e4fe2c
7: 0x558c003e5114 - std::panicking::begin_panic_handler::{{closure}}::h00d785e82757ce3c
8: 0x558c003e41c9 - std::sys_common::backtrace::__rust_end_short_backtrace::h1628d957bcd06996
9: 0x558c003e4e47 - rust_begin_unwind
10: 0x558c00317df3 - core::panicking::panic_fmt::hdc63834ffaaefae5
11: 0x558c00318206 - core::result::unwrap_failed::h82b551e0ff2b2176
12: 0x558c0032c80b - nvidia_oc::main::h5d23fbf1455d3ebc
13: 0x558c0033bcd3 - std::sys_common::backtrace::__rust_begin_short_backtrace::h62cf2716c9e91ff3
14: 0x558c0033c949 - std::rt::lang_start::{{closure}}::h84d874c721ca05fb
15: 0x558c003dd110 - std::rt::lang_start_internal::h3ed4fe7b2f419135
16: 0x558c0032f045 - main
17: 0x7fc213fccc8a - __libc_start_call_main
at ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
18: 0x7fc213fccd45 - __libc_start_main_impl
at ./csu/../csu/libc-start.c:360:3
19: 0x558c003183e5 - _start
20: 0x0 - <unknown>
aragorn@Aragorn:~$ nvidia-smi
Wed Jul 31 16:04:52 2024
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 545.23.06 Driver Version: 545.23.06 CUDA Version: 12.3 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce GTX 1050 On | 00000000:01:00.0 Off | N/A |
| N/A 47C P8 N/A / ERR! | 5MiB / 4096MiB | 0% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 4115 G /usr/bin/gnome-shell 1MiB |
+---------------------------------------------------------------------------------------+
aragorn@Aragorn:~$
Upgraded to the 560 driver and it resolved the issue.
I can alter memory offset but if I try to alter freq offset I get an error:
sudo ./nvidia_oc set --index 0 --freq-offset 272 --mem-offset 1000 thread 'main' panicked at src/main.rs:73:26: Failed to set GPU frequency offset: "Error code: 6" note: run with
RUST_BACKTRACE=1
environment variable to display a backtraceTried contacting you on discord. Not a friend.