Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
ohai.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A cozy, fast and secure Mastodon server where everyone is welcome. Run by the folks at ohai.is.

Administered by:

Server stats:

1.8K
active users

ohai.social: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.4.0-nightly.2025-04-01

#nvidia

106 posts93 participants8 posts today
Public *
EmpathicQubit

I setup #systemd-boot on my computer, but there were a couple of annoyances:

  • The #kernel isn't signed for whatever reason. I wired in sbctl sign to the install script, but that wasn't completely straightforward.

  • Since #Windows is installed in a separate #EFI (because Windows likes to fuck up the entire EFI partition sometimes) it couldn't "see" the Windows Boot Manager. I copied it to the other partition, but it will have to be manually updated whenever Microsoft changes it. Maybe that doesn't happen that often idk.

Side problem is that #Ubuntu's nvidia-lowlatency kernel isn't set up to reject unsigned modules, so it's a bit of a security hole, but also means that I don't have to figure out getting #DKMS to use the correct key. Right now I don't have a dependency on a DKMS-built driver. I used to use one for my dock and NVidia, but it seems that #NVidia doesn't need it now? and I don't use the video in my dock anymore because the #DisplayLink driver is annoying regardless of secure boot issues, because it doesn't work from power on. This isn't a problem with Linux, just DisplayLink in general.

Replied in thread
Public
NV Access

@pafurijaz Hi! Is there any chance we can encourage you to trend #NVIDIA when that's the company you mean, please? I know what their NASDAQ handle is, but #NVDA is much more widely known as the name of the screen reader we make and the #NVDA hashtag is very widely used for the screen reader. It will save both our communities polluting each other's feeds. If you'd like to find out more about the screen reader, our website is nvaccess.org/ - Thank you!

NV AccessNV AccessEmpowering lives through non-visual access to technology
Replied in thread
Public
NV Access

@Dhmspector Hi! Is there any chance we can encourage you to trend #NVIDIA when that's the company you mean, please? I know what their NASDAQ handle is, but #NVDA is much more widely known as the name of the screen reader we make and the #NVDA hashtag is very widely used for the screen reader. It will save both our communities polluting each other's feeds. If you'd like to find out more about the screen reader, our website is nvaccess.org/ - Thank you!

NV AccessNV AccessEmpowering lives through non-visual access to technology
Public *
KM6ECC

Good or bad? My laptop with a 12th gen i7-12800H and an #nvidia A1000 gpu can sustain 3.8ghz turbo on all cores, almost a 1 ghz turbo (sustained) over the 2.9ghz promise from #intel. I am using Ollama and Gemma3:27b to beat on it. GPU is a bit of a lap potato of course and hovers around 65C, while the CPU rides 96C.

Tokens per second is about 25.

#ollama#AI#LLM
Public
Benjamin Carr, Ph.D. 👨🏻‍💻🧬

#UALink debuts its first #AI interconnect spec – usable in just 18 short months
Consortium (UAC) was formed in May 2024 by vendors – AMD, AWS, Broadcom, Cisco, Google, HPE, Intel, Meta, Microsoft, and Astera Labs – who think world needs an open alternative to #Nvidia's #NVLink tech that allows creation of networked GPU clusters needed to run AI workloads at scale.
The Ultra Accelerator Link Consortium has delivered first GPU interconnect specification: UALink 200G 1.0.
theregister.com/2025/04/08/ual

The Register · UALink debuts its first AI interconnect spec – usable in just 18 short monthsBy Simon Sharwood
Public
♡ Eva Winterschön ♡

🌇 Blog Post Forthcoming for GPUs 🌇

ok, it's morning nap time starting 5min ago... but I'm decently pleased that this morning I've wrangled the (yes, linux) kernel into behaving more decently (a little like FreeBSD's sanity) when allocating similarly ID'd PCIe devices to VMs for GPU passthrough.

Short version:
- many GPUs of same PCI ID type in host
- want certain VMs to get certain GPUs
- want main host to get one GPU for compute (server, no display)

FreeBSD:
- the above requires two loader.conf lines and one rc.conf line

Linux:
- oh lordy, how many modprobe and modules-load.d files, and grub entries, and kmod blacklists, rebuild initramfs, and rebuild grub menu, and don't forget to purge the entire nouveau driver from the kernel so it doens't try to backstab nvidia and steal your first born...

#nvidia#gpu#linux
ExploreLive feeds
About