corvus @ corvus @lemmy.ml

Posts

12

Comments

156

Joined

2 yr. ago

I've been using Debian with KDE Plasma for over a decade and I can count the crashes with the fingers of one hand.

Here you can find hardware for linux that requires no proprietary driver or firmware, in your case is ASUS BT400. I was in the same situation as yours so I bought it and it works.

I don't know, but I really enjoyed reading his books.

It gives me exactly the same message but I'm not using a VPN. When I use the external viewer option with mpv using yt-dlp I only get video without audio. I can download the video fine using yt-dlp and then watch it with mpv, but if I try to stream to mpv while downloading to watch it real-time it gives an ffmpeg error: can't recognize format... weird.

Then it's not the new C, maybe the new C++

Then I'll wait for Rust++

Imagine having that great idea and carefully crafting the edition of the book for that to happen. Or may be just unscrupulously fill up the book with white pages with the sentence "this page is intentionally left blank (see page 269)"

I ended up buying an ASUS BT400, it works out of the box in Linux. I found it here

Oh great, thanks

Yeah I tested with lower numbers and it works, I just wanted to offload the whole model thinking it will work, 2GB it's a lot. With other models it prints about 250MB when fails and if you sum up the model size it's still well below the iGPU free memory so I dont get it... anyway, I was thinking about upgrading the memory to 32GB or may be 64GB but I hesitate because with models around 7GB and CPU only I get around 5 t/s and with 14GB 2-3 t/s, so I run one of around 30GB I guess it will get around 1 t/s? My supposition is that increasing RAM doesn't increase performance per se, just let's you upload bigger models to memory, so performance is approximately linear on model size... what do you think?

I get an error when offloading the whole model to GPU

./build/bin/llama-cli -m ~/software/ai/models/deepseek-math-7b-instruct.Q8_0.gguf -n 200 -t 10 -ngl 31 -if

The relevant output is:

....

llama_model_load_from_file_impl: using device Vulkan0 (Intel(R) Iris(R) Xe Graphics (RPL-U)) - 7759 MiB free

...

print_info: file size = 6.84 GiB (8.50 BPW)

....

load_tensors: loading model tensors, this can take a while... (mmap = true) load_tensors: offloading 30 repeating layers to GPU load_tensors: offloading output layer to GPU load_tensors: offloaded 31/31 layers to GPU load_tensors: Vulkan0 model buffer size = 6577.83 MiB load_tensors: CPU_Mapped model buffer size = 425.00 MiB

.....

ggml_vulkan: Device memory allocation of size 2013265920 failed ggml_vulkan: vk::Device::allocateMemory: ErrorOutOfDeviceMemory llama_kv_cache_init: failed to allocate buffer for kv cache llama_init_from_model: llama_kv_cache_init() failed for self-attention cache common_init_from_params: failed to create context with model '~/software/ai/models/deepseek-math-7b-instruct.Q8_0.gguf' main: error: unable to load model

It seems to me that there is enough room for the model, but I don't know what "Device memory allocation of size 2013265920" means.

Is BLAS faster with CPU only than Vulkan with CPU+iGPU? After failing to make work the SYCL backend in llama.cpp apparently because of a Debian driver issue I ended up using the Vulkan backend but after many tests offloadding to the iGPU doesn't seem to make much difference.

Is BLAS faster with CPU only than Vulkan with CPU+iGPU? After failing to make work the SYCL backend of llama.cpp apparently because a Debian driver issue I tried the Vulkan backend successfuly but offloading to iGPU doesn't seems to make much difference.

I don't like intermediaries ;) Fortunately I compiled llama.cpp with the Vulkan backend and everything went smooth and now I have the option to offload to the GPU. Now I will test performance CPU vs CPU+GPU. Downloaded deepseek 14b and is really good, the best I could run so far in my limited hardware.

Yes, gpt4all runs it in cpu mode, the gpu option does not appear in the drop-down menu, which means the gpu it's not supported or there is an error. I'm trying to run the models with the SyCL backend implemented in llama.cpp that performs specific optimizations for cpu+gpu with the Intel DPC++/C++ Compiler and the OneAPI Toolkit.

Also try Deepseek 14b. It will be much faster.

ok, I'll test it out.

I tried llama.cpp but I was having some errors about not finding some library so I tried gpt4all and it worked. I'll try to recompilte and test it again. I have a thinkbook with Intel i5-1335u and integrated Xe graphics. I installed the Intel OneAPI toolkit so llama.cpp could take advantage of the SYCL backend for Intel GPUs, but I had an execution error that I was unable to solve after many days. I installed the Vulkan SDK needed to compile gpt4all with the hope to being able to use the GPU but gpt4all-chat doesn't show the option to run from it, so from what I read it means that it's not supported, but from some posts that I read I should not expect a big performance boost from that GPU.

What do you mean by "average number of pages"? Average over what?

Searching "github download webpage video" gives this and more results to try.