Cannot start llama.cpp on Windows, no output

[waltersamson]

My setup:

• Multiple AVX2 Intel and AMD Windows boxes
• llama-b4372-bin-win-avx2-x64
• Windows 10

I feel stupid now. I've been using llama.cpp on Linux for quite some time without issue. Today I tried running it on Windows, somehow I ran into a strange problem:

• Download llama-b4372-bin-win-avx2-x64.zip from Releases
• Execute llama-cli.exe or llama-server.exe

Expected:
The executable should print something, such as "build: 4337 (160bc03) with cc (Ubuntu 13.2.0-23ubuntu4) 13.2.0 for x86_64-linux-gnu"

However, in my case, the program just returns, with no output whatsoever, it looks like this:

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>llama-server.exe

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>llama-cli.exe

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>llama-gguf.exe
usage: llama-gguf.exe data.gguf r|w [n]
r: read data.gguf file
w: write data.gguf file
n: no check of tensor data

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>

The binary llama-gguf.exe did run, but not llama-server.exe and llama-cli.exe.

Then I tried giving it a complete cmdline, but still nothing happened:

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>llama-cli.exe -m ..\..\..\models\volatile\Qwen2.5-14B-Instruct-Q8_0.gguf -co -cnv -p "You are a helpful assistant." -fa -n 512 -c 8192

F:\llm\apps\llama.cpp\llama-b4372-bin-win-avx2-x64>

I feel stupid now, I must be missing an obvious piece.

Thanks for any help!

[Martin-HZK]

Same question, have you figure out why? Everything works fine on windows but not windows