5.36.406.459 I srv get_availabl: prompt cache update took 134.10 ms
5.36.427.780 I slot launch_slot_: id 2 | task 921 | processing task, is_child = 0
5.42.339.503 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 4096, progress = 0.15, t = 5.91 s / 692.86 tokens per second
5.54.506.021 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 8192, progress = 0.31, t = 18.08 s / 453.14 tokens per second
6.07.322.975 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 12288, progress = 0.46, t = 30.90 s / 397.73 tokens per second
6.20.693.688 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 16384, progress = 0.62, t = 44.27 s / 370.13 tokens per second
6.34.698.298 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 20480, progress = 0.77, t = 58.27 s / 351.46 tokens per second
6.49.277.350 I slot print_timing: id 1 | task 920 | prompt processing, n_tokens = 24576, progress = 0.93, t = 72.85 s / 337.35 tokens per second
6.56.773.219 I begin: ngram_mod occupancy = 83346/4194304 (0.02)
7.04.163.020 I reasoning-budget: activated, budget=2147483647 tokens
7.04.167.331 I begin: ngram_mod occupancy = 83429/4194304 (0.02)
/tmp/pi-github-repos/Anbeeld/beellama.cpp/common/speculative.cpp:4780: GGML_ASSERT(impl) failed
[New LWP 1973552]
[New LWP 1973551]
[New LWP 1973550]
[New LWP 1973549]
[New LWP 1973548]
[New LWP 1973547]
[New LWP 1973546]
[New LWP 1973545]
[New LWP 1973544]
[New LWP 1973543]
[New LWP 1973542]
[New LWP 1973541]
[New LWP 1973540]
[New LWP 1973539]
[New LWP 1973538]
[New LWP 1973537]
[New LWP 1973536]
[New LWP 1973535]
[New LWP 1973534]
[New LWP 1973533]
[New LWP 1973532]
[New LWP 1973531]
[New LWP 1973530]
This GDB supports auto-downloading debuginfo from the following URLs:
<https://debuginfod.ubuntu.com>
Enable debuginfod for this session? (y or [n]) [answered N; input not from terminal]
Debuginfod has been disabled.
To make this setting permanent, add 'set debuginfod enabled off' to .gdbinit.
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
0x0000ee3e681e7b74 in __GI___wait4 (pid=1973930, stat_loc=0x0, options=0, usage=0x0) at ../sysdeps/unix/sysv/linux/wait4.c:30
warning: 30 ../sysdeps/unix/sysv/linux/wait4.c: No such file or directory
#0 0x0000ee3e681e7b74 in __GI___wait4 (pid=1973930, stat_loc=0x0, options=0, usage=0x0) at ../sysdeps/unix/sysv/linux/wait4.c:30
30 in ../sysdeps/unix/sysv/linux/wait4.c
#1 0x0000ee3e677494bc in ggml_print_backtrace () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libggml-base.so.0
#2 0x0000ee3e67749660 in ggml_abort () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libggml-base.so.0
#3 0x0000ee3e67f50d28 in common_speculative_accept(common_speculative*, unsigned short) () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libllama-common.so.0
#4 0x0000ee3e686ebb24 in server_context_impl::update_slots() () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libllama-server-impl.so
#5 0x0000ee3e68781bc4 in server_queue::start_loop(long) () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libllama-server-impl.so
#6 0x0000ee3e68639e34 in llama_server(int, char**) () from /tmp/pi-github-repos/Anbeeld/beellama.cpp/build/bin/libllama-server-impl.so
#7 0x0000ee3e681384c4 in __libc_start_call_main (main=main@entry=0xb810339e1180 <main>, argc=argc@entry=58, argv=argv@entry=0xffffce0e4528) at ../sysdeps/nptl/libc_start_call_main.h:58
warning: 58 ../sysdeps/nptl/libc_start_call_main.h: No such file or directory
#8 0x0000ee3e68138598 in __libc_start_main_impl (main=0xb810339e1180 <main>, argc=58, argv=0xffffce0e4528, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=<optimized out>) at ../csu/libc-start.c:360
warning: 360 ../csu/libc-start.c: No such file or directory
#9 0x0000b810339e11f0 in _start ()
[Inferior 1 (process 1973526) detached]
Aborted (core dumped)
Name and Version
./llama-cli --version
version: 1 (85e22ea)
built with GNU 13.3.0 for Linux aarch64
Operating systems
Linux
GGML backends
CUDA
Hardware
DGX Spark
Models
Mistral Small 4 NVFP4
Problem description & steps to reproduce
Well—Agent Pi was working on the prompt when the server suddenly crashed. I don't know if we can draw any conclusions from this log, but I'm reporting it because BeeLlama has potential :P
First Bad Commit
No response
Relevant log output
Logs