Bug: llama-server with --system-prompt-file stops abruptly without any error

[pritam-dey3]

What happened?

following @ggerganov 's advice here I tried to run server with the following command

docker run -v /home/pritam/llm/models:/models -v /home/pritam/lipika/:/lipika -p 8080:8080 ghcr.io/ggerganov/llama.cpp:server -m /models/gemma-2-2b-it-Q4_K_M.gguf --port 8080 --host 0.0.0.0 --system-prompt-file /lipika/system.txt --logdir /lipika/logs --log-enable --log-new

The system.txt file contains around 2.3k tokens. Shortly after starting the server, the process stops abruptly without showing any error. I mentioned logdir and enabled logs but no logs were generated.

This is the the full output before the server stops

docker run -v /home/pritam/llm/models:/models -v /home/pritam/lipika/:/lipika -p 8080:8080 ghcr.io/ggerganov/llama.cpp:server -m /models/gemma-2-2b-it-Q4_K_M.gguf --port 8080 --host 0.0.0.0 --system-prompt-file /lipika/system.txt --logdir /lipika/logs --log-enable --log-new
INFO [                    main] build info | tid="140734278524576" timestamp=1723315495 build=0 commit="unknown"
INFO [                    main] system info | tid="140734278524576" timestamp=1723315495 n_threads=4 n_threads_batch=-1 total_threads=4 system_info="AVX = 0 | AVX_VNNI = 0 | AVX2 = 0 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 0 | NEON = 1 | SVE = 0 | ARM_FMA = 1 | F16C = 0 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 0 | SSE3 = 0 | SSSE3 = 0 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | "
llama_model_loader: loaded meta data with 39 key-value pairs and 288 tensors from /models/gemma-2-2b-it-Q4_K_M.gguf (version GGUF V3 (latest))
llama_model_loader: Dumping metadata keys/values. Note: KV overrides do not apply in this output.
llama_model_loader: - kv   0:                       general.architecture str              = gemma2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Gemma 2 2b It
llama_model_loader: - kv   3:                           general.finetune str              = it
llama_model_loader: - kv   4:                           general.basename str              = gemma-2
llama_model_loader: - kv   5:                         general.size_label str              = 2B
llama_model_loader: - kv   6:                            general.license str              = gemma
llama_model_loader: - kv   7:                               general.tags arr[str,2]       = ["conversational", "text-generation"]
llama_model_loader: - kv   8:                      gemma2.context_length u32              = 8192
llama_model_loader: - kv   9:                    gemma2.embedding_length u32              = 2304
llama_model_loader: - kv  10:                         gemma2.block_count u32              = 26
llama_model_loader: - kv  11:                 gemma2.feed_forward_length u32              = 9216
llama_model_loader: - kv  12:                gemma2.attention.head_count u32              = 8
llama_model_loader: - kv  13:             gemma2.attention.head_count_kv u32              = 4
llama_model_loader: - kv  14:    gemma2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  15:                gemma2.attention.key_length u32              = 256
llama_model_loader: - kv  16:              gemma2.attention.value_length u32              = 256
llama_model_loader: - kv  17:                          general.file_type u32              = 15
llama_model_loader: - kv  18:              gemma2.attn_logit_softcapping f32              = 50.000000
llama_model_loader: - kv  19:             gemma2.final_logit_softcapping f32              = 30.000000
llama_model_loader: - kv  20:            gemma2.attention.sliding_window u32              = 4096
llama_model_loader: - kv  21:                       tokenizer.ggml.model str              = llama
llama_model_loader: - kv  22:                         tokenizer.ggml.pre str              = default
llama_model_loader: - kv  23:                      tokenizer.ggml.tokens arr[str,256000]  = ["<pad>", "<eos>", "<bos>", "<unk>", ...
llama_model_loader: - kv  24:                      tokenizer.ggml.scores arr[f32,256000]  = [-1000.000000, -1000.000000, -1000.00...
llama_model_loader: - kv  25:                  tokenizer.ggml.token_type arr[i32,256000]  = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, ...
llama_model_loader: - kv  26:                tokenizer.ggml.bos_token_id u32              = 2
llama_model_loader: - kv  27:                tokenizer.ggml.eos_token_id u32              = 1
llama_model_loader: - kv  28:            tokenizer.ggml.unknown_token_id u32              = 3
llama_model_loader: - kv  29:            tokenizer.ggml.padding_token_id u32              = 0
llama_model_loader: - kv  30:               tokenizer.ggml.add_bos_token bool             = true
llama_model_loader: - kv  31:               tokenizer.ggml.add_eos_token bool             = false
llama_model_loader: - kv  32:                    tokenizer.chat_template str              = {{ bos_token }}{% if messages[0]['rol...
llama_model_loader: - kv  33:            tokenizer.ggml.add_space_prefix bool             = false
llama_model_loader: - kv  34:               general.quantization_version u32              = 2
llama_model_loader: - kv  35:                      quantize.imatrix.file str              = /models_out/gemma-2-2b-it-GGUF/gemma-...
llama_model_loader: - kv  36:                   quantize.imatrix.dataset str              = /training_dir/calibration_datav3.txt
llama_model_loader: - kv  37:             quantize.imatrix.entries_count i32              = 182
llama_model_loader: - kv  38:              quantize.imatrix.chunks_count i32              = 128
llama_model_loader: - type  f32:  105 tensors
llama_model_loader: - type q4_K:  156 tensors
llama_model_loader: - type q6_K:   27 tensors
llm_load_vocab: special tokens cache size = 249
llm_load_vocab: token to piece cache size = 1.6014 MB
llm_load_print_meta: format           = GGUF V3 (latest)
llm_load_print_meta: arch             = gemma2
llm_load_print_meta: vocab type       = SPM
llm_load_print_meta: n_vocab          = 256000
llm_load_print_meta: n_merges         = 0
llm_load_print_meta: vocab_only       = 0
llm_load_print_meta: n_ctx_train      = 8192
llm_load_print_meta: n_embd           = 2304
llm_load_print_meta: n_layer          = 26
llm_load_print_meta: n_head           = 8
llm_load_print_meta: n_head_kv        = 4
llm_load_print_meta: n_rot            = 256
llm_load_print_meta: n_swa            = 4096
llm_load_print_meta: n_embd_head_k    = 256
llm_load_print_meta: n_embd_head_v    = 256
llm_load_print_meta: n_gqa            = 2
llm_load_print_meta: n_embd_k_gqa     = 1024
llm_load_print_meta: n_embd_v_gqa     = 1024
llm_load_print_meta: f_norm_eps       = 0.0e+00
llm_load_print_meta: f_norm_rms_eps   = 1.0e-06
llm_load_print_meta: f_clamp_kqv      = 0.0e+00
llm_load_print_meta: f_max_alibi_bias = 0.0e+00
llm_load_print_meta: f_logit_scale    = 0.0e+00
llm_load_print_meta: n_ff             = 9216
llm_load_print_meta: n_expert         = 0
llm_load_print_meta: n_expert_used    = 0
llm_load_print_meta: causal attn      = 1
llm_load_print_meta: pooling type     = 0
llm_load_print_meta: rope type        = 2
llm_load_print_meta: rope scaling     = linear
llm_load_print_meta: freq_base_train  = 10000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn  = 8192
llm_load_print_meta: rope_finetuned   = unknown
llm_load_print_meta: ssm_d_conv       = 0
llm_load_print_meta: ssm_d_inner      = 0
llm_load_print_meta: ssm_d_state      = 0
llm_load_print_meta: ssm_dt_rank      = 0
llm_load_print_meta: model type       = 2B
llm_load_print_meta: model ftype      = Q4_K - Medium
llm_load_print_meta: model params     = 2.61 B
llm_load_print_meta: model size       = 1.59 GiB (5.21 BPW)
llm_load_print_meta: general.name     = Gemma 2 2b It
llm_load_print_meta: BOS token        = 2 '<bos>'
llm_load_print_meta: EOS token        = 1 '<eos>'
llm_load_print_meta: UNK token        = 3 '<unk>'
llm_load_print_meta: PAD token        = 0 '<pad>'
llm_load_print_meta: LF token         = 227 '<0x0A>'
llm_load_print_meta: EOT token        = 107 '<end_of_turn>'
llm_load_print_meta: max token length = 48
llm_load_tensors: ggml ctx size =    0.13 MiB
llm_load_tensors:        CPU buffer size =  1623.67 MiB
..........................................................
llama_new_context_with_model: n_ctx      = 8192
llama_new_context_with_model: n_batch    = 2048
llama_new_context_with_model: n_ubatch   = 512
llama_new_context_with_model: flash_attn = 0
llama_new_context_with_model: freq_base  = 10000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init:        CPU KV buffer size =   832.00 MiB
llama_new_context_with_model: KV self size  =  832.00 MiB, K (f16):  416.00 MiB, V (f16):  416.00 MiB
llama_new_context_with_model:        CPU  output buffer size =     1.95 MiB
llama_new_context_with_model:        CPU compute buffer size =   504.50 MiB
llama_new_context_with_model: graph nodes  = 1050
llama_new_context_with_model: graph splits = 1
INFO [                    init] initializing slots | tid="140734278524576" timestamp=1723315496 n_slots=1
INFO [                    init] new slot | tid="140734278524576" timestamp=1723315496 id_slot=0 n_ctx_slot=8192
INFO [                    main] model loaded | tid="140734278524576" timestamp=1723315496
INFO [                    main] chat template | tid="140734278524576" timestamp=1723315496 chat_example="<start_of_turn>user\nYou are a helpful assistant\n\nHello<end_of_turn>\n<start_of_turn>model\nHi there<end_of_turn>\n<start_of_turn>user\nHow are you?<end_of_turn>\n<start_of_turn>model\n" built_in=true
INFO [                    main] HTTP server listening | tid="140734278524576" timestamp=1723315496 n_threads_http="3" port="8080" hostname="0.0.0.0"
VERB [              start_loop] new task may arrive | tid="140732260343456" timestamp=1723367556
VERB [              start_loop] update_multitasks | tid="140732260343456" timestamp=1723367556
VERB [              start_loop] callback_update_slots | tid="140732260343456" timestamp=1723367556
VERB [    system_prompt_update] system prompt update | tid="140732260343456" timestamp=1723367556 system_prompt="You are..."
VERB [          kv_cache_clear] clearing KV cache | tid="140732260343456" timestamp=1723367556

If I don't mention the system-prompt file, the server runs just fine.

Name and Version

I pulled the most recent version of ghcr.io/ggerganov/llama.cpp:server and tested it in my laptop and raspberry pi 5. In both cases the server stopped after about 30 seconds when system.txt was provided.

windows laptop:

version: 0 (unknown)
built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for x86_64-linux-gnu

raspberry pi 5:

version: 0 (unknown)
built with cc (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0 for aarch64-linux-gnu

What operating system are you seeing the problem on?

Windows, Other? (Please let us know in description)

Relevant log output

No response

[pritam-dey3]

I tried to do some debugging and it seems this line is causing a seg fault. (Segmentation fault (core dumped))
I am too weak in cpp to figure out why

Edit: The default maximum batch_size is 2048 whereas my system.txt token length was 2330, which caused the segmentation fault

[pritam-dey3]

@compilade can you please tell me why we are setting n_tokens = n_batch in this line?

In this case since I provided a system prompt more than n_batch it caused segmentation fault.