Why when Flash Attention enabled ,the model is loaded to CPU and inferring process run on CPU ?

[hzgdeerHo]

When Flash Attention enabled ,the model is load to CPU ？

###1.Flash Attention disabled:

CUDA_VISIBLE_DEVICES=0 ./llama-server -m /home/ubuntu/.cache/huggingface/hub/models--MaziyarPanahi--Meta-Llama-3.1-8B-Instruct-GGUF/snapshots/1f301d86d760b435a11a56de3863bc0121bfb98f/Meta-Llama-3.1-8B-Instruct.Q8_0.gguf --n_gpu_layers -1 -cb -np 4 --port 8866 --host 0.0.0.0
INFO [ main] build info | tid="140011552022528" timestamp=1723011875 build=3530 commit="0bf16de0"
INFO [ main] system info | tid="140011552022528" timestamp=1723011875 n_threads=8 n_threads_batch=-1 total_threads=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | "
llama_model_loader: loaded meta data with 33 key-value pairs and 291 tensors from /home/ubuntu/.cache/huggingface/hub/models--MaziyarPanahi--Meta-Llama-3.1-8B-Instruct-GGUF/snapshots/1f301d86d760b435a11a56de3863bc0121bfb98f/Meta-Llama-3.1-8B-Instruct.Q8_0.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 = llama
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Models Meta Llama Meta Llama 3.1 8B I...
llama_model_loader: - kv 3: general.finetune str = Instruct
llama_model_loader: - kv 4: general.basename str = models-meta-llama-Meta-Llama-3.1
llama_model_loader: - kv 5: general.size_label str = 8B
llama_model_loader: - kv 6: general.license str = llama3.1
llama_model_loader: - kv 7: general.tags arr[str,6] = ["facebook", "meta", "pytorch", "llam...
llama_model_loader: - kv 8: general.languages arr[str,8] = ["en", "de", "fr", "it", "pt", "hi", ...
llama_model_loader: - kv 9: llama.block_count u32 = 32
llama_model_loader: - kv 10: llama.context_length u32 = 131072
llama_model_loader: - kv 11: llama.embedding_length u32 = 4096
llama_model_loader: - kv 12: llama.feed_forward_length u32 = 14336
llama_model_loader: - kv 13: llama.attention.head_count u32 = 32
llama_model_loader: - kv 14: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 15: llama.rope.freq_base f32 = 500000.000000
llama_model_loader: - kv 16: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 17: general.file_type u32 = 7
llama_model_loader: - kv 18: llama.vocab_size u32 = 128256
llama_model_loader: - kv 19: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 20: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 21: tokenizer.ggml.pre str = smaug-bpe
llama_model_loader: - kv 22: tokenizer.ggml.tokens arr[str,128256] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 23: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 24: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv 25: tokenizer.ggml.bos_token_id u32 = 128000
llama_model_loader: - kv 26: tokenizer.ggml.eos_token_id u32 = 128009
llama_model_loader: - kv 27: tokenizer.chat_template str = {% set loop_messages = messages %}{% ...
llama_model_loader: - kv 28: general.quantization_version u32 = 2
llama_model_loader: - kv 29: quantize.imatrix.file str = ./Meta-Llama-3.1-8B-Instruct-GGUF_ima...
llama_model_loader: - kv 30: quantize.imatrix.dataset str = group_40.txt
llama_model_loader: - kv 31: quantize.imatrix.entries_count i32 = 224
llama_model_loader: - kv 32: quantize.imatrix.chunks_count i32 = 68
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q8_0: 226 tensors
llm_load_vocab: special tokens cache size = 256
llm_load_vocab: token to piece cache size = 0.7999 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
llm_load_print_meta: vocab type = BPE
llm_load_print_meta: n_vocab = 128256
llm_load_print_meta: n_merges = 280147
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 131072
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_head = 32
llm_load_print_meta: n_head_kv = 8
llm_load_print_meta: n_rot = 128
llm_load_print_meta: n_swa = 0
llm_load_print_meta: n_embd_head_k = 128
llm_load_print_meta: n_embd_head_v = 128
llm_load_print_meta: n_gqa = 4
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-05
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 = 14336
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 = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 131072
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 = 8B
llm_load_print_meta: model ftype = Q8_0
llm_load_print_meta: model params = 8.03 B
llm_load_print_meta: model size = 7.95 GiB (8.50 BPW)
llm_load_print_meta: general.name = Models Meta Llama Meta Llama 3.1 8B Instruct
llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token = 128009 '<|eot_id|>'
llm_load_print_meta: LF token = 128 'Ä'
llm_load_print_meta: EOT token = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: Tesla T4, compute capability 7.5, VMM: yes
llm_load_tensors: ggml ctx size = 0.14 MiB
llm_load_tensors: offloading 0 repeating layers to GPU
llm_load_tensors: offloaded 0/33 layers to GPU
llm_load_tensors: CPU buffer size = 8137.64 MiB
.........................................................................................
llama_new_context_with_model: n_ctx = 131072
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 = 500000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init: CUDA_Host KV buffer size = 16384.00 MiB
llama_new_context_with_model: KV self size = 16384.00 MiB, K (f16): 8192.00 MiB, V (f16): 8192.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 2.45 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 8993.00 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 264.01 MiB
llama_new_context_with_model: graph nodes = 1030
llama_new_context_with_model: graph splits = 356

###1.Flash Attention enabled:

CUDA_VISIBLE_DEVICES=0 ./llama-server -m /home/ubuntu/.cache/huggingface/hub/models--MaziyarPanahi--Meta-Llama-3.1-8B-Instruct-GGUF/snapshots/1f301d86d760b435a11a56de3863bc0121bfb98f/Meta-Llama-3.1-8B-Instruct.Q8_0.gguf --n_gpu_layers -1 -cb -np 4 --port 8866 --host 0.0.0.0 -fa
INFO [ main] build info | tid="140629271105536" timestamp=1723013333 build=3530 commit="0bf16de0"
INFO [ main] system info | tid="140629271105536" timestamp=1723013333 n_threads=8 n_threads_batch=-1 total_threads=8 system_info="AVX = 1 | AVX_VNNI = 0 | AVX2 = 1 | AVX512 = 0 | AVX512_VBMI = 0 | AVX512_VNNI = 0 | AVX512_BF16 = 0 | FMA = 1 | NEON = 0 | SVE = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | BLAS = 1 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | MATMUL_INT8 = 0 | LLAMAFILE = 1 | "
llama_model_loader: loaded meta data with 33 key-value pairs and 291 tensors from /home/ubuntu/.cache/huggingface/hub/models--MaziyarPanahi--Meta-Llama-3.1-8B-Instruct-GGUF/snapshots/1f301d86d760b435a11a56de3863bc0121bfb98f/Meta-Llama-3.1-8B-Instruct.Q8_0.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 = llama
llama_model_loader: - kv 1: general.type str = model
llama_model_loader: - kv 2: general.name str = Models Meta Llama Meta Llama 3.1 8B I...
llama_model_loader: - kv 3: general.finetune str = Instruct
llama_model_loader: - kv 4: general.basename str = models-meta-llama-Meta-Llama-3.1
llama_model_loader: - kv 5: general.size_label str = 8B
llama_model_loader: - kv 6: general.license str = llama3.1
llama_model_loader: - kv 7: general.tags arr[str,6] = ["facebook", "meta", "pytorch", "llam...
llama_model_loader: - kv 8: general.languages arr[str,8] = ["en", "de", "fr", "it", "pt", "hi", ...
llama_model_loader: - kv 9: llama.block_count u32 = 32
llama_model_loader: - kv 10: llama.context_length u32 = 131072
llama_model_loader: - kv 11: llama.embedding_length u32 = 4096
llama_model_loader: - kv 12: llama.feed_forward_length u32 = 14336
llama_model_loader: - kv 13: llama.attention.head_count u32 = 32
llama_model_loader: - kv 14: llama.attention.head_count_kv u32 = 8
llama_model_loader: - kv 15: llama.rope.freq_base f32 = 500000.000000
llama_model_loader: - kv 16: llama.attention.layer_norm_rms_epsilon f32 = 0.000010
llama_model_loader: - kv 17: general.file_type u32 = 7
llama_model_loader: - kv 18: llama.vocab_size u32 = 128256
llama_model_loader: - kv 19: llama.rope.dimension_count u32 = 128
llama_model_loader: - kv 20: tokenizer.ggml.model str = gpt2
llama_model_loader: - kv 21: tokenizer.ggml.pre str = smaug-bpe
llama_model_loader: - kv 22: tokenizer.ggml.tokens arr[str,128256] = ["!", """, "#", "$", "%", "&", "'", ...
llama_model_loader: - kv 23: tokenizer.ggml.token_type arr[i32,128256] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv 24: tokenizer.ggml.merges arr[str,280147] = ["Ġ Ġ", "Ġ ĠĠĠ", "ĠĠ ĠĠ", "...
llama_model_loader: - kv 25: tokenizer.ggml.bos_token_id u32 = 128000
llama_model_loader: - kv 26: tokenizer.ggml.eos_token_id u32 = 128009
llama_model_loader: - kv 27: tokenizer.chat_template str = {% set loop_messages = messages %}{% ...
llama_model_loader: - kv 28: general.quantization_version u32 = 2
llama_model_loader: - kv 29: quantize.imatrix.file str = ./Meta-Llama-3.1-8B-Instruct-GGUF_ima...
llama_model_loader: - kv 30: quantize.imatrix.dataset str = group_40.txt
llama_model_loader: - kv 31: quantize.imatrix.entries_count i32 = 224
llama_model_loader: - kv 32: quantize.imatrix.chunks_count i32 = 68
llama_model_loader: - type f32: 65 tensors
llama_model_loader: - type q8_0: 226 tensors
llm_load_vocab: special tokens cache size = 256
llm_load_vocab: token to piece cache size = 0.7999 MB
llm_load_print_meta: format = GGUF V3 (latest)
llm_load_print_meta: arch = llama
llm_load_print_meta: vocab type = BPE
llm_load_print_meta: n_vocab = 128256
llm_load_print_meta: n_merges = 280147
llm_load_print_meta: vocab_only = 0
llm_load_print_meta: n_ctx_train = 131072
llm_load_print_meta: n_embd = 4096
llm_load_print_meta: n_layer = 32
llm_load_print_meta: n_head = 32
llm_load_print_meta: n_head_kv = 8
llm_load_print_meta: n_rot = 128
llm_load_print_meta: n_swa = 0
llm_load_print_meta: n_embd_head_k = 128
llm_load_print_meta: n_embd_head_v = 128
llm_load_print_meta: n_gqa = 4
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-05
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 = 14336
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 = 0
llm_load_print_meta: rope scaling = linear
llm_load_print_meta: freq_base_train = 500000.0
llm_load_print_meta: freq_scale_train = 1
llm_load_print_meta: n_ctx_orig_yarn = 131072
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 = 8B
llm_load_print_meta: model ftype = Q8_0
llm_load_print_meta: model params = 8.03 B
llm_load_print_meta: model size = 7.95 GiB (8.50 BPW)
llm_load_print_meta: general.name = Models Meta Llama Meta Llama 3.1 8B Instruct
llm_load_print_meta: BOS token = 128000 '<|begin_of_text|>'
llm_load_print_meta: EOS token = 128009 '<|eot_id|>'
llm_load_print_meta: LF token = 128 'Ä'
llm_load_print_meta: EOT token = 128009 '<|eot_id|>'
llm_load_print_meta: max token length = 256
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: Tesla T4, compute capability 7.5, VMM: yes
llm_load_tensors: ggml ctx size = 0.14 MiB
llm_load_tensors: offloading 0 repeating layers to GPU
llm_load_tensors: offloaded 0/33 layers to GPU
llm_load_tensors: CPU buffer size = 8137.64 MiB
.........................................................................................
llama_new_context_with_model: n_ctx = 131072
llama_new_context_with_model: n_batch = 2048
llama_new_context_with_model: n_ubatch = 512
llama_new_context_with_model: flash_attn = 1
llama_new_context_with_model: freq_base = 500000.0
llama_new_context_with_model: freq_scale = 1
llama_kv_cache_init: CUDA_Host KV buffer size = 16384.00 MiB
llama_new_context_with_model: KV self size = 16384.00 MiB, K (f16): 8192.00 MiB, V (f16): 8192.00 MiB
llama_new_context_with_model: CUDA_Host output buffer size = 2.45 MiB
llama_new_context_with_model: CUDA0 compute buffer size = 929.00 MiB
llama_new_context_with_model: CUDA_Host compute buffer size = 264.01 MiB
llama_new_context_with_model: graph nodes = 903
llama_new_context_with_model: graph splits = 356
INFO [ init] initializing slots | tid="140629271105536" timestamp=1723013347 n_slots=4
INFO [ init] new slot | tid="140629271105536" timestamp=1723013347 id_slot=0 n_ctx_slot=32768
INFO [ init] new slot | tid="140629271105536" timestamp=1723013347 id_slot=1 n_ctx_slot=32768
INFO [ init] new slot | tid="140629271105536" timestamp=1723013347 id_slot=2 n_ctx_slot=32768
INFO [ init] new slot | tid="140629271105536" timestamp=1723013347 id_slot=3 n_ctx_slot=32768
INFO [ main] model loaded | tid="140629271105536" timestamp=1723013347
INFO [ main] chat template | tid="140629271105536" timestamp=1723013347 chat_example="<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHello<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nHi there<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nHow are you?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n" built_in=true
INFO [ main] HTTP server listening | tid="140629271105536" timestamp=1723013347 n_threads_http="7" port="8866" hostname="0.0.0.0"
INFO [ update_slots] all slots are idle | tid="140629271105536" timestamp=1723013347

####and If flash-attention enabled the GPU is not been utilized.

#####I installed the llama.cpp newest version ,like this:

sudo apt update
sudo apt install nvidia-cuda-toolkit

git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp

sudo apt install g++-9

export CUDA_DOCKER_ARCH=compute_70

export NVCCFLAGS="-ccbin g++-9"

cd ~/llama.cpp
make clean
GGML_CUDA_FA_ALL_QUANTS=true make GGML_CUDA=1 -j16

[hzgdeerHo]

And on the same configure of hardware, I fell llama.cpp server is more slower than llama-cpp-python api server?

[AllyourBaseBelongToUs]

Maybe I'm seeing it wrong, but in both examples, not a single Layer is offloaded to your GPU?

this:

llama_new_context_with_model: CUDA0 compute buffer size = 929.00 MiB

is significantly lower though, which would theoretically give a performance boost when using GPU

[ggerganov]

Use --n_gpu_layers 99 to offload the entire model to the GPU