github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11388] Unable to allocate CUDA0 buffer when num_gpu > 1 #33277

New Issue
Closed
opened 2026-04-22 15:48:04 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-22 15:48:04 -05:00
Owner
Copy Link

Originally created by @Ji-Peng on GitHub (Jul 12, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11388

What is the issue?

Ollama version: 0.9.6
GPU: NVIDIA GeForce RTX 4070 SUPER 12GiB
CUDA version (by nvcc --version): Cuda compilation tools, release 12.6, V12.6.85
OS: Windows 11 24H2 26100.4652

Environment variables:
OLLAMA_GPU_LAYER: cuda
OLLAMA_GPU_OVERHEAD: 536870912
OLLAMA_NUM_PARALLEL: 1

Modelfile:

FROM qwen2.5-coder:7b
PARAMETER num_ctx 1024
PARAMETER num_batch 1
PARAMETER num_gpu 2

Note: My attempts found that when num_gpu is set to a value > 1, an error will be reported: "Error: llama runner process has terminated: error loading model: unable to allocate CUDA0 buffer". When num_gpu=1, it can run, but the CPU/GPU ratio is about 60%/40%. I want the GPU ratio to be higher.
The memory part in the log: library=cuda layers.requested=2 layers.model=29 layers.offload=2 layers.split="" memory.available="[9.3 GiB]" memory.gpu_overhead="512.0 MiB" memory.required.full="5.2 GiB" memory.required.partial="1.3 GiB" memory.required.kv="112.0 MiB" memory.required.allocations="[1.3 GiB]" memory.weights.total="4.1 GiB" memory.weights.repeating="3.7 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="608.0 KiB" memory.graph.partial="427.0 MiB"

Among them: memory.available="[9.3 GiB]" is obviously larger than memory.required.full="5.2 GiB", but why is it wrong?

Relevant log output

Full log:

time=2025-07-12T10:02:51.868+08:00 level=INFO source=routes.go:1235 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:536870912 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:D:\\Ollama\\LLMInstall OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]"
time=2025-07-12T10:02:51.879+08:00 level=INFO source=images.go:476 msg="total blobs: 23"
time=2025-07-12T10:02:51.881+08:00 level=INFO source=images.go:483 msg="total unused blobs removed: 0"
time=2025-07-12T10:02:51.882+08:00 level=INFO source=routes.go:1288 msg="Listening on 127.0.0.1:11434 (version 0.9.6)"
time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu_windows.go:167 msg=packages count=1
time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=14
time=2025-07-12T10:02:52.064+08:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4070 SUPER" total="12.0 GiB" available="10.8 GiB"
[GIN] 2025/07/12 - 10:03:09 | 200 |            0s |       127.0.0.1 | HEAD     "/"
[GIN] 2025/07/12 - 10:03:09 | 200 |     56.3713ms |       127.0.0.1 | POST     "/api/show"
time=2025-07-12T10:03:09.908+08:00 level=INFO source=sched.go:788 msg="new model will fit in available VRAM in single GPU, loading" model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b parallel=1 available=10004086784 required="1.3 GiB"
time=2025-07-12T10:03:09.934+08:00 level=INFO source=server.go:135 msg="system memory" total="31.8 GiB" free="16.8 GiB" free_swap="30.1 GiB"
time=2025-07-12T10:03:09.935+08:00 level=INFO source=server.go:175 msg=offload library=cuda layers.requested=2 layers.model=29 layers.offload=2 layers.split="" memory.available="[9.3 GiB]" memory.gpu_overhead="512.0 MiB" memory.required.full="5.2 GiB" memory.required.partial="1.3 GiB" memory.required.kv="112.0 MiB" memory.required.allocations="[1.3 GiB]" memory.weights.total="4.1 GiB" memory.weights.repeating="3.7 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="608.0 KiB" memory.graph.partial="427.0 MiB"
llama_model_loader: loaded meta data with 34 key-value pairs and 339 tensors from D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 (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              = qwen2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Qwen2.5 Coder 7B Instruct
llama_model_loader: - kv   3:                           general.finetune str              = Instruct
llama_model_loader: - kv   4:                           general.basename str              = Qwen2.5-Coder
llama_model_loader: - kv   5:                         general.size_label str              = 7B
llama_model_loader: - kv   6:                            general.license str              = apache-2.0
llama_model_loader: - kv   7:                       general.license.link str              = https://huggingface.co/Qwen/Qwen2.5-C...
llama_model_loader: - kv   8:                   general.base_model.count u32              = 1
llama_model_loader: - kv   9:                  general.base_model.0.name str              = Qwen2.5 Coder 7B
llama_model_loader: - kv  10:          general.base_model.0.organization str              = Qwen
llama_model_loader: - kv  11:              general.base_model.0.repo_url str              = https://huggingface.co/Qwen/Qwen2.5-C...
llama_model_loader: - kv  12:                               general.tags arr[str,6]       = ["code", "codeqwen", "chat", "qwen", ...
llama_model_loader: - kv  13:                          general.languages arr[str,1]       = ["en"]
llama_model_loader: - kv  14:                          qwen2.block_count u32              = 28
llama_model_loader: - kv  15:                       qwen2.context_length u32              = 32768
llama_model_loader: - kv  16:                     qwen2.embedding_length u32              = 3584
llama_model_loader: - kv  17:                  qwen2.feed_forward_length u32              = 18944
llama_model_loader: - kv  18:                 qwen2.attention.head_count u32              = 28
llama_model_loader: - kv  19:              qwen2.attention.head_count_kv u32              = 4
llama_model_loader: - kv  20:                       qwen2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  21:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  22:                          general.file_type u32              = 15
llama_model_loader: - kv  23:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  24:                         tokenizer.ggml.pre str              = qwen2
llama_model_loader: - kv  25:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  26:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  27:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv  28:                tokenizer.ggml.eos_token_id u32              = 151645
llama_model_loader: - kv  29:            tokenizer.ggml.padding_token_id u32              = 151643
llama_model_loader: - kv  30:                tokenizer.ggml.bos_token_id u32              = 151643
llama_model_loader: - kv  31:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  32:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
llama_model_loader: - kv  33:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  141 tensors
llama_model_loader: - type q4_K:  169 tensors
llama_model_loader: - type q6_K:   29 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q4_K - Medium
print_info: file size   = 4.36 GiB (4.91 BPW) 
load: special tokens cache size = 22
load: token to piece cache size = 0.9310 MB
print_info: arch             = qwen2
print_info: vocab_only       = 1
print_info: model type       = ?B
print_info: model params     = 7.62 B
print_info: general.name     = Qwen2.5 Coder 7B Instruct
print_info: vocab type       = BPE
print_info: n_vocab          = 152064
print_info: n_merges         = 151387
print_info: BOS token        = 151643 '<|endoftext|>'
print_info: EOS token        = 151645 '<|im_end|>'
print_info: EOT token        = 151645 '<|im_end|>'
print_info: PAD token        = 151643 '<|endoftext|>'
print_info: LF token         = 198 'Ċ'
print_info: FIM PRE token    = 151659 '<|fim_prefix|>'
print_info: FIM SUF token    = 151661 '<|fim_suffix|>'
print_info: FIM MID token    = 151660 '<|fim_middle|>'
print_info: FIM PAD token    = 151662 '<|fim_pad|>'
print_info: FIM REP token    = 151663 '<|repo_name|>'
print_info: FIM SEP token    = 151664 '<|file_sep|>'
print_info: EOG token        = 151643 '<|endoftext|>'
print_info: EOG token        = 151645 '<|im_end|>'
print_info: EOG token        = 151662 '<|fim_pad|>'
print_info: EOG token        = 151663 '<|repo_name|>'
print_info: EOG token        = 151664 '<|file_sep|>'
print_info: max token length = 256
llama_model_load: vocab only - skipping tensors
time=2025-07-12T10:03:10.336+08:00 level=INFO source=server.go:438 msg="starting llama server" cmd="D:\\Ollama\\OllamaInstall\\ollama.exe runner --model D:\\Ollama\\LLMInstall\\blobs\\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 --ctx-size 1024 --batch-size 1 --n-gpu-layers 2 --threads 8 --no-mmap --parallel 1 --port 53356"
time=2025-07-12T10:03:10.343+08:00 level=INFO source=sched.go:483 msg="loaded runners" count=1
time=2025-07-12T10:03:10.344+08:00 level=INFO source=server.go:598 msg="waiting for llama runner to start responding"
time=2025-07-12T10:03:10.345+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error"
time=2025-07-12T10:03:10.394+08:00 level=INFO source=runner.go:815 msg="starting go runner"
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: NVIDIA GeForce RTX 4070 SUPER, compute capability 8.9, VMM: yes
load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cuda.dll
load_backend: loaded CPU backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cpu-icelake.dll
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: NVIDIA GeForce RTX 4070 SUPER, compute capability 8.9, VMM: yes
load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\cuda_v12\ggml-cuda.dll
time=2025-07-12T10:03:10.516+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 CUDA.1.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.1.USE_GRAPHS=1 CUDA.1.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang)
time=2025-07-12T10:03:10.518+08:00 level=INFO source=runner.go:874 msg="Server listening on 127.0.0.1:53356"
time=2025-07-12T10:03:10.598+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server loading model"
llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 4070 SUPER) - 11053 MiB free
llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 4070 SUPER) - 11053 MiB free
llama_model_loader: loaded meta data with 34 key-value pairs and 339 tensors from D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 (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              = qwen2
llama_model_loader: - kv   1:                               general.type str              = model
llama_model_loader: - kv   2:                               general.name str              = Qwen2.5 Coder 7B Instruct
llama_model_loader: - kv   3:                           general.finetune str              = Instruct
llama_model_loader: - kv   4:                           general.basename str              = Qwen2.5-Coder
llama_model_loader: - kv   5:                         general.size_label str              = 7B
llama_model_loader: - kv   6:                            general.license str              = apache-2.0
llama_model_loader: - kv   7:                       general.license.link str              = https://huggingface.co/Qwen/Qwen2.5-C...
llama_model_loader: - kv   8:                   general.base_model.count u32              = 1
llama_model_loader: - kv   9:                  general.base_model.0.name str              = Qwen2.5 Coder 7B
llama_model_loader: - kv  10:          general.base_model.0.organization str              = Qwen
llama_model_loader: - kv  11:              general.base_model.0.repo_url str              = https://huggingface.co/Qwen/Qwen2.5-C...
llama_model_loader: - kv  12:                               general.tags arr[str,6]       = ["code", "codeqwen", "chat", "qwen", ...
llama_model_loader: - kv  13:                          general.languages arr[str,1]       = ["en"]
llama_model_loader: - kv  14:                          qwen2.block_count u32              = 28
llama_model_loader: - kv  15:                       qwen2.context_length u32              = 32768
llama_model_loader: - kv  16:                     qwen2.embedding_length u32              = 3584
llama_model_loader: - kv  17:                  qwen2.feed_forward_length u32              = 18944
llama_model_loader: - kv  18:                 qwen2.attention.head_count u32              = 28
llama_model_loader: - kv  19:              qwen2.attention.head_count_kv u32              = 4
llama_model_loader: - kv  20:                       qwen2.rope.freq_base f32              = 1000000.000000
llama_model_loader: - kv  21:     qwen2.attention.layer_norm_rms_epsilon f32              = 0.000001
llama_model_loader: - kv  22:                          general.file_type u32              = 15
llama_model_loader: - kv  23:                       tokenizer.ggml.model str              = gpt2
llama_model_loader: - kv  24:                         tokenizer.ggml.pre str              = qwen2
llama_model_loader: - kv  25:                      tokenizer.ggml.tokens arr[str,152064]  = ["!", "\"", "#", "$", "%", "&", "'", ...
llama_model_loader: - kv  26:                  tokenizer.ggml.token_type arr[i32,152064]  = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ...
llama_model_loader: - kv  27:                      tokenizer.ggml.merges arr[str,151387]  = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",...
llama_model_loader: - kv  28:                tokenizer.ggml.eos_token_id u32              = 151645
llama_model_loader: - kv  29:            tokenizer.ggml.padding_token_id u32              = 151643
llama_model_loader: - kv  30:                tokenizer.ggml.bos_token_id u32              = 151643
llama_model_loader: - kv  31:               tokenizer.ggml.add_bos_token bool             = false
llama_model_loader: - kv  32:                    tokenizer.chat_template str              = {%- if tools %}\n    {{- '<|im_start|>...
llama_model_loader: - kv  33:               general.quantization_version u32              = 2
llama_model_loader: - type  f32:  141 tensors
llama_model_loader: - type q4_K:  169 tensors
llama_model_loader: - type q6_K:   29 tensors
print_info: file format = GGUF V3 (latest)
print_info: file type   = Q4_K - Medium
print_info: file size   = 4.36 GiB (4.91 BPW) 
load: special tokens cache size = 22
load: token to piece cache size = 0.9310 MB
print_info: arch             = qwen2
print_info: vocab_only       = 0
print_info: n_ctx_train      = 32768
print_info: n_embd           = 3584
print_info: n_layer          = 28
print_info: n_head           = 28
print_info: n_head_kv        = 4
print_info: n_rot            = 128
print_info: n_swa            = 0
print_info: n_swa_pattern    = 1
print_info: n_embd_head_k    = 128
print_info: n_embd_head_v    = 128
print_info: n_gqa            = 7
print_info: n_embd_k_gqa     = 512
print_info: n_embd_v_gqa     = 512
print_info: f_norm_eps       = 0.0e+00
print_info: f_norm_rms_eps   = 1.0e-06
print_info: f_clamp_kqv      = 0.0e+00
print_info: f_max_alibi_bias = 0.0e+00
print_info: f_logit_scale    = 0.0e+00
print_info: f_attn_scale     = 0.0e+00
print_info: n_ff             = 18944
print_info: n_expert         = 0
print_info: n_expert_used    = 0
print_info: causal attn      = 1
print_info: pooling type     = -1
print_info: rope type        = 2
print_info: rope scaling     = linear
print_info: freq_base_train  = 1000000.0
print_info: freq_scale_train = 1
print_info: n_ctx_orig_yarn  = 32768
print_info: rope_finetuned   = unknown
print_info: ssm_d_conv       = 0
print_info: ssm_d_inner      = 0
print_info: ssm_d_state      = 0
print_info: ssm_dt_rank      = 0
print_info: ssm_dt_b_c_rms   = 0
print_info: model type       = 7B
print_info: model params     = 7.62 B
print_info: general.name     = Qwen2.5 Coder 7B Instruct
print_info: vocab type       = BPE
print_info: n_vocab          = 152064
print_info: n_merges         = 151387
print_info: BOS token        = 151643 '<|endoftext|>'
print_info: EOS token        = 151645 '<|im_end|>'
print_info: EOT token        = 151645 '<|im_end|>'
print_info: PAD token        = 151643 '<|endoftext|>'
print_info: LF token         = 198 'Ċ'
print_info: FIM PRE token    = 151659 '<|fim_prefix|>'
print_info: FIM SUF token    = 151661 '<|fim_suffix|>'
print_info: FIM MID token    = 151660 '<|fim_middle|>'
print_info: FIM PAD token    = 151662 '<|fim_pad|>'
print_info: FIM REP token    = 151663 '<|repo_name|>'
print_info: FIM SEP token    = 151664 '<|file_sep|>'
print_info: EOG token        = 151643 '<|endoftext|>'
print_info: EOG token        = 151645 '<|im_end|>'
print_info: EOG token        = 151662 '<|fim_pad|>'
print_info: EOG token        = 151663 '<|repo_name|>'
print_info: EOG token        = 151664 '<|file_sep|>'
print_info: max token length = 256
load_tensors: loading model tensors, this can take a while... (mmap = false)
alloc_tensor_range: failed to initialize tensor blk.27.attn_q.weight
llama_model_load: error loading model: unable to allocate CUDA0 buffer
llama_model_load_from_file_impl: failed to load model
panic: unable to load model: D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463

goroutine 12 [running]:
github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc00010c360, {0x2, 0x0, 0x0, {0x0, 0x0, 0x0}, 0xc000483da0, 0x0}, {0xc00003a1c0, ...}, ...)
	C:/a/ollama/ollama/runner/llamarunner/runner.go:751 +0x395
created by github.com/ollama/ollama/runner/llamarunner.Execute in goroutine 1
	C:/a/ollama/ollama/runner/llamarunner/runner.go:848 +0xb57
time=2025-07-12T10:03:10.869+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error"
time=2025-07-12T10:03:10.893+08:00 level=ERROR source=server.go:464 msg="llama runner terminated" error="exit status 2"
time=2025-07-12T10:03:11.120+08:00 level=ERROR source=sched.go:489 msg="error loading llama server" error="llama runner process has terminated: error loading model: unable to allocate CUDA0 buffer"
[GIN] 2025/07/12 - 10:03:11 | 500 |    1.4203973s |       127.0.0.1 | POST     "/api/generate"

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.9.6

Originally created by @Ji-Peng on GitHub (Jul 12, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11388 ### What is the issue? Ollama version: 0.9.6 GPU: NVIDIA GeForce RTX 4070 SUPER 12GiB CUDA version (by nvcc --version): Cuda compilation tools, release 12.6, V12.6.85 OS: Windows 11 24H2 26100.4652 Environment variables: OLLAMA_GPU_LAYER: cuda OLLAMA_GPU_OVERHEAD: 536870912 OLLAMA_NUM_PARALLEL: 1 Modelfile: ``` FROM qwen2.5-coder:7b PARAMETER num_ctx 1024 PARAMETER num_batch 1 PARAMETER num_gpu 2 ``` Note: My attempts found that when num_gpu is set to a value > 1, an error will be reported: "Error: llama runner process has terminated: error loading model: unable to allocate CUDA0 buffer". When num_gpu=1, it can run, but the CPU/GPU ratio is about 60%/40%. I want the GPU ratio to be higher. The memory part in the log: `library=cuda layers.requested=2 layers.model=29 layers.offload=2 layers.split="" memory.available="[9.3 GiB]" memory.gpu_overhead="512.0 MiB" memory.required.full="5.2 GiB" memory.required.partial="1.3 GiB" memory.required.kv="112.0 MiB" memory.required.allocations="[1.3 GiB]" memory.weights.total="4.1 GiB" memory.weights.repeating="3.7 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="608.0 KiB" memory.graph.partial="427.0 MiB"` Among them: memory.available="[9.3 GiB]" is obviously larger than memory.required.full="5.2 GiB", but why is it wrong? ### Relevant log output ```shell Full log: time=2025-07-12T10:02:51.868+08:00 level=INFO source=routes.go:1235 msg="server config" env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_CONTEXT_LENGTH:4096 OLLAMA_DEBUG:INFO OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:536870912 OLLAMA_HOST:http://127.0.0.1:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:D:\\Ollama\\LLMInstall OLLAMA_MULTIUSER_CACHE:false OLLAMA_NEW_ENGINE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:1 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://* vscode-file://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]" time=2025-07-12T10:02:51.879+08:00 level=INFO source=images.go:476 msg="total blobs: 23" time=2025-07-12T10:02:51.881+08:00 level=INFO source=images.go:483 msg="total unused blobs removed: 0" time=2025-07-12T10:02:51.882+08:00 level=INFO source=routes.go:1288 msg="Listening on 127.0.0.1:11434 (version 0.9.6)" time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs" time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu_windows.go:167 msg=packages count=1 time=2025-07-12T10:02:51.882+08:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=8 efficiency=0 threads=14 time=2025-07-12T10:02:52.064+08:00 level=INFO source=types.go:130 msg="inference compute" id=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b library=cuda variant=v12 compute=8.9 driver=12.7 name="NVIDIA GeForce RTX 4070 SUPER" total="12.0 GiB" available="10.8 GiB" [GIN] 2025/07/12 - 10:03:09 | 200 | 0s | 127.0.0.1 | HEAD "/" [GIN] 2025/07/12 - 10:03:09 | 200 | 56.3713ms | 127.0.0.1 | POST "/api/show" time=2025-07-12T10:03:09.908+08:00 level=INFO source=sched.go:788 msg="new model will fit in available VRAM in single GPU, loading" model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b parallel=1 available=10004086784 required="1.3 GiB" time=2025-07-12T10:03:09.934+08:00 level=INFO source=server.go:135 msg="system memory" total="31.8 GiB" free="16.8 GiB" free_swap="30.1 GiB" time=2025-07-12T10:03:09.935+08:00 level=INFO source=server.go:175 msg=offload library=cuda layers.requested=2 layers.model=29 layers.offload=2 layers.split="" memory.available="[9.3 GiB]" memory.gpu_overhead="512.0 MiB" memory.required.full="5.2 GiB" memory.required.partial="1.3 GiB" memory.required.kv="112.0 MiB" memory.required.allocations="[1.3 GiB]" memory.weights.total="4.1 GiB" memory.weights.repeating="3.7 GiB" memory.weights.nonrepeating="426.4 MiB" memory.graph.full="608.0 KiB" memory.graph.partial="427.0 MiB" llama_model_loader: loaded meta data with 34 key-value pairs and 339 tensors from D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 (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 = qwen2 llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Qwen2.5 Coder 7B Instruct llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.basename str = Qwen2.5-Coder llama_model_loader: - kv 5: general.size_label str = 7B llama_model_loader: - kv 6: general.license str = apache-2.0 llama_model_loader: - kv 7: general.license.link str = https://huggingface.co/Qwen/Qwen2.5-C... llama_model_loader: - kv 8: general.base_model.count u32 = 1 llama_model_loader: - kv 9: general.base_model.0.name str = Qwen2.5 Coder 7B llama_model_loader: - kv 10: general.base_model.0.organization str = Qwen llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen2.5-C... llama_model_loader: - kv 12: general.tags arr[str,6] = ["code", "codeqwen", "chat", "qwen", ... llama_model_loader: - kv 13: general.languages arr[str,1] = ["en"] llama_model_loader: - kv 14: qwen2.block_count u32 = 28 llama_model_loader: - kv 15: qwen2.context_length u32 = 32768 llama_model_loader: - kv 16: qwen2.embedding_length u32 = 3584 llama_model_loader: - kv 17: qwen2.feed_forward_length u32 = 18944 llama_model_loader: - kv 18: qwen2.attention.head_count u32 = 28 llama_model_loader: - kv 19: qwen2.attention.head_count_kv u32 = 4 llama_model_loader: - kv 20: qwen2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 21: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 llama_model_loader: - kv 22: general.file_type u32 = 15 llama_model_loader: - kv 23: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 24: tokenizer.ggml.pre str = qwen2 llama_model_loader: - kv 25: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 26: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 27: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 151643 llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 151643 llama_model_loader: - kv 31: tokenizer.ggml.add_bos_token bool = false llama_model_loader: - kv 32: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>... llama_model_loader: - kv 33: general.quantization_version u32 = 2 llama_model_loader: - type f32: 141 tensors llama_model_loader: - type q4_K: 169 tensors llama_model_loader: - type q6_K: 29 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q4_K - Medium print_info: file size = 4.36 GiB (4.91 BPW) load: special tokens cache size = 22 load: token to piece cache size = 0.9310 MB print_info: arch = qwen2 print_info: vocab_only = 1 print_info: model type = ?B print_info: model params = 7.62 B print_info: general.name = Qwen2.5 Coder 7B Instruct print_info: vocab type = BPE print_info: n_vocab = 152064 print_info: n_merges = 151387 print_info: BOS token = 151643 '<|endoftext|>' print_info: EOS token = 151645 '<|im_end|>' print_info: EOT token = 151645 '<|im_end|>' print_info: PAD token = 151643 '<|endoftext|>' print_info: LF token = 198 'Ċ' print_info: FIM PRE token = 151659 '<|fim_prefix|>' print_info: FIM SUF token = 151661 '<|fim_suffix|>' print_info: FIM MID token = 151660 '<|fim_middle|>' print_info: FIM PAD token = 151662 '<|fim_pad|>' print_info: FIM REP token = 151663 '<|repo_name|>' print_info: FIM SEP token = 151664 '<|file_sep|>' print_info: EOG token = 151643 '<|endoftext|>' print_info: EOG token = 151645 '<|im_end|>' print_info: EOG token = 151662 '<|fim_pad|>' print_info: EOG token = 151663 '<|repo_name|>' print_info: EOG token = 151664 '<|file_sep|>' print_info: max token length = 256 llama_model_load: vocab only - skipping tensors time=2025-07-12T10:03:10.336+08:00 level=INFO source=server.go:438 msg="starting llama server" cmd="D:\\Ollama\\OllamaInstall\\ollama.exe runner --model D:\\Ollama\\LLMInstall\\blobs\\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 --ctx-size 1024 --batch-size 1 --n-gpu-layers 2 --threads 8 --no-mmap --parallel 1 --port 53356" time=2025-07-12T10:03:10.343+08:00 level=INFO source=sched.go:483 msg="loaded runners" count=1 time=2025-07-12T10:03:10.344+08:00 level=INFO source=server.go:598 msg="waiting for llama runner to start responding" time=2025-07-12T10:03:10.345+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error" time=2025-07-12T10:03:10.394+08:00 level=INFO source=runner.go:815 msg="starting go runner" 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: NVIDIA GeForce RTX 4070 SUPER, compute capability 8.9, VMM: yes load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cuda.dll load_backend: loaded CPU backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cpu-icelake.dll 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: NVIDIA GeForce RTX 4070 SUPER, compute capability 8.9, VMM: yes load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\cuda_v12\ggml-cuda.dll time=2025-07-12T10:03:10.516+08:00 level=INFO source=ggml.go:104 msg=system CPU.0.SSE3=1 CPU.0.SSSE3=1 CPU.0.AVX=1 CPU.0.AVX2=1 CPU.0.F16C=1 CPU.0.FMA=1 CPU.0.BMI2=1 CPU.0.AVX512=1 CPU.0.AVX512_VBMI=1 CPU.0.AVX512_VNNI=1 CPU.0.LLAMAFILE=1 CPU.1.LLAMAFILE=1 CUDA.0.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.0.USE_GRAPHS=1 CUDA.0.PEER_MAX_BATCH_SIZE=128 CUDA.1.ARCHS=500,600,610,700,750,800,860,870,890,900,1200 CUDA.1.USE_GRAPHS=1 CUDA.1.PEER_MAX_BATCH_SIZE=128 compiler=cgo(clang) time=2025-07-12T10:03:10.518+08:00 level=INFO source=runner.go:874 msg="Server listening on 127.0.0.1:53356" time=2025-07-12T10:03:10.598+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server loading model" llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 4070 SUPER) - 11053 MiB free llama_model_load_from_file_impl: using device CUDA0 (NVIDIA GeForce RTX 4070 SUPER) - 11053 MiB free llama_model_loader: loaded meta data with 34 key-value pairs and 339 tensors from D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 (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 = qwen2 llama_model_loader: - kv 1: general.type str = model llama_model_loader: - kv 2: general.name str = Qwen2.5 Coder 7B Instruct llama_model_loader: - kv 3: general.finetune str = Instruct llama_model_loader: - kv 4: general.basename str = Qwen2.5-Coder llama_model_loader: - kv 5: general.size_label str = 7B llama_model_loader: - kv 6: general.license str = apache-2.0 llama_model_loader: - kv 7: general.license.link str = https://huggingface.co/Qwen/Qwen2.5-C... llama_model_loader: - kv 8: general.base_model.count u32 = 1 llama_model_loader: - kv 9: general.base_model.0.name str = Qwen2.5 Coder 7B llama_model_loader: - kv 10: general.base_model.0.organization str = Qwen llama_model_loader: - kv 11: general.base_model.0.repo_url str = https://huggingface.co/Qwen/Qwen2.5-C... llama_model_loader: - kv 12: general.tags arr[str,6] = ["code", "codeqwen", "chat", "qwen", ... llama_model_loader: - kv 13: general.languages arr[str,1] = ["en"] llama_model_loader: - kv 14: qwen2.block_count u32 = 28 llama_model_loader: - kv 15: qwen2.context_length u32 = 32768 llama_model_loader: - kv 16: qwen2.embedding_length u32 = 3584 llama_model_loader: - kv 17: qwen2.feed_forward_length u32 = 18944 llama_model_loader: - kv 18: qwen2.attention.head_count u32 = 28 llama_model_loader: - kv 19: qwen2.attention.head_count_kv u32 = 4 llama_model_loader: - kv 20: qwen2.rope.freq_base f32 = 1000000.000000 llama_model_loader: - kv 21: qwen2.attention.layer_norm_rms_epsilon f32 = 0.000001 llama_model_loader: - kv 22: general.file_type u32 = 15 llama_model_loader: - kv 23: tokenizer.ggml.model str = gpt2 llama_model_loader: - kv 24: tokenizer.ggml.pre str = qwen2 llama_model_loader: - kv 25: tokenizer.ggml.tokens arr[str,152064] = ["!", "\"", "#", "$", "%", "&", "'", ... llama_model_loader: - kv 26: tokenizer.ggml.token_type arr[i32,152064] = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, ... llama_model_loader: - kv 27: tokenizer.ggml.merges arr[str,151387] = ["Ġ Ġ", "ĠĠ ĠĠ", "i n", "Ġ t",... llama_model_loader: - kv 28: tokenizer.ggml.eos_token_id u32 = 151645 llama_model_loader: - kv 29: tokenizer.ggml.padding_token_id u32 = 151643 llama_model_loader: - kv 30: tokenizer.ggml.bos_token_id u32 = 151643 llama_model_loader: - kv 31: tokenizer.ggml.add_bos_token bool = false llama_model_loader: - kv 32: tokenizer.chat_template str = {%- if tools %}\n {{- '<|im_start|>... llama_model_loader: - kv 33: general.quantization_version u32 = 2 llama_model_loader: - type f32: 141 tensors llama_model_loader: - type q4_K: 169 tensors llama_model_loader: - type q6_K: 29 tensors print_info: file format = GGUF V3 (latest) print_info: file type = Q4_K - Medium print_info: file size = 4.36 GiB (4.91 BPW) load: special tokens cache size = 22 load: token to piece cache size = 0.9310 MB print_info: arch = qwen2 print_info: vocab_only = 0 print_info: n_ctx_train = 32768 print_info: n_embd = 3584 print_info: n_layer = 28 print_info: n_head = 28 print_info: n_head_kv = 4 print_info: n_rot = 128 print_info: n_swa = 0 print_info: n_swa_pattern = 1 print_info: n_embd_head_k = 128 print_info: n_embd_head_v = 128 print_info: n_gqa = 7 print_info: n_embd_k_gqa = 512 print_info: n_embd_v_gqa = 512 print_info: f_norm_eps = 0.0e+00 print_info: f_norm_rms_eps = 1.0e-06 print_info: f_clamp_kqv = 0.0e+00 print_info: f_max_alibi_bias = 0.0e+00 print_info: f_logit_scale = 0.0e+00 print_info: f_attn_scale = 0.0e+00 print_info: n_ff = 18944 print_info: n_expert = 0 print_info: n_expert_used = 0 print_info: causal attn = 1 print_info: pooling type = -1 print_info: rope type = 2 print_info: rope scaling = linear print_info: freq_base_train = 1000000.0 print_info: freq_scale_train = 1 print_info: n_ctx_orig_yarn = 32768 print_info: rope_finetuned = unknown print_info: ssm_d_conv = 0 print_info: ssm_d_inner = 0 print_info: ssm_d_state = 0 print_info: ssm_dt_rank = 0 print_info: ssm_dt_b_c_rms = 0 print_info: model type = 7B print_info: model params = 7.62 B print_info: general.name = Qwen2.5 Coder 7B Instruct print_info: vocab type = BPE print_info: n_vocab = 152064 print_info: n_merges = 151387 print_info: BOS token = 151643 '<|endoftext|>' print_info: EOS token = 151645 '<|im_end|>' print_info: EOT token = 151645 '<|im_end|>' print_info: PAD token = 151643 '<|endoftext|>' print_info: LF token = 198 'Ċ' print_info: FIM PRE token = 151659 '<|fim_prefix|>' print_info: FIM SUF token = 151661 '<|fim_suffix|>' print_info: FIM MID token = 151660 '<|fim_middle|>' print_info: FIM PAD token = 151662 '<|fim_pad|>' print_info: FIM REP token = 151663 '<|repo_name|>' print_info: FIM SEP token = 151664 '<|file_sep|>' print_info: EOG token = 151643 '<|endoftext|>' print_info: EOG token = 151645 '<|im_end|>' print_info: EOG token = 151662 '<|fim_pad|>' print_info: EOG token = 151663 '<|repo_name|>' print_info: EOG token = 151664 '<|file_sep|>' print_info: max token length = 256 load_tensors: loading model tensors, this can take a while... (mmap = false) alloc_tensor_range: failed to initialize tensor blk.27.attn_q.weight llama_model_load: error loading model: unable to allocate CUDA0 buffer llama_model_load_from_file_impl: failed to load model panic: unable to load model: D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 goroutine 12 [running]: github.com/ollama/ollama/runner/llamarunner.(*Server).loadModel(0xc00010c360, {0x2, 0x0, 0x0, {0x0, 0x0, 0x0}, 0xc000483da0, 0x0}, {0xc00003a1c0, ...}, ...) C:/a/ollama/ollama/runner/llamarunner/runner.go:751 +0x395 created by github.com/ollama/ollama/runner/llamarunner.Execute in goroutine 1 C:/a/ollama/ollama/runner/llamarunner/runner.go:848 +0xb57 time=2025-07-12T10:03:10.869+08:00 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error" time=2025-07-12T10:03:10.893+08:00 level=ERROR source=server.go:464 msg="llama runner terminated" error="exit status 2" time=2025-07-12T10:03:11.120+08:00 level=ERROR source=sched.go:489 msg="error loading llama server" error="llama runner process has terminated: error loading model: unable to allocate CUDA0 buffer" [GIN] 2025/07/12 - 10:03:11 | 500 | 1.4203973s | 127.0.0.1 | POST "/api/generate" ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.9.6
GiteaMirror added the bug label 2026-04-22 15:48:04 -05:00
GiteaMirror commented 2026-04-22 15:48:07 -05:00
Author
Owner
Copy Link

@Ji-Peng commented on GitHub (Jul 12, 2025):

More log with level=DEBUG:

time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:491 msg="triggering expiration for failed load" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024
[GIN] 2025/07/12 - 10:34:55 | 500 |    1.4827842s |       127.0.0.1 | POST     "/api/generate"
time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:364 msg="runner expired event received" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024
time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:379 msg="got lock to unload expired event" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024
time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:402 msg="starting background wait for VRAM recovery" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024
time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.2 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.2 GiB"
time=2025-07-12T10:34:55.651+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB"
releasing nvml library
time=2025-07-12T10:34:55.685+08:00 level=DEBUG source=server.go:1031 msg="stopping llama server" pid=23564
time=2025-07-12T10:34:55.685+08:00 level=DEBUG source=sched.go:407 msg="runner terminated and removed from list, blocking for VRAM recovery" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463
time=2025-07-12T10:34:55.903+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.2 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB"
time=2025-07-12T10:34:55.927+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB"
releasing nvml library
time=2025-07-12T10:34:56.153+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB"
time=2025-07-12T10:34:56.179+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB"
releasing nvml library
time=2025-07-12T10:34:56.403+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB"
time=2025-07-12T10:34:56.413+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB"
releasing nvml library
time=2025-07-12T10:34:56.653+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB"
time=2025-07-12T10:34:56.678+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB"
releasing nvml library
time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:700 msg="gpu VRAM free memory converged after 1.05 seconds" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463
time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:410 msg="sending an unloaded event" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463
time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:312 msg="ignoring unload event with no pending requests"
<!-- gh-comment-id:3064544841 --> @Ji-Peng commented on GitHub (Jul 12, 2025): More log with level=DEBUG: ``` time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:491 msg="triggering expiration for failed load" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024 [GIN] 2025/07/12 - 10:34:55 | 500 | 1.4827842s | 127.0.0.1 | POST "/api/generate" time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:364 msg="runner expired event received" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024 time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:379 msg="got lock to unload expired event" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024 time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=sched.go:402 msg="starting background wait for VRAM recovery" runner.name=registry.ollama.ai/library/qwen2.5-coder:7b-16k-gpu runner.inference=cuda runner.devices=1 runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 runner.num_ctx=1024 time=2025-07-12T10:34:55.629+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.2 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.2 GiB" time=2025-07-12T10:34:55.651+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB" releasing nvml library time=2025-07-12T10:34:55.685+08:00 level=DEBUG source=server.go:1031 msg="stopping llama server" pid=23564 time=2025-07-12T10:34:55.685+08:00 level=DEBUG source=sched.go:407 msg="runner terminated and removed from list, blocking for VRAM recovery" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 time=2025-07-12T10:34:55.903+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.2 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB" time=2025-07-12T10:34:55.927+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB" releasing nvml library time=2025-07-12T10:34:56.153+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB" time=2025-07-12T10:34:56.179+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB" releasing nvml library time=2025-07-12T10:34:56.403+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB" time=2025-07-12T10:34:56.413+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB" releasing nvml library time=2025-07-12T10:34:56.653+08:00 level=DEBUG source=gpu.go:391 msg="updating system memory data" before.total="31.8 GiB" before.free="16.1 GiB" before.free_swap="28.3 GiB" now.total="31.8 GiB" now.free="16.1 GiB" now.free_swap="28.3 GiB" time=2025-07-12T10:34:56.678+08:00 level=DEBUG source=gpu.go:441 msg="updating cuda memory data" gpu=GPU-cb15de68-b178-c58d-6cac-5fa0b4b5b91b name="NVIDIA GeForce RTX 4070 SUPER" overhead="0 B" before.total="12.0 GiB" before.free="9.3 GiB" now.total="12.0 GiB" now.free="9.3 GiB" now.used="2.7 GiB" releasing nvml library time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:700 msg="gpu VRAM free memory converged after 1.05 seconds" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:410 msg="sending an unloaded event" runner.size="5.2 GiB" runner.vram="1.3 GiB" runner.parallel=1 runner.pid=23564 runner.model=D:\Ollama\LLMInstall\blobs\sha256-60e05f2100071479f596b964f89f510f057ce397ea22f2833a0cfe029bfc2463 time=2025-07-12T10:34:56.680+08:00 level=DEBUG source=sched.go:312 msg="ignoring unload event with no pending requests" ```
GiteaMirror commented 2026-04-22 15:48:07 -05:00
Author
Owner
Copy Link

@Ji-Peng commented on GitHub (Jul 12, 2025):

I now switch to LM studio, and with the same model and similar configuration, I can fully utilize the GPU.

<!-- gh-comment-id:3064737665 --> @Ji-Peng commented on GitHub (Jul 12, 2025): I now switch to LM studio, and with the same model and similar configuration, I can fully utilize the GPU.
GiteaMirror commented 2026-04-22 15:48:09 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 12, 2025):

load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cuda.dll
load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\cuda_v12\ggml-cuda.dll

ollama is loading the cuda library twice. https://github.com/ollama/ollama/issues/11211#issuecomment-3014589554

<!-- gh-comment-id:3065688753 --> @rick-github commented on GitHub (Jul 12, 2025): ``` load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\ggml-cuda.dll load_backend: loaded CUDA backend from D:\Ollama\OllamaInstall\lib\ollama\cuda_v12\ggml-cuda.dll ``` ollama is loading the cuda library twice. https://github.com/ollama/ollama/issues/11211#issuecomment-3014589554
GiteaMirror commented 2026-04-22 15:48:11 -05:00
Author
Owner
Copy Link

@Ji-Peng commented on GitHub (Jul 13, 2025):

Solved! Thanks!

<!-- gh-comment-id:3066346430 --> @Ji-Peng commented on GitHub (Jul 13, 2025): Solved! Thanks!
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#33277
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?