Bug: Vulkan not compile #9582
Labels
bug-unconfirmed
critical severity
Used to report critical severity bugs in llama.cpp (e.g. Crashing, Corrupted, Dataloss)
Comments
|
I experience the same issue when trying to compile whisper.cpp with Vulkan enabled on macOS arm m1 with Vulkan |
|
same thing with me |
|
I also got the "Failed to create pipes" error. On Linux for me it was resolved by:
Why I think this works... This ulimit command increases the files that can be open by the OS at any time to 2048 and pipes require virtual files to function. This ulimit command is temporary and should reset to the default after a reboot (my system defaulted to 1024). This does not resolve whisper.cpp's vulkan compilation for me. |
|
the ulimit -n 2048 is fix the issue on my hardware and distro |
jeffbolznv
added a commit
to jeffbolznv/llama.cpp
that referenced
this issue
Nov 8, 2024
Fixes ggerganov#9582 Spawning too many concurrent copies of glslc leads to "Failed to create pipes" errors on Linux. This change applies the same throttling we use for multithreaded pipeline creation.
4 tasks
|
I think #10222 should fix this, but I only had a repro in a branch that had more shaders. Can somebody who has a repro give it a try? |
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this issue
Nov 15, 2024
…ggerganov#10222) Fixes ggerganov#9582 Spawning too many concurrent copies of glslc leads to "Failed to create pipes" errors on Linux. This change applies the same throttling we use for multithreaded pipeline creation.
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this issue
Nov 18, 2024
…ggerganov#10222) Fixes ggerganov#9582 Spawning too many concurrent copies of glslc leads to "Failed to create pipes" errors on Linux. This change applies the same throttling we use for multithreaded pipeline creation.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
What happened?
~/llama.cpp (master)> cmake -B build -DGGML_VULKAN=1
cmake --build build --config Release -j 16
-- The C compiler identification is GNU 14.2.1
-- The CXX compiler identification is GNU 14.2.1
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/cc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Found Git: /usr/bin/git (found version "2.46.1")
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success
-- Found Threads: TRUE
-- Found OpenMP_C: -fopenmp (found version "4.5")
-- Found OpenMP_CXX: -fopenmp (found version "4.5")
-- Found OpenMP: TRUE (found version "4.5")
-- OpenMP found
-- Using llamafile
-- Found Vulkan: /lib/libvulkan.so (found version "1.3.295") found components: glslc glslangValidator
-- Vulkan found
-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with GGML_CCACHE=OFF
-- CMAKE_SYSTEM_PROCESSOR: x86_64
-- x86 detected
-- Configuring done (1.0s)
-- Generating done (0.1s)
-- Build files have been written to: /home/vecna/llama.cpp/build
[ 0%] Generating build details from Git
[ 1%] Building C object examples/gguf-hash/CMakeFiles/sha256.dir/deps/sha256/sha256.c.o
[ 1%] Building CXX object ggml/src/vulkan-shaders/CMakeFiles/vulkan-shaders-gen.dir/vulkan-shaders-gen.cpp.o
[ 1%] Building C object examples/gguf-hash/CMakeFiles/sha1.dir/deps/sha1/sha1.c.o
[ 2%] Building C object examples/gguf-hash/CMakeFiles/xxhash.dir/deps/xxhash/xxhash.c.o
-- Found Git: /usr/bin/git (found version "2.46.1")
[ 3%] Building CXX object common/CMakeFiles/build_info.dir/build-info.cpp.o
[ 3%] Built target build_info
In function ‘SHA1Update’,
inlined from ‘SHA1Final’ at /home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:265:5:
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:219:13: warning: ‘SHA1Transform’ reading 64 bytes from a region of size 0 [-Wstringop-overread]
219 | SHA1Transform(context->state, &data[i]);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:219:13: note: referencing argument 2 of type ‘const unsigned char[64]’
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c: In function ‘SHA1Final’:
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:54:6: note: in a call to function ‘SHA1Transform’
54 | void SHA1Transform(
| ^~~~~~~~~~~~~
In function ‘SHA1Update’,
inlined from ‘SHA1Final’ at /home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:269:9:
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:219:13: warning: ‘SHA1Transform’ reading 64 bytes from a region of size 0 [-Wstringop-overread]
219 | SHA1Transform(context->state, &data[i]);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:219:13: note: referencing argument 2 of type ‘const unsigned char[64]’
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c: In function ‘SHA1Final’:
/home/vecna/llama.cpp/examples/gguf-hash/deps/sha1/sha1.c:54:6: note: in a call to function ‘SHA1Transform’
54 | void SHA1Transform(
| ^~~~~~~~~~~~~
[ 3%] Built target sha1
[ 3%] Built target sha256
[ 3%] Built target xxhash
[ 4%] Linking CXX executable ../../../bin/vulkan-shaders-gen
[ 4%] Built target vulkan-shaders-gen
[ 5%] Generate vulkan shaders
ggml_vulkan: Generating and compiling shaders to SPIR-V
Error executing command for Error executing command for Error executing command for Error executing command for Error executing command for matmul_id_q5_k_f32_aligned_fp32Error executing command for Error executing command for matmul_f16_alignedError executing command for mul_mat_vec_id_q3_k_f32: Failed to create pipes
Error executing command for dequant_q6_k: Failed to create pipes
dequant_iq4_nl: Failed to create pipes
dequant_q2_k: Failed to create pipes
Error executing command for Error executing command for matmul_id_q4_0_f32_fp32: Failed to create pipes
matmul_q4_1_f32_alignedError executing command for mul_mat_vec_q3_k_f32_f32: Error executing command for mul_mat_vec_id_iq4_nl_f32: Failed to create pipes
mul_mat_vec_iq4_nl_f16_f32: Error executing command for Failed to create pipes
: Error executing command for Error executing command for mul_mat_vec_iq4_nl_f32_f32: Failed to create pipes
: Failed to create pipes
matmul_f32_f32mul_mat_vec_id_q6_k_f32: Failed to create pipes
: Failed to create pipesget_rows_iq4_nl: Error executing command for Failed to create pipes
mul_mat_vec_q3_k_f16_f32: Failed to create pipes
Failed to create pipes
: Failed to create pipes
mul_mat_vec_q6_k_f16_f32: Failed to create pipes
Failed to create pipes
[ 6%] Building C object ggml/src/CMakeFiles/ggml.dir/ggml-quants.c.o
[ 9%] Building C object ggml/src/CMakeFiles/ggml.dir/ggml-aarch64.c.o
[ 9%] Building CXX object ggml/src/CMakeFiles/ggml.dir/llamafile/sgemm.cpp.o
[ 9%] Building C object ggml/src/CMakeFiles/ggml.dir/ggml-backend.c.o
[ 7%] Building CXX object ggml/src/CMakeFiles/ggml.dir/ggml-vulkan-shaders.cpp.o
[ 9%] Building C object ggml/src/CMakeFiles/ggml.dir/ggml.c.o
[ 9%] Building CXX object ggml/src/CMakeFiles/ggml.dir/ggml-vulkan.cpp.o
[ 9%] Building C object ggml/src/CMakeFiles/ggml.dir/ggml-alloc.c.o
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp: In function ‘void ggml_vk_load_shaders(vk_device&)’:
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1197:89: error: ‘matmul_f32_f32_len’ was not declared in this scope; did you mean ‘matmul_f32_f16_len’?
1197 | ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->l, "matmul_f32_l", matmul_f32_f32_len, matmul_f32_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1);
| ^~~~~~~~~~~~~~~~~~
| matmul_f32_f16_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1197:109: error: ‘matmul_f32_f32_data’ was not declared in this scope; did you mean ‘matmul_f32_f16_data’?
1197 | ggml_vk_create_pipeline(device, device->pipeline_matmul_f32->l, "matmul_f32_l", matmul_f32_f32_len, matmul_f32_f32_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, 1);
| ^~~~~~~~~~~~~~~~~~~
| matmul_f32_f16_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1214:99: error: ‘matmul_f16_aligned_len’ was not declared in this scope; did you mean ‘matmul_id_f16_aligned_len’?
1214 | ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_l, "matmul_f16_aligned_l", matmul_f16_aligned_len, matmul_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~
| matmul_id_f16_aligned_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1214:123: error: ‘matmul_f16_aligned_data’ was not declared in this scope; did you mean ‘matmul_id_f16_aligned_data’?
1214 | ggml_vk_create_pipeline(device, device->pipeline_matmul_f16->a_l, "matmul_f16_aligned_l", matmul_f16_aligned_len, matmul_f16_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~
| matmul_id_f16_aligned_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1235:129: error: ‘matmul_q4_1_f32_aligned_len’ was not declared in this scope; did you mean ‘matmul_q5_1_f32_aligned_len’?
1235 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_l, "matmul_q4_1_f32_aligned_l", matmul_q4_1_f32_aligned_len, matmul_q4_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_q5_1_f32_aligned_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1235:158: error: ‘matmul_q4_1_f32_aligned_data’ was not declared in this scope; did you mean ‘matmul_q5_1_f32_aligned_data’?
1235 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat[GGML_TYPE_Q4_1]->a_l, "matmul_q4_1_f32_aligned_l", matmul_q4_1_f32_aligned_len, matmul_q4_1_f32_aligned_data, "main", 3, sizeof(vk_mat_mat_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_q5_1_f32_aligned_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1526:125: error: ‘matmul_id_q4_0_f32_fp32_len’ was not declared in this scope; did you mean ‘matmul_id_q8_0_f32_fp32_len’?
1526 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->l, "matmul_id_q4_0_f32_l", matmul_id_q4_0_f32_fp32_len, matmul_id_q4_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_id_q8_0_f32_fp32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1526:154: error: ‘matmul_id_q4_0_f32_fp32_data’ was not declared in this scope; did you mean ‘matmul_id_q8_0_f32_fp32_data’?
1526 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q4_0]->l, "matmul_id_q4_0_f32_l", matmul_id_q4_0_f32_fp32_len, matmul_id_q4_0_f32_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_id_q8_0_f32_fp32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1585:135: error: ‘matmul_id_q5_k_f32_aligned_fp32_len’ was not declared in this scope; did you mean ‘matmul_id_q6_k_f32_aligned_fp32_len’?
1585 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_l, "matmul_id_q5_k_f32_aligned_l", matmul_id_q5_k_f32_aligned_fp32_len, matmul_id_q5_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_id_q6_k_f32_aligned_fp32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1585:172: error: ‘matmul_id_q5_k_f32_aligned_fp32_data’ was not declared in this scope; did you mean ‘matmul_id_q6_k_f32_aligned_fp32_data’?
1585 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_mat_id[GGML_TYPE_Q5_K]->a_l, "matmul_id_q5_k_f32_aligned_l", matmul_id_q5_k_f32_aligned_fp32_len, matmul_id_q5_k_f32_aligned_fp32_data, "main", 4, sizeof(vk_mat_mat_id_push_constants), l_wg_denoms, warptile_mmq_l, l_align);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| matmul_id_q6_k_f32_aligned_fp32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1613:127: error: ‘mul_mat_vec_q3_k_f32_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_q6_k_f32_f32_len’?
1613 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f32_f32", mul_mat_vec_q3_k_f32_f32_len, mul_mat_vec_q3_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q6_k_f32_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1613:157: error: ‘mul_mat_vec_q3_k_f32_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_q6_k_f32_f32_data’?
1613 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f32_f32", mul_mat_vec_q3_k_f32_f32_len, mul_mat_vec_q3_k_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q6_k_f32_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1617:131: error: ‘mul_mat_vec_iq4_nl_f32_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_q4_k_f32_f32_len’?
1617 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f32_f32", mul_mat_vec_iq4_nl_f32_f32_len, mul_mat_vec_iq4_nl_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q4_k_f32_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1617:163: error: ‘mul_mat_vec_iq4_nl_f32_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_q4_k_f32_f32_data’?
1617 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f32_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f32_f32", mul_mat_vec_iq4_nl_f32_f32_len, mul_mat_vec_iq4_nl_f32_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q4_k_f32_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1627:127: error: ‘mul_mat_vec_q3_k_f16_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_q5_k_f16_f32_len’?
1627 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f16_f32", mul_mat_vec_q3_k_f16_f32_len, mul_mat_vec_q3_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q5_k_f16_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1627:157: error: ‘mul_mat_vec_q3_k_f16_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_q5_k_f16_f32_data’?
1627 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q3_K], "mul_mat_vec_q3_k_f16_f32", mul_mat_vec_q3_k_f16_f32_len, mul_mat_vec_q3_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q5_k_f16_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1630:127: error: ‘mul_mat_vec_q6_k_f16_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_q5_k_f16_f32_len’?
1630 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q6_K], "mul_mat_vec_q6_k_f16_f32", mul_mat_vec_q6_k_f16_f32_len, mul_mat_vec_q6_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q5_k_f16_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1630:157: error: ‘mul_mat_vec_q6_k_f16_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_q5_k_f16_f32_data’?
1630 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_Q6_K], "mul_mat_vec_q6_k_f16_f32", mul_mat_vec_q6_k_f16_f32_len, mul_mat_vec_q6_k_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q5_k_f16_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1631:131: error: ‘mul_mat_vec_iq4_nl_f16_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_q4_k_f16_f32_len’?
1631 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f16_f32", mul_mat_vec_iq4_nl_f16_f32_len, mul_mat_vec_iq4_nl_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q4_k_f16_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1631:163: error: ‘mul_mat_vec_iq4_nl_f16_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_q4_k_f16_f32_data’?
1631 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_f16_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_iq4_nl_f16_f32", mul_mat_vec_iq4_nl_f16_f32_len, mul_mat_vec_iq4_nl_f16_f32_data, "main", 3, sizeof(vk_mat_vec_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_q4_k_f16_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1641:125: error: ‘mul_mat_vec_id_q3_k_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q5_k_f32_len’?
1641 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q3_K], "mul_mat_vec_id_q3_k_f32", mul_mat_vec_id_q3_k_f32_len, mul_mat_vec_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q5_k_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1641:154: error: ‘mul_mat_vec_id_q3_k_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q5_k_f32_data’?
1641 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q3_K], "mul_mat_vec_id_q3_k_f32", mul_mat_vec_id_q3_k_f32_len, mul_mat_vec_id_q3_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q5_k_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1644:125: error: ‘mul_mat_vec_id_q6_k_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q5_k_f32_len’?
1644 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q6_K], "mul_mat_vec_id_q6_k_f32", mul_mat_vec_id_q6_k_f32_len, mul_mat_vec_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q5_k_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1644:154: error: ‘mul_mat_vec_id_q6_k_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q5_k_f32_data’?
1644 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_Q6_K], "mul_mat_vec_id_q6_k_f32", mul_mat_vec_id_q6_k_f32_len, mul_mat_vec_id_q6_k_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q5_k_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1645:129: error: ‘mul_mat_vec_id_iq4_nl_f32_len’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q4_k_f32_len’?
1645 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_id_iq4_nl_f32", mul_mat_vec_id_iq4_nl_f32_len, mul_mat_vec_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q4_k_f32_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1645:160: error: ‘mul_mat_vec_id_iq4_nl_f32_data’ was not declared in this scope; did you mean ‘mul_mat_vec_id_q4_k_f32_data’?
1645 | ggml_vk_create_pipeline(device, device->pipeline_dequant_mul_mat_vec_id_f32[GGML_TYPE_IQ4_NL], "mul_mat_vec_id_iq4_nl_f32", mul_mat_vec_id_iq4_nl_f32_len, mul_mat_vec_id_iq4_nl_f32_data, "main", 4, sizeof(vk_mat_vec_id_push_constants), {1, 1, 1}, { device->subgroup_size }, 1);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| mul_mat_vec_id_q4_k_f32_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1654:95: error: ‘dequant_q2_k_len’ was not declared in this scope; did you mean ‘dequant_q5_k_len’?
1654 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_Q2_K], "dequant_q2_k", dequant_q2_k_len, dequant_q2_k_data, "main", 2, 5 * sizeof(uint32_t), {256 * 64, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~
| dequant_q5_k_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1654:113: error: ‘dequant_q2_k_data’ was not declared in this scope; did you mean ‘dequant_q5_k_data’?
1654 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_Q2_K], "dequant_q2_k", dequant_q2_k_len, dequant_q2_k_data, "main", 2, 5 * sizeof(uint32_t), {256 * 64, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~
| dequant_q5_k_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1658:95: error: ‘dequant_q6_k_len’ was not declared in this scope; did you mean ‘dequant_q5_k_len’?
1658 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_Q6_K], "dequant_q6_k", dequant_q6_k_len, dequant_q6_k_data, "main", 2, 5 * sizeof(uint32_t), {256 * 64, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~
| dequant_q5_k_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1658:113: error: ‘dequant_q6_k_data’ was not declared in this scope; did you mean ‘dequant_q5_k_data’?
1658 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_Q6_K], "dequant_q6_k", dequant_q6_k_len, dequant_q6_k_data, "main", 2, 5 * sizeof(uint32_t), {256 * 64, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~
| dequant_q5_k_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1659:99: error: ‘dequant_iq4_nl_len’ was not declared in this scope; did you mean ‘dequant_q4_k_len’?
1659 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_IQ4_NL], "dequant_iq4_nl", dequant_iq4_nl_len, dequant_iq4_nl_data, "main", 2, 5 * sizeof(uint32_t), {256 * 16, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~~
| dequant_q4_k_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1659:119: error: ‘dequant_iq4_nl_data’ was not declared in this scope; did you mean ‘dequant_q4_k_data’?
1659 | ggml_vk_create_pipeline(device, device->pipeline_dequant[GGML_TYPE_IQ4_NL], "dequant_iq4_nl", dequant_iq4_nl_len, dequant_iq4_nl_data, "main", 2, 5 * sizeof(uint32_t), {256 * 16, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~~~
| dequant_q4_k_data
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1669:101: error: ‘get_rows_iq4_nl_len’ was not declared in this scope; did you mean ‘get_rows_q4_1_len’?
1669 | ggml_vk_create_pipeline(device, device->pipeline_get_rows[GGML_TYPE_IQ4_NL], "get_rows_iq4_nl", get_rows_iq4_nl_len, get_rows_iq4_nl_data, "main", 3, sizeof(vk_op_binary_push_constants), {1024, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~~~
| get_rows_q4_1_len
/home/vecna/llama.cpp/ggml/src/ggml-vulkan.cpp:1669:122: error: ‘get_rows_iq4_nl_data’ was not declared in this scope; did you mean ‘get_rows_q4_1_data’?
1669 | ggml_vk_create_pipeline(device, device->pipeline_get_rows[GGML_TYPE_IQ4_NL], "get_rows_iq4_nl", get_rows_iq4_nl_len, get_rows_iq4_nl_data, "main", 3, sizeof(vk_op_binary_push_constants), {1024, 1, 1}, {}, 1);
| ^~~~~~~~~~~~~~~~~~~~
| get_rows_q4_1_data
make[2]: *** [ggml/src/CMakeFiles/ggml.dir/build.make:201: ggml/src/CMakeFiles/ggml.dir/ggml-vulkan.cpp.o] Error 1
make[2]: *** Waiting for unfinished jobs....
make[1]: *** [CMakeFiles/Makefile2:1626: ggml/src/CMakeFiles/ggml.dir/all] Error 2
make: *** [Makefile:146: all] Error 2
Name and Version
it just don't compile and i don't found a special issue option for it
What operating system are you seeing the problem on?
Linux
Relevant log output
The text was updated successfully, but these errors were encountered: