ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-01-26 17:20:01 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
t0002
ik_llama.cpp/Makefile
firecoperana a3e78038d0 Merge vulkan code from mainline up to commit of 6/28/2025 (#563) 
* Merge vulkan code from mainline up to commit of 6/28/2025

* Vulkan Optimizations and Fixes (#8959)

* Optimize Vulkan REPEAT performance

* Use Vulkan GLSL fused multiply-add instruction where possible

* Add GGML_VULKAN_PERF option to output performance data per operator

* Rework and fix Vulkan descriptor set and descriptor pool handling

* Fix float32 concat f16 shader validation error

* Add Vulkan GROUP_NORM eps parameter

* Fix validation error with transfer queue memory barrier flags

* Remove trailing whitespaces

vulkan : do not use tensor->extra (#9407)

* vulkan : do not use tensor->extra

This patch allows using the Vulkan backend with the RPC backend as
tensor->extra is no longer used.

Ref: #8536

* Adapt GGML_VULKAN_CHECK_RESULTS to extra removal (#2)

---------

Co-authored-by: 0cc4m <picard12@live.de>
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan : fix build (#0)

ggml-ci

Improve Vulkan shader build system (#9239)

* Improve Vulkan shader builds system

- Add dependency to vulkan-shaders-gen to rebuild shaders when changing the shader compilation utility.
- Add option to generate debug info for Vulkan shaders to provide shader source to Vulkan shader profiling tools

* remove not required self dependency

ggml : fix build break for the vulkan-debug (#9265)

- windows build : Ok.
- linux build : Ok.

Signed-off-by: Changyeon Kim <cyzero.kim@samsung.com>

vulkan: correctly report support for OP_CONT (ggml/946)

test-backend-ops fails because ggml_cont aborts
when invoked passing an unsupported type.

This commit makes ggml_cont tests pass

Signed-off-by: Salvatore Mesoraca <s.mesoraca16@gmail.com>

vulkan: add dryrun support to sin and cos ops (ggml/947)

sin and cos failed test-backend-ops because they
tried to dereference a context pointer that is null
on dry runs.

This commit prevents that segfault.

Signed-off-by: Salvatore Mesoraca <s.mesoraca16@gmail.com>

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

Overlap cmdbuffer creation and cmdbuffer execution in Vulkan backend by submitting smaller cmdbuffers early. (#9118)

* Overlap cmdbuffer creation and cmdbuffer execution in Vulkan backend by submitting smaller cmdbuffers early.

* fix compile issues

* Fix issues where the last submit wasn't executed or handled properly.

* remove trailing whitespace

* Repair GGML_VULKAN_CHECK_RESULTS

* Increase submit counter only if actual work has been submitted and increase submit count to 100.

* Fix some nodes are not checked with GGML_VULKAN_CHECK_RESULTS enabled.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

Enable use to the rebar feature to upload buffers to the device. (#9251)

vulkan : argsort barriers must be under uniform control flow (ggml/951)

a return before a barrier (that happens only in some threads in
a workgroup) leads to UB.
While the old code actually works on some devices,
it fails on some others (i.e. "smaller" GPUs).

BTW, I think it would be better to set specialization constants
when the graph is built, in that way the local workgroup
could be sized appropriately.
But it would take a lot of work.

Signed-off-by: Salvatore Mesoraca <s.mesoraca16@gmail.com>

vulkan : fix build for GGML_VULKAN_RUN_TESTS, add TFLOPS to log (ggml/961)

vulkan : multithread pipeline creation (ggml/963)

vulkan : mul_mat: fix UB with small warps (ggml/952)

When the device's warp size is less than 16,
it is possible for loadstride_a (mul_mm.comp:114)
and loadstride_b (mul_mm.comp:115) to be set to 0.
Because they are calculated as: the workgroup size,
multiplied by LOAD_VEC_* (which can be 1) and divided by 16.
And the workgroup size is set to be the same as the
warp/subgroup size.

The loadstride_* variables are used as increments in the
loops that populate the buffers used for the multiplication.

When they are 0 they cause an infinite loop.
But infinite loops without side-effects are UB and the
values of loadstride_* are known at compile time.
So, the compiler quietly optimizes all the loops away.
As a consequence, the buffers are not populated and
the multiplication result is just a matrix with all elements
set to 0.

We prevent the UB by making sure that the workgroup size
will never be less than 16, even if our device has a
smaller warp size (e.g. 8).

Signed-off-by: Salvatore Mesoraca <s.mesoraca16@gmail.com>

vulkan : retry allocation with fallback flags (whisper/2451)

Co-authored-by: Samuel Morris <samuel.morris@artlist.io>

vulkan : improve ggml_vk_create_buffer error handling (#9898)

vulkan: Fix newly added tests for permuted mul_mat and 1D im2col (#10226)

vulkan: Throttle the number of shader compiles during the build step. (#10222)

Fixes #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.
# Conflicts:
#	ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp

vulkan: Optimize contiguous copies (#10254)

* tests: Fix memory bandwidth calculation for perf tests

Add a flops calculation for flash attention.

Add one GGML_OP_CPY perf test.

* vulkan: Optimize contiguous copies

Add a variant of the copy shader for when the tensors are contiguous. Avoid
the complex addressing calculations, and do four elements per invocation
to hide some other overhead.

Apply similar changes to the scale shader, since scale is always contiguous.

Add a "progress bar" for shader compiles.
# Conflicts:
#	tests/test-backend-ops.cpp

vulkan: Use macros to make the mat mul pipeline creation more concise (#10259)

Also add vk_matmul_pipeline2 to hold f16/f32 accumulator versions of a
pipeline. This isn't really used yet.

vulkan: Optimize binary ops (#10270)

Reuse the index calculations across all of src0/src1/dst. Add a shader
variant for when src0/src1 are the same dimensions and additional modulus
for src1 aren't needed. Div/mod are slow, so add "fast" div/mod that
have a fast path when the calculation isn't needed or can be done more
cheaply.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/acc.comp

ggml : vulkan logs (whisper/2547)

vulkan: Optimize some mat-vec mul quant shaders (#10296)

Compute two result elements per workgroup (for Q{4,5}_{0,1}). This reuses
the B loads across the rows and also reuses some addressing calculations.
This required manually partially unrolling the loop, since the compiler
is less willing to unroll outer loops.

Add bounds-checking on the last iteration of the loop. I think this was at
least partly broken before.

Optimize the Q4_K shader to vectorize most loads and reduce the number of
bit twiddling instructions.

Vulkan: Fix device info output format specifiers (#10366)

* Vulkan: Fix device info output format specifiers

* Vulkan: Use zu printf specifier for size_t instead of ld

vulkan: remove use of null initializer (#10372)

Seems like this isn't working for vulkan-over-metal when the array is sized
by a spec constant. Maybe a spirv-cross limitation?

vulkan: Optimize soft_max (#10301)

* vulkan: Optimize soft_max

Large soft_max could already saturate memory, but small/medium sizes were
pretty slow. The bulk of the gains for them comes from using a smaller
workgroup size, and making the workgroup size match the subgroup size also
makes the barriers much cheaper.

Cache some values in locals to avoid refetching/recomputing. And stamp
out a few "template instantiations" so smaller cases will fully unroll.

Add a missing early return for OOB rows. This happens when there are more
than 512 rows and the dispatch is 512 x H.

* vulkan: Further soft_max optimizations

Restore the workgroup size of 512 case, use it for >1024.

Use unrollable loops for more iteration counts.

vulkan: further optimize mul_mat_vec using larger loads (#10387)

* vulkan: Use pipeline_robustness to disable robustness in mul_mat_vec.

Add some early returns for nonexistent rows in mul_mat_vec shaders. These
can only be hit when dispatching a 2D grid of workgroups. Fix the logic
for the 2D grid of workgroups to round up.

Enable the pipeline robustness extension if it's available, and use it to
disable robustness for these pipelines. The instructions to do the bounds
checking contend for the same ALU resources as the bit twiddling dequant
instructions.

* vulkan: Add GLSL structure aliases for quant types to allow larger loads

In Vulkan it's not possible to cast pointer types, so instead you have to
declare an aliased binding for the memory with a different type. This
commit adds aliases for the quant formats using 16b ints, and in a few
places where the struct size is a multiple of 4 also using 32b ints.
Currently only q4_k's aliases are used, but others will be used in
subsequent commits.

* vulkan: use larger loads in q5_k and q6_k shaders.

Similar to the optimization I did in q4_k recently, this vectorizes some loads
and reduces the number of bit twiddling instructions.

* vulkan: use larger K step per iteration in mul_mat_vec.

Add vec4 dequantization functions, and use them to do K=8 per iteration in
mul_mat_vec. This uses 16b loads for the quant values and 128b loads for B
which helps reduce the load on the memory system.

The K_PER_ITER==2 logic is still there, just for F16/F32, and really only
because they support unaligned sizes.

Tweak the num_iters/unrolling logic to be simpler and catch a couple missed
unrolling opportunities.

vulkan: copy iq4_nl LUT into shared memory (#10409)

vulkan: predicate max operation in soft_max shaders/soft_max (#10437)

Fixes #10434

vulkan: Fix a vulkan-shaders-gen arugment parsing error (#10484)

The vulkan-shaders-gen was not parsing the --no-clean argument correctly.
Because the previous code was parsing the arguments which have a value only
and the --no-clean argument does not have a value, it was not being parsed
correctly. This commit can now correctly parse arguments that don't have values.

vulkan: fix group_norm (#10496)

Fix bad calculation of the end of the range. Add a backend test that
covers the bad case (taken from stable diffusion).

Fixes https://github.com/leejet/stable-diffusion.cpp/issues/439.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: optimize Q2_K and Q3_K mul_mat_vec (#10459)

vulkan: skip integer div/mod in get_offsets for batch_idx==0 (#10506)

vulkan: further optimize q5_k mul_mat_vec (#10479)

vulkan: Handle GPUs with less shared memory (#10468)

There have been reports of failure to compile on systems with <= 32KB
of shared memory (e.g. #10037). This change makes the large tile size
fall back to a smaller size if necessary, and makes mul_mat_id fall
back to CPU if there's only 16KB of shared memory.

vulkan: define all quant data structures in types.comp (#10440)

vulkan: get the first command buffer submitted sooner (#10499)

This is an incremental improvement over #9118 to get work to the GPU a bit
sooner. The first part is to start with a smaller number of nodes before
the first submit, and ramp it up to the current 100 nodes/submit. The
second part is to reduce the dryrun overhead for all the nodes that just
need to request descriptor space.

With these changes I get around 1-2% speedup on RTX 4070 combined with my
old Haswell-era CPU.

vulkan: Dynamic subgroup size support for Q6_K mat_vec (#10536)

* subgroup 64 version with subgroup add. 15% faster

scalable version

tested for subgroup sizes 16-128

* check for subgroup multiple of 16 and greater than 16

* subgroup sizes are always a power of 2 (https://github.com/KhronosGroup/GLSL/issues/45)

* force 16 sequential threads per block

* make 16 subgroup size a constant

vulkan: optimize and reenable split_k (#10637)

Use vector loads when possible in mul_mat_split_k_reduce. Use split_k
when there aren't enough workgroups to fill the shaders.

vulkan: Implement "fast divide" (mul+shift) for unary ops like copy (#10642)

vulkan: Add VK_NV_cooperative_matrix2 support for mul_mat and flash attention (#10206)

# Conflicts:
#	ggml/src/vulkan-shaders/dequant_funcs_cm2.comp
#	ggml/src/vulkan-shaders/flash_attn_cm2.comp
#	ggml/src/vulkan-shaders/mul_mm_cm2.comp

Vulkan: VK_KHR_cooperative_matrix support to speed up prompt processing (#10597)

* Vulkan: Implement VK_KHR_cooperative_matrix support in the matrix matrix multiplication shader

* Improve performance with better q4_k and q5_k dequant and store unrolling

* Add Vulkan MUL_MAT and MUL_MAT_ID accumulator precision selection

* Rework mulmat shader selection and compilation logic, avoid compiling shaders that won't get used by device

* Vulkan: Implement accumulator switch for specific mul mat mat shaders

* Vulkan: Unroll more loops for more mul mat mat performance

* Vulkan: Add VK_AMD_shader_core_properties2 support to read Compute Unit count for split_k logic

* Disable coopmat support on AMD proprietary driver

* Remove redundant checks

* Add environment variable GGML_VK_DISABLE_COOPMAT to disable VK_KHR_cooperative_matrix support

* Fix rebase typo

* Fix coopmat2 MUL_MAT_ID pipeline selection
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: compile a test shader in cmake to check for coopmat2 support (#10713)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/ggml-vulkan/CMakeLists.txt
#	ggml/src/vulkan-shaders/test_coopmat2_support.comp

Vulkan: fix NaN in tanh.comp with AMD proprietary driver on Windows (#10723)

* Vulkan: fix NaN in tanh.comp

* Faster NaN-free tanh

vulkan: fix compile warnings (#10731)

vulkan: disable spirv-opt for coopmat shaders (#10763)

There are some bugs in the 1.3.296 SDK, so disable this. It isn't strictly
necessary anyway.

Add missing dependency on vulkan-shaders-gen, so shaders get recompiled when it
changes.

Fix coopmat support reporting when glslc doesn't support NV_coopmat2.

vulkan: dynamic subgroup size for the remaining k quants (#10745)

* q5_k

q4_k

q3_k

q2_k

q6_k multi row example

* revert as multi row isnt faster for k quants

vulkan: request round-to-even for fp16 in im2col/rope_head (#10767)

Vulkan doesn't mandate a specific rounding mode, but the shader_float_controls
feature allows rounding mode to be requested if the implementation supports it.

Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats (#10721)

* Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats

* Fix subgroup size control extension support check

Add accf32 and accf16 checks for coopmats

* Also disable coopmats on amdvlk

Vulkan: Use improved q4_k and q5_k dequant code in dequant shaders (#10798)

vulkan: small mul_mat_vec optimizations (#10665)

* double the number of rows per workgroup

* Update ggml-vulkan.cpp

* Vulkan: Add VK_EXT_subgroup_size_control support to ensure full subgroups for coopmats

* only increase the number of rows for amd and subgroup size 64

* fix missing NUM_ROWS for mul_mat_vec_iq4_nl_f16_f32, untested

* use subgroup min and max to check for gcn (requires https://github.com/ggerganov/llama.cpp/pull/10721)

* manual merge ggml-vulkan.cpp

* set min and max subgroup size in any case

* Also double the number of rows for Intel GPUs

Change Debug print name

add GGML_ROPE_TYPE_MROPE

rwkv6: add wkv6 support for Vulkan backend (#10829)

* rwkv_wkv6 vulkan shader

* RWKV_WKV6 Vulkan op tests passed

Signed-off-by: Molly Sophia <mollysophia379@gmail.com>

* Apply code format changes

Signed-off-by: Molly Sophia <mollysophia379@gmail.com>

* add [[unroll]] and remove unnecessary conditions

* add uma support

* fix erros in EditorConfig Checker

---------

Signed-off-by: Molly Sophia <mollysophia379@gmail.com>
Co-authored-by: Molly Sophia <mollysophia379@gmail.com>
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/wkv6.comp

vulkan: bugfixes for small subgroup size systems + llvmpipe test (#10809)

* ensure mul mat shaders work on systems with subgroup size less than 32

more fixes

add test

* only s_warptile_mmq needs to be run with 32 threads or more
# Conflicts:
#	.github/workflows/build.yml

vulkan : fix soft_max.comp division by zero (whisper/2633)

This change prevents a division by zero error when p.KY is 0.

vulkan: optimize coopmat2 dequant functions (#10855)

Change the code to do 16b loads when possible and extract the appropriate
component late, so the code is effectively decoding a pair of elements and
then selecting one. This can allow more commoning to happen in the compiler
when neighboring elements are loaded.

vulkan: build fixes for 32b (#10927)

* vulkan: build fixes for 32b

Should fix #10923

* vulkan: initialize some buffer/offset variables

examples, ggml : fix GCC compiler warnings (#10983)

Warning types fixed (observed under MSYS2 GCC 14.2.0):
* format '%ld' expects argument of type 'long int', but argument has type 'size_t'
* llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp:81:46: warning: missing initializer for member '_STARTUPINFOA::lpDesktop' [-Wmissing-field-initializers]  (emitted for all struct field except first)
# Conflicts:
#	examples/export-lora/export-lora.cpp

vulkan: multi-row k quants (#10846)

* multi row k quant shaders!

* better row selection

* more row choices

* readjust row selection

* rm_kq=2 by default

vulkan: Use push constant offset to handle misaligned descriptors (#10987)

vulkan: im2col and matmul optimizations for stable diffusion (#10942)

* tests: Add im2col perf tests

* vulkan: optimize im2col, more elements per thread

* vulkan: increase small tile size for NV_coopmat2

* vulkan: change im2col to 512 elements per workgroup

vulkan: optimize mul_mat for small values of N (#10991)

Make the mul_mat_vec shaders support N>1 (as a spec constant, NUM_COLS) where
the batch_strides are overloaded to hold the row strides. Put the loads from the
B matrix in the innermost loop because it should cache better.

Share some code for reducing the result values to memory in mul_mat_vec_base.
# Conflicts:
#	tests/test-backend-ops.cpp

fix: Vulkan shader gen binary path (#11037)

Vulkan: Add device-specific blacklist for coopmat for the AMD proprietary driver (#11074)

* Vulkan: Add device-specific blacklist for coopmat for the AMD proprietary driver

* Add (TM) to AMD name check

fix lora print

Disable GL_KHR_cooperative_matrix Vulkan extension if not available. (#11117)

* Disable GL_KHR_cooperative_matrix Vulkan extension if not available.

* Perform Vulkan extensions checks in a more sensible order

* Remove unnecessary #ifdef directive
# Conflicts:
#	ggml/src/vulkan-shaders/test_coopmat_support.comp

llama: add support for QRWKV6 model architecture (#11001)

Vulkan: Fix float16 use on devices without float16 support + fix subgroup_size_control validation error (#11161)

* Vulkan: Remove float16 use in shaders

* Fix validation error about subgroup_size_control extension

fix: ggml: fix vulkan-shaders-gen build (#10448)

* fix: ggml: fix vulkan-shaders-gen build

The vulkan-shaders-gen target was not being built correctly
in case of cross-compilation.
Other outputs need to be built for the cross compile target,
but vulkan-shaders-gen needs to be built for the host.

* refactor: ggml: Improve vulkan-shaders-gen toolchain setup

- Add GGML_SHADERS_GEN_TOOLCHAIN CMake option.
- Auto-detect host toolchain if not set.

* refactor: ggml: Improve vulkan-shaders-gen toolchain setup

Use configure_file to generate host_toolchain.cmake from template

* fix: ggml: Fix compile error

Fix compile error not finding vulkan-shaders-gen

* fix: vulkan-shaders-gen build and path handling

Fix build issues with vulkan-shaders-gen:
- Add target dependency for correct build order
- Use CMAKE_HOST_SYSTEM_NAME for executable suffix
- Fix MSVC output directory in host toolchain
- Normalize path handling for cross-compilation

* fix: improve host compiler detection in vulkan shader build

Improve host compiler detection for vulkan shader generation:
- Add NO_CMAKE_FIND_ROOT_PATH to all compiler searches
- Consolidate compiler detection logic
- Fix Windows-specific MSVC detection
- Ensure correct compiler search in cross-compilation

* refactor: Simplify CMake function for detecting host compiler

Simplified the CMake function to improve the process of detecting the host compiler.

* fix: Remove unnecessary Vulkan library linkage in CMakeLists.txt

Since `vulkan-shader-gen.cpp` only requires the `glslc` executable
and not the Vulkan headers or libraries, CMakeLists.txt needs to
be corrected.
(See: ecc93d0558fc3ecb8a5af69d2ece02fae4710ade)

* refactor: Rename host_toolchain.cmake.in

- Rename host_toolchain.cmake.in to cmake/host-toolchain.cmake.in

* refactor: GGML_VULKAN_SHADERS_GEN_TOOLCHAIN

Rename the macro GGML_SHADERS_GEN_TOOLCHAIN to GGML_VULKAN_SHADERS_GEN_TOOLCHAIN
# Conflicts:
#	ggml/src/ggml-vulkan/CMakeLists.txt

vulkan: scale caching for k quants + misc fixes (#11081)

* q6_k scale caching

* 16 bit unpack

* q4_k test (slow)

* revert it

* q3_k

* q2_k

* little stuff

* try precalculating products of a and q2_k scales

* Revert "try precalculating products of a and q2_k scales"

This reverts commit 65110b81f23f66331a50c6e889a7c1ab9470a86b.

* unpack should be u16, add vim swap to gitignore (about time)

* better q4_k scales

* q5_k

* better q6_k with separate paths for all threads and partial threads in use, plus some more optimizations

* q2_k better dequant

* q3_k optimizations

* q3_k use hmask simd from cpu avx version

* make the caches happy

* q3_k separate out calculation

* q2_k separate out

* little stuff

* use calc_superblock everywhere

* q2_k optimize scale calculation

* more barriers

vulkan: optimize coopmat2 q2_k dequant function (#11130)

vulkan: optimize coopmat2 q4_k/q5_k dequant functions. (#11206)

Do masking on whole dwords, fetch all scales at once.

vulkan: support copy from f32 to q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl (#11166)

* vulkan: support copy from f32 to q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl

Shaders are based on cpy.cu.

* vulkan: support copy from q4_0/q4_1/q5_0/q5_1/q8_0/iq4_nl to f32

* ggml: copy q->f32 assumes some contiguity in the destination
# Conflicts:
#	ggml/src/ggml-cpu/ggml-cpu.c
#	ggml/src/vulkan-shaders/copy_from_quant.comp
#	ggml/src/vulkan-shaders/copy_to_quant.comp

vulkan: fix coopmat2 flash attention for non-contiguous inputs (#11281)

Add code similar to mul_mm_cm2 to force alignment of strides, to avoid
a performance regression.

Add noncontiguous FA tests in test-backend-ops.

Fixes #11268.
# Conflicts:
#	tests/test-backend-ops.cpp

vulkan: fix coopmat2 validation failures (#11284)

mul mat and flash attention shaders were loading f32 types directly into
A/B matrices, which happens to work but is technically invalid usage.
For FA, we can load it as an Accumulator matrix and convert and this
is not in the inner loop and is cheap enough. For mul mat, it's more
efficient to do this conversion in a separate pass and have the input(s)
be f16.

coopmat2 requires SPIR-V 1.6 (related using to LocalSizeId). LocalSizeId
requires maintenance4 be enabled, and SPIR-V 1.6 requires Vulkan 1.3.

vulkan: fix diag_mask_inf (#11323)

With robustbufferaccess disabled, this shader was showing OOB stores. There
is a bounds check in the code, but the workgrouop dimensions were reversed vs
CUDA and it was running the wrong number of threads. So fix the workgroup
dimensions and disable robustness for this pipeline.

vulkan: sort shaders for more deterministic binary (#11315)

Fixes #11306.

Vulkan-run-test: fix mmq_wg_denoms (#11343)

There should be a copy-and-paste error here.

*mmq_wg_denoms should be used together with *warptile_mmq, instead of
wg_denoms.

vulkan: compile shaders on-demand (#11406)

Reduce first-run startup time and memory consumption.

Should fix #11339.

vulkan: Catch pipeline creation failure and print an error message (#11436)

* vulkan: Catch pipeline creation failure and print an error message

Also, fix some warnings from my on-demand compile change.

* vulkan: fix pipeline creation logging

vulkan: implement initial support for IQ2 and IQ3 quantizations (#11360)

* vulkan: initial support for IQ3_S

* vulkan: initial support for IQ3_XXS

* vulkan: initial support for IQ2_XXS

* vulkan: initial support for IQ2_XS

* vulkan: optimize Q3_K by removing branches

* vulkan: implement dequantize variants for coopmat2

* vulkan: initial support for IQ2_S

* vulkan: vertically realign code

* port failing dequant callbacks from mul_mm

* Fix array length mismatches

* vulkan: avoid using workgroup size before it is referenced

* tests: increase timeout for Vulkan llvmpipe backend

---------

Co-authored-by: Jeff Bolz <jbolz@nvidia.com>
# Conflicts:
#	ggml/src/vulkan-shaders/dequant_iq2_s.comp
#	ggml/src/vulkan-shaders/dequant_iq2_xs.comp
#	ggml/src/vulkan-shaders/dequant_iq2_xxs.comp
#	ggml/src/vulkan-shaders/dequant_iq3_s.comp
#	ggml/src/vulkan-shaders/dequant_iq3_xxs.comp

CUDA: non-contiguous (RMS) norm support (#11659)

vulkan: use smaller combined allocations to avoid fragmentation (#11551)

# Conflicts:
#	ggml/src/ggml-alloc.c

vulkan: initial support for IQ4_XS quantization (#11501)

# Conflicts:
#	ggml/src/vulkan-shaders/dequant_iq4_xs.comp

vulkan: optimize coopmat2 iq2/iq3 callbacks (#11521)

* vulkan: optimize coopmat2 iq2/iq3 callbacks

* build: trigger CI on GLSL compute shader changes

vulkan: print shared memory size (#11719)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: account for lookup tables when checking shared memory size (#11502)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: add environment variable GGML_VK_PREFER_HOST_MEMORY to avoid VRAM allocation (#11592)

vulkan: linux builds + small subgroup size fixes (#11767)

* mm subgroup size

* upload vulkan x86 builds

vulkan: initial support for IQ1_S and IQ1_M quantizations (#11528)

* vulkan: initial support for IQ1_S and IQ1_M quantizations

* vulkan: define MMV kernels for IQ1 quantizations

* devops: increase timeout of Vulkan tests again

* vulkan: simplify ifdef for init_iq_shmem
# Conflicts:
#	ggml/src/vulkan-shaders/dequant_iq1_m.comp
#	ggml/src/vulkan-shaders/dequant_iq1_s.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq1_m.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq1_s.comp

vulkan: support multi/vision rope, and noncontiguous rope (#11902)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/rope_multi.comp
#	ggml/src/vulkan-shaders/rope_vision.comp

vulkan: implement several ops relevant for ggml_opt (#11769)

* vulkan: support memset_tensor

* vulkan: support GGML_OP_SUM

* vulkan: implement GGML_OP_ARGMAX

* vulkan: implement GGML_OP_SUB

* vulkan: implement GGML_OP_COUNT_EQUAL

* vulkan: implement GGML_OP_OPT_STEP_ADAMW

* vulkan: fix check_results RWKV_WKV6 crash and memory leaks

* vulkan: implement GGML_OP_REPEAT_BACK

* tests: remove invalid test-backend-ops REPEAT_BACK tests

* vulkan: fix COUNT_EQUAL memset using a fillBuffer command
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/argmax.comp
#	ggml/src/vulkan-shaders/count_equal.comp
#	ggml/src/vulkan-shaders/opt_step_adamw.comp
#	ggml/src/vulkan-shaders/repeat_back.comp
#	ggml/src/vulkan-shaders/sub.comp
#	tests/test-backend-ops.cpp

vulkan: implement more backpropagation operators (#11914)

* vulkan: implement GGML_OP_ROPE_BACK

* vulkan: implement GGML_OP_RMS_NORM_BACK

* vulkan: implement GGML_OP_SILU_BACK

* vulkan: implement GGML_OP_SOFTMAX_BACK
# Conflicts:
#	ggml/src/vulkan-shaders/rms_norm_back.comp
#	ggml/src/vulkan-shaders/silu_back.comp
#	ggml/src/vulkan-shaders/soft_max_back.comp

Add memset tensor in all backend interface

SYCL: implement memset ggml backend buffer interface (#12580)

* SYCL: implement memset ggml backend buffer interface

* use GGML_ABORT macro

* Do not wait for all queues to finish for memset operation
# Conflicts:
#	ggml/src/ggml-sycl.cpp

add OP sigmoid (#12056)

Co-authored-by: Judd <foldl@boxvest.com>
# Conflicts:
#	ggml/src/vulkan-shaders/sigmoid.comp

vulkan: fix assertion when qy_needs_dequant (#12068)

Looks like a copy/paste bug from qx_needs_dequant.

vulkan: improve im2col (#11826)

* vulkan: improve im2col performance

vulkan: matmul dequantization improvements (#12015)

* faster dequant for old quants

* dont use unpack for iq4_nl

* vec2 unpack for q8

vulkan: add specific MMV kernels for IQ2 and IQ3 quants + optimizations (#11595)

* vulkan: implement specialized MMV kernels for IQ2 quantizations

* vulkan: add MMV kernels for IQ3 quants

* vulkan: Increase MMV batch size and unroll IQ LUT setup

* vulkan: fix init_iq_shmem for WG sizes larger than tables

* vulkan: common batch size for all I-quants
# Conflicts:
#	ggml/src/vulkan-shaders/mul_mat_vec_iq2_s.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq2_xs.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq2_xxs.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq3_s.comp
#	ggml/src/vulkan-shaders/mul_mat_vec_iq3_xxs.comp

cuda/vulkan: specify fp32-only support for some operations in supports_op (ggml/1129)

ggml-ci

# Conflicts:
#	ggml/src/ggml-cuda.cu
#	tests/test-backend-ops.cpp

mat vec double buffer (#12188)

vulkan: fix bug in coopmat1 mul_mat_id (#12316)

* tests: run mul_mat_id with a larger N

* vulkan: fix bug in coopmat1 mul_mat_id

Update build.yml for Windows Vulkan builder to use Vulkan 1.4.304 SDK for VK_NV_cooperative_matrix2 support (#12301)

vulkan: Adjust coopmat2 tile sizes and selection heuristic (#12258)

vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bounds checking (#12273)

* vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bounds checking

vulkan: use fp32 in coopmat2 q4_k dequant function (#12309)

vulkan: subgroup size tuning (#12087)

* vulkan: subgroup size test

* Vulkan: Add device architecture enum and logic to recognize AMD generations

* vulkan: use new architecture logic to specify subgroup size

* Initial vulkan subgroup size tuning for RDNA3

* vulkan: commonize RDNA subgroup tuning

* vulkan: override subgroup size if required_subgroup_size = 0

* vulkan: disable warp 32 for RDNA3

* vulkan: fine tuned RDNA1 subgroup sizes

* vulkan: adjusted subgroup size map

* vulkan: fixed RDNA2 subgroup map

---------

Co-authored-by: 0cc4m <picard12@live.de>

vulkan: Add N/2 and N/4 optimized paths in coopmat2 shader (#12312)

ggml-vulkan: remove unused find_program(glslc) (#12416)

It's already found by FindVulkan.cmake in the parent CMakeLists

Vulkan: Default to 1GB allocations instead of 4GB to avoid fragmentation and driver issues (#12434)

vulkan: Submit once enough matmul work has been recorded (#12406)

I've been seeing significantly worse performance for tg with flash attention
enabled vs disabled, and it seems to be related to the submit heuristic.
Change the heuristic to check how many bytes worth of weight matrix are
used and flush every 100MB, and ramp up after the first few submits.
This seems to resolve the issue, and also increases perf for non-FA a bit.

vulkan: optimize iq1 coopmat2 dequant functions (#12427)

vulkan: workaround for AMD Windows driver 16 bit unpack8 bug (#12472)

Vulkan: RTE rounding for cpy to quant (#12480)

* Vulkan: RTE rounding for cpy to quant

Co-Authored-By: Jeff Bolz <jbolz@nvidia.com>

* remove trailing whitespace

* avoid duplicating pipeline_cpy_f32_quant

* fix copypasting issue

* remove duplicated code

---------

Co-authored-by: Jeff Bolz <jbolz@nvidia.com>

vulkan: Optimize mul_mat_vec p021 and nc shaders (#12505)

* tests: add mul_mat perf/functional tests for p021/nc vulkan shaders

* vulkan: Optimize mul_mat_vec p021 and nc shaders.

These shaders are used in attention calculations, and when the KV cache grows
large they start to dominate the run time. For the nc shader (which is called
with large 'k' dimension), use unrolling and vector loads. For the p021 shader
(which is called with large 'm' and small 'k' dimensions), take advantage of
grouped query attention to reuse loads from the A matrix for the whole group,
and reduce the number of workgroups (too much overhead from tiny dispatches).

Using subgroupAdd in the p021 shader also helps, use that conditionally.
# Conflicts:
#	tests/test-backend-ops.cpp

vulkan: fix mul_mat_vec failure in backend tests (#12529)

The OOB calculation could be wrong if the last iteration was during one of
the unrolled loops. Adjust the unrolling counts to avoid this. Add a couple
new backend tests that hit this failure on NVIDIA GPUs.

vulkan: fix coopmat shader generation when cross-compiling (#12272)

* vulkan: fix coopmat shader generation when cross-compiling

Previously the status of coopmat{,2} support isn't passed to the
vulkan-shaders-gen project building on the host, which leads to build
failure because of the cross-compiling code expecting coopmat{,2}
shaders that didn't get generated.

Fix this by passing the coopmat{,2} support status to vulkan-shaders
subproject.

Signed-off-by: Icenowy Zheng <uwu@icenowy.me>

* Only call coop-mat shaders once

* Fix whitespace

---------

Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
Co-authored-by: bandoti <141645996+bandoti@users.noreply.github.com>

cmake: improve Vulkan cooperative matrix support checks (whisper/2966)

Co-authored-by: Sandro Hanea <me@sandro.rocks>

cmake : fix whitespace (#0)

Vulkan: Add DP4A MMQ and Q8_1 quantization shader (#12135)

* Vulkan: Add DP4A MMQ and Q8_1 quantization shader

* Add q4_0 x q8_1 matrix matrix multiplication support

* Vulkan: Add int8 coopmat MMQ support

* Vulkan: Add q4_1, q5_0 and q5_1 quants, improve integer dot code

* Add GL_EXT_integer_dot_product check

* Remove ggml changes, fix mmq pipeline picker

* Remove ggml changes, restore Intel coopmat behaviour

* Fix glsl compile attempt when integer vec dot is not supported

* Remove redundant code, use non-saturating integer dot, enable all matmul sizes for mmq

* Remove redundant comment

* Fix integer dot check

* Fix compile issue with unsupported int dot glslc

* Update Windows build Vulkan SDK version
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/mul_mmq.comp
#	ggml/src/vulkan-shaders/mul_mmq_funcs.comp
#	ggml/src/vulkan-shaders/quantize_q8_1.comp
#	ggml/src/vulkan-shaders/test_integer_dot_support.comp

vulkan: fix build when glslc doesn't support coopmat (#12683)

Vulkan: Fix mmq int dot float cache size (#12722)

vulkan: Implement grouped query attention in the coopmat2 FA shader (#12559)

When adjacent batches of Q share the same batches of K/V, batch them into
the same workgroup. For example, when:

dst(128,32,1,1) = FA(q(128,1,32,1), k(128,16640,8,1), v(128,16640,8,1))

previously we would run 32 workgroups computing 1 result each, now we will
run 8 workgroups computing 4 results each.

This doesn't directly translate to better performance (at least when you have
>=32 SMs), but in a subsequent change I'll enable split_k which will scale much
better with 4x fewer workgroups.

cmake: remove caching from vulkan coopmat checks (#12719)

vulkan: Implement split_k for coopmat2 flash attention. (#12627)

When using group query attention, we have one workgroup per KV batch and this
can be very few workgroups (e.g. just 8 in some models). Enable split_k to
spread the work across SMs. This helps a lot when the KV cache is large.
# Conflicts:
#	ggml/src/vulkan-shaders/flash_attn_split_k_reduce.comp

vulkan: Fix missing cmake logic for dot product extension (#12721)

vulkan: set cmake minimum and project name in vulkan-shaders (#12744)

vulkan: Hybrid waitForFences/getFenceStatus to reduce fence latency (#12630)

There seems to be a bubble waking up from waitForFences, which costs a few
percent performance and also increased variance in performance. This change
inserts an "almost_ready" fence when the graph is about 80% complete and we
waitForFences for the almost_ready fence and then spin (with _mm_pauses) waiting
for the final fence to be signaled.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

cmake: fix ggml-shaders-gen compiler paths containing spaces (#12747)

fixes error for compiler paths with spaces

Vulkan: Tune Vulkan mmq int dot shader for performance (#12767)

vulkan: Use unclamped loads for flash attention mask (#12720)

nem1 must be a multiple of GGML_KQ_MASK_PAD, and GGML_KQ_MASK_PAD is a multiple
of the number of rows in the matrix. The KV dim is a multiple of the number of
columns for the aligned shader.

vulkan: fix NaN issue in flash attention shader (#12776)

Use -FLT_MAX/2 rather than -inf as the initial value for computing the maximum.

vulkan: Use fp16 for the flash attention P*V multiplication (#12783)

This is consistent with the ggml-cuda behavior and the mul_mat fallback.

vulkan: In coopmat2 mmq, load q4_k/q5_k scales through shared memory (#12833)

q4_k and q5_k had a lot of redundant global loads where the same 16B of
scale information is repeatedly loaded and decoded during each loop iteration.
This change restructures the loops to more explicitly iterate over whole
blocks in the outer loop (with unrolled inner loop) and to copy/decode the
scale data into shared memory once at the start of each outer loop. The copy
is pipelined so the scale load from global memory is relatively cheap.

This improves q4_k/q5_k model prompt processing performance by around 5-7%.
I briefly tried applying this to q6_k and q4_0, and it didn't help for q6_k
and hurt for q4_0.

The big "else" path in mul_mm_cm2.comp that had all the clamped/unclamped
variants isn't used as often as it originally was (e.g. due to the padded_N
change), so I trimmed it down to offset some of the new complexity of the
semi-manual loop unrolling.

vulkan: use aligned loads for flash attention mask (#12853)

Rewrite the stride logic for the mask tensor in the FA shader to force the
stride to be aligned, to allow using more efficient loads.

vulkan: enable coopmat2 FA gqa and split_k optimizations more often (#12931)

The grouped query attention optmization doesn't require a power of two ratio,
the only thing relying on it was the modulo operation written as bitwise &.

split_k need not depend on gqa_ratio - enable it any time there's only one
workgroup in the X dimension. The shader gets the split index from the x coord,
and multiple workgroups in the X dimension (pre-split) indicates a larger
FA operation that wouldn't need splitting.

vulkan: support noncontiguous rms_norm (#13031)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: matmul gcn tuning (#13016)

* tune matmul for gcn

* this one is more power efficient

* Update ggml/src/ggml-vulkan/ggml-vulkan.cpp

Co-authored-by: 0cc4m <picard12@live.de>

* disable this tune for the proprietary driver

---------

Co-authored-by: 0cc4m <picard12@live.de>

vulkan: use uint array index to avoid glslang bug (#13193)

vulkan: Handle src1 batch dimension in non-contiguous mat-vec-mul shader (#13191)

* vulkan: Handle src1 batch dimension in non-contiguous mat-vec-mul shader

vulkan: Add bfloat16 support (#12554)

* vulkan: Add bfloat16 support

This adds bfloat16 matrix multiply support based on VK_KHR_shader_bfloat16.
The extension is required for coopmat multiply support, but matrix-vector
multiply trivially promotes bf16 to fp32 and doesn't require the extension.
The copy/get_rows shaders also don't require the extension.

It's probably possible to fall back to non-coopmat and promote to fp32 when
the extension isn't supported, but this change doesn't do that.

The coopmat support also requires a glslc that supports the extension, which
currently requires a custom build.

* vulkan: Support bf16 tensors without the bf16 extension or coopmat support

Compile a variant of the scalar mul_mm shader that will promote the bf16
values to float, and use that when either the bf16 extension or the coopmat
extensions aren't available.

* vulkan: bfloat16 fixes (really works without bfloat16 support now)

* vulkan: fix spirv-val failure and reenable -O
# Conflicts:
#	ggml/src/vulkan-shaders/test_bfloat16_support.comp

vulkan: Additional type support for unary, binary, and copy (#13266)

Support f16->f32 copy.
Support f16->f16 and f32->f32 unary ops.
Support all combinations of f16/f32 for src0/src1/dst for add/sub/mul/div.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: Allow up to 4096 elements for mul_mat_id row_ids (#13326)

This assert fired running Qwen_Qwen3-30B-A3B-Q2_K.gguf:

GGML_ASSERT(nei0 * nei1 <= 3072);

The tensor is 8 x 512. Increase this array size to accommodate.

vulkan: scalar flash attention implementation (#13324)

* vulkan: scalar flash attention implementation

* vulkan: always use fp32 for scalar flash attention

* vulkan: use vector loads in scalar flash attention shader

* vulkan: remove PV matrix, helps with register usage

* vulkan: reduce register usage in scalar FA, but perf may be slightly worse

* vulkan: load each Q value once. optimize O reduction. more tuning

* vulkan: support q4_0/q8_0 KV in scalar FA

* CI: increase timeout to accommodate newly-supported tests

* vulkan: for scalar FA, select between 1 and 8 rows

* vulkan: avoid using Float16 capability in scalar FA
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/flash_attn.comp

vulkan: workaround FA compile failures on macos (#13517)

vulkan: KHR_coopmat flash attention (#13506)

This shader uses coopmat1 to do the Q*K^T multiply. The P*V multiply is more
difficult for various reasons so I haven't done it. Performance for this
shader is around 2.5x better than for the scalar shader when doing prompt
processing. Some of the benefit may be from other optimizations like staging
through shared memory, or splitting by rows.
# Conflicts:
#	ggml/src/vulkan-shaders/flash_attn_cm1.comp

cmake: simplify vulkan shader test logic (#13263)

vulkan: use scalar FA rather than coopmat2 when N==1 (#13554)

Add pipeline_acc_f32

vulkan: move common FA code to flash_attn_base.comp (#13556)

* vulkan: move common FA code to flash_attn_base.comp

* vulkan: move common FA index/stride setup code to flash_attn_base.comp

* build fix
# Conflicts:
#	ggml/src/vulkan-shaders/flash_attn_base.comp

cmake: use the current build config for vulkan-shaders-gen (#13595)

* fix: use the current build config for `vulkan-shaders-gen`

* fix: only pass a valid build type to `--config`

Vulkan: Add f32 accumulator support to quantized mul mat to fix GLM4 32B incoherence (#13607)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: fix warnings (#13626)

* small fixes

* remove ifdef

use LOG_WARN to replace `std::cerr` (#13657)

vulkan: Disable coopmat/coopmat2/bfloat extensions if glslc doesn't support it (#13696)

vulkan: support CPY from any type to itself (#13695)

Reuse the f16/f32 copy shaders, and just scale the number of elements
according to the type size.

add GGML_LOG_WARN

vulkan: mark IM2COL as supporting non-contig (#13783)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: use timestamp queries for GGML_VULKAN_PERF (#13817)

Also change it to be controlled by an env var rather than cmake flag

vulkan : Remove unexpected ; (ggml/1253)

vulkan: fix warnings in perf logger querypool code (#13937)

ggml-vulkan: adds support for op CONV_TRANSPOSE_1D (#13813)

* * ggml-vulkan: adds op CONV_TRANSPOSE_1D

* test-backend-ops: adds more spohisticated tests for CONV_TRANSPOSE_1D

* Missing barrier added to shader.
Number of additional tests reduced to 108.

* * Fixes typo in variable name.

* Removes extra whitespaces.

* Adds int64->int32 casts to prevent possible warnings.

* Problem size reduced in tests to pass tests with llvmpipe.

* supports_op condition moved from unintended position
# Conflicts:
#	ggml/src/ggml-vulkan.cpp
#	ggml/src/vulkan-shaders/conv_transpose_1d.comp

vulkan: Enable VK_KHR_cooperative_matrix extension for Intel Xe2 GPUs (#14001)

* allowing B580 and U9-288V

* experimenting code to detect Xe2

* allowing coopmat only for Xe2 GPUs

* fixed comment wording

* fixed comment wording

* removed unnecessary driver check

Vulkan: Don't default to CPU device (like llvmpipe), even if no other device is available, to allow fallback to CPU backend (#14099)

# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: force device 0 in CI (#14106)

Add GGML_LOG_INFO

vulkan: Track descriptor pools/sets per-context (#14109)

Use the same descriptor set layout for all pipelines (MAX_PARAMETER_COUNT == 8)
and move it to the vk_device. Move all the descriptor pool and set tracking to
the context - none of it is specific to pipelines anymore. It has a single vector
of pools and vector of sets, and a single counter to track requests and a single
counter to track use.

vulkan: Better thread-safety for command pools/buffers (#14116)

This change moves the command pool/buffer tracking into a vk_command_pool
structure. There are two instances per context (for compute+transfer) and
two instances per device for operations that don't go through a context.
This should prevent separate contexts from stomping on each other.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: mutex around vkQueueSubmit (#14127)

This fixes the remaining crash in test-thread-safety on my system.

cmake: clean up external project logic for vulkan-shaders-gen (#14179)

* Remove install step for vulkan-shaders-gen

* Add install step to normalize msvc with make

* Regenerate modified shaders at build-time
# Conflicts:
#	.github/workflows/build.yml

cmake: remove shader-gen step-targets from ggml-vulkan (#14226)

* Remove step-targets from vulkan-shaders-gen

* Unset DESTDIR when building vulkan-shaders-gen

Vulkan: Set device max size for host memory to avoid OOM warning and fallback to CPU buffer (#14249)

Add support for VK_EXT_debug_utils to add labels to Vulkan objects. (#13792)

* Add support for VK_EXT_debug_utils to add labels to Vulkan objects. In step 1 compute pipelines are getting labeled.

* remove #ifdef for debug utils and add queue marker.
# Conflicts:
#	ggml/src/ggml-vulkan.cpp

vulkan: update windows SDK in CI (#14334)

vulkan: update windows SDK in release.yml (#14344)

# Conflicts:
#	.github/workflows/release.yml

cmake: regen vulkan shaders when shaders-gen sources change (#14398)

* Add shaders-gen sources as target deps

vulkan: Fix GGML_VULKAN_SHADER_DEBUG_INFO (#14427)

This setting needs to be passed through to vulkan-shaders-gen

vulkan: lock accesses of pinned_memory vector (#14333)

vulkan: handle noncontig in the final case of ggml_vk_get_cpy_pipeline (#14378)

Fix cuda build error

test

* remove  new cpu backend and yml files

* remove new op and GGML_ROPE_TYPE_NEOX

* fix build error

* change cmake file to add matrix operation

* remove coopmat2 check in flash attention

* print gpu info for vulkan

* disable fuse to recover vulkan performance

---------

Co-authored-by: 0cc4m <picard12@live.de>
Co-authored-by: firecoperana <firecoperana>
2025-07-02 08:49:42 +02:00

1686 lines
50 KiB
Makefile
Raw Permalink Blame History

# Define the default target now so that it is always the first target
BUILD_TARGETS = \
libllava.a \
llama-baby-llama \
llama-batched \
llama-batched-bench \
llama-bench \
llama-benchmark-matmult \
llama-cli \
llama-convert-llama2c-to-ggml \
llama-embedding \
llama-eval-callback \
llama-export-lora \
llama-gbnf-validator \
llama-gguf \
llama-gguf-hash \
llama-gguf-split \
llama-gritlm \
llama-imatrix \
llama-infill \
llama-llava-cli \
llama-minicpmv-cli\
llama-lookahead \
llama-lookup \
llama-lookup-create \
llama-lookup-merge \
llama-lookup-stats \
llama-parallel \
llama-passkey \
llama-perplexity \
llama-q8dot \
llama-quantize \
llama-quantize-stats \
llama-retrieval \
llama-save-load-state \
llama-server \
llama-simple \
llama-speculative \
llama-tokenize \
llama-vdot \
llama-cvector-generator \
tests/test-c.o
# Binaries only useful for tests
TEST_TARGETS = \
tests/test-autorelease \
tests/test-backend-ops \
tests/test-chat-template \
tests/test-double-float \
tests/test-grad0 \
tests/test-grammar-integration \
tests/test-grammar-parser \
tests/test-json-schema-to-grammar \
tests/test-llama-grammar \
tests/test-model-load-cancel \
tests/test-opt \
tests/test-quantize-fns \
tests/test-quantize-perf \
tests/test-rope \
tests/test-sampling \
tests/test-tokenizer-0 \
tests/test-tokenizer-1-bpe \
tests/test-tokenizer-1-spm
# Legacy build targets that were renamed in #7809, but should still be removed when the project is cleaned
LEGACY_TARGETS_CLEAN = main quantize quantize-stats perplexity imatrix embedding vdot q8dot convert-llama2c-to-ggml \
simple batched batched-bench save-load-state server gguf gguf-split eval-callback llama-bench libllava.a llava-cli baby-llama \
retrieval speculative infill tokenize benchmark-matmult parallel export-lora lookahead lookup passkey gritlm
# Legacy build targets that were renamed in #7809, but we want to build binaries that for them that output a deprecation warning if people try to use them.
# We don't want to clutter things too much, so we only build replacements for the most commonly used binaries.
LEGACY_TARGETS_BUILD = main quantize perplexity embedding server
# Deprecation aliases
ifdef LLAMA_CUBLAS
$(error LLAMA_CUBLAS is removed. Use GGML_CUDA instead.)
endif
ifdef LLAMA_CUDA
GGML_CUDA := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_KOMPUTE
GGML_KOMPUTE := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_METAL
GGML_METAL := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_OPENMP
GGML_OPENMP := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_RPC
GGML_RPC := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_SYCL
GGML_SYCL := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_SYCL_F16
GGML_SYCL_F16 := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_OPENBLAS
GGML_OPENBLAS := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_OPENBLAS64
GGML_OPENBLAS64 := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_BLIS
GGML_BLIS := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_NO_LLAMAFILE
GGML_NO_LLAMAFILE := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_NO_ACCELERATE
GGML_NO_ACCELERATE := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_NO_OPENMP
GGML_NO_OPENMP := 1
DEPRECATE_WARNING := 1
endif
ifdef LLAMA_NO_METAL
GGML_NO_METAL := 1
DEPRECATE_WARNING := 1
endif
ifndef UNAME_S
UNAME_S := $(shell uname -s)
endif
ifndef UNAME_P
UNAME_P := $(shell uname -p)
endif
ifndef UNAME_M
UNAME_M := $(shell uname -m)
endif
# In GNU make default CXX is g++ instead of c++. Let's fix that so that users
# of non-gcc compilers don't have to provide g++ alias or wrapper.
DEFCC := cc
DEFCXX := c++
ifeq ($(origin CC),default)
CC := $(DEFCC)
endif
ifeq ($(origin CXX),default)
CXX := $(DEFCXX)
endif
# Mac OS + Arm can report x86_64
# ref: https://github.com/ggerganov/whisper.cpp/issues/66#issuecomment-1282546789
ifeq ($(UNAME_S),Darwin)
ifndef GGML_NO_METAL
GGML_METAL := 1
endif
GGML_NO_OPENMP := 1
ifneq ($(UNAME_P),arm)
SYSCTL_M := $(shell sysctl -n hw.optional.arm64 2>/dev/null)
ifeq ($(SYSCTL_M),1)
# UNAME_P := arm
# UNAME_M := arm64
warn := $(warning Your arch is announced as x86_64, but it seems to actually be ARM64. Not fixing that can lead to bad performance. For more info see: https://github.com/ggerganov/whisper.cpp/issues/66\#issuecomment-1282546789)
endif
endif
endif
ifdef GGML_METAL
GGML_METAL_EMBED_LIBRARY := 1
endif
ifdef GGML_RPC
BUILD_TARGETS += rpc-server
endif
ifdef GGML_VULKAN
BUILD_TARGETS += vulkan-shaders-gen
endif
default: $(BUILD_TARGETS) $(LEGACY_TARGETS_BUILD)
test: $(TEST_TARGETS)
@failures=0; \
for test_target in $(TEST_TARGETS); do \
if [ "$$test_target" = "tests/test-tokenizer-0" ]; then \
./$$test_target $(CURDIR)/models/ggml-vocab-llama-spm.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-llama-bpe.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-phi-3.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-falcon.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-bert-bge.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-starcoder.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-gpt-2.gguf; \
./$$test_target $(CURDIR)/models/ggml-vocab-refact.gguf; \
elif [ "$$test_target" = "tests/test-tokenizer-1-spm" ]; then \
continue; \
elif [ "$$test_target" = "tests/test-tokenizer-1-bpe" ]; then \
continue; \
else \
echo "Running test $$test_target..."; \
./$$test_target; \
fi; \
if [ $$? -ne 0 ]; then \
printf 'Test %s FAILED!\n\n' $$test_target; \
failures=$$(( failures + 1 )); \
else \
printf 'Test %s passed.\n\n' $$test_target; \
fi; \
done; \
if [ $$failures -gt 0 ]; then \
printf '\n%s tests failed.\n' $$failures; \
exit 1; \
fi
@echo 'All tests passed.'
all: $(BUILD_TARGETS) $(TEST_TARGETS) $(LEGACY_TARGETS_BUILD)
ifdef RISCV_CROSS_COMPILE
CC := riscv64-unknown-linux-gnu-gcc
CXX := riscv64-unknown-linux-gnu-g++
endif
#
# Compile flags
#
# keep standard at C11 and C++17
MK_CPPFLAGS = -Iggml/include -Iggml/src -Iinclude -Isrc -Icommon
MK_CFLAGS = -std=c11 -fPIC
MK_CXXFLAGS = -std=c++17 -fPIC
MK_NVCCFLAGS = -std=c++17
ifdef LLAMA_NO_CCACHE
GGML_NO_CCACHE := 1
DEPRECATE_WARNING := 1
endif
ifndef GGML_NO_CCACHE
CCACHE := $(shell which ccache)
ifdef CCACHE
export CCACHE_SLOPPINESS = time_macros
$(info I ccache found, compilation results will be cached. Disable with GGML_NO_CCACHE.)
CC := $(CCACHE) $(CC)
CXX := $(CCACHE) $(CXX)
else
$(info I ccache not found. Consider installing it for faster compilation.)
endif # CCACHE
endif # GGML_NO_CCACHE
# clock_gettime came in POSIX.1b (1993)
# CLOCK_MONOTONIC came in POSIX.1-2001 / SUSv3 as optional
# posix_memalign came in POSIX.1-2001 / SUSv3
# M_PI is an XSI extension since POSIX.1-2001 / SUSv3, came in XPG1 (1985)
MK_CPPFLAGS += -D_XOPEN_SOURCE=600
# Somehow in OpenBSD whenever POSIX conformance is specified
# some string functions rely on locale_t availability,
# which was introduced in POSIX.1-2008, forcing us to go higher
ifeq ($(UNAME_S),OpenBSD)
MK_CPPFLAGS += -U_XOPEN_SOURCE -D_XOPEN_SOURCE=700
endif
# Data types, macros and functions related to controlling CPU affinity and
# some memory allocation are available on Linux through GNU extensions in libc
ifeq ($(UNAME_S),Linux)
MK_CPPFLAGS += -D_GNU_SOURCE
endif
# RLIMIT_MEMLOCK came in BSD, is not specified in POSIX.1,
# and on macOS its availability depends on enabling Darwin extensions
# similarly on DragonFly, enabling BSD extensions is necessary
ifeq ($(UNAME_S),Darwin)
MK_CPPFLAGS += -D_DARWIN_C_SOURCE
endif
ifeq ($(UNAME_S),DragonFly)
MK_CPPFLAGS += -D__BSD_VISIBLE
endif
# alloca is a non-standard interface that is not visible on BSDs when
# POSIX conformance is specified, but not all of them provide a clean way
# to enable it in such cases
ifeq ($(UNAME_S),FreeBSD)
MK_CPPFLAGS += -D__BSD_VISIBLE
endif
ifeq ($(UNAME_S),NetBSD)
MK_CPPFLAGS += -D_NETBSD_SOURCE
endif
ifeq ($(UNAME_S),OpenBSD)
MK_CPPFLAGS += -D_BSD_SOURCE
endif
ifdef GGML_SCHED_MAX_COPIES
MK_CPPFLAGS += -DGGML_SCHED_MAX_COPIES=$(GGML_SCHED_MAX_COPIES)
endif
ifdef LLAMA_DEBUG
MK_CFLAGS += -O0 -g
MK_CXXFLAGS += -O0 -g
MK_LDFLAGS += -g
MK_NVCCFLAGS += -O0 -g
ifeq ($(UNAME_S),Linux)
MK_CPPFLAGS += -D_GLIBCXX_ASSERTIONS
endif
else
MK_CPPFLAGS += -DNDEBUG
MK_CFLAGS += -O3
MK_CXXFLAGS += -O3
MK_NVCCFLAGS += -O3
endif
ifdef LLAMA_SANITIZE_THREAD
MK_CFLAGS += -fsanitize=thread -g
MK_CXXFLAGS += -fsanitize=thread -g
MK_LDFLAGS += -fsanitize=thread -g
endif
ifdef LLAMA_SANITIZE_ADDRESS
MK_CFLAGS += -fsanitize=address -fno-omit-frame-pointer -g
MK_CXXFLAGS += -fsanitize=address -fno-omit-frame-pointer -g
MK_LDFLAGS += -fsanitize=address -fno-omit-frame-pointer -g
endif
ifdef LLAMA_SANITIZE_UNDEFINED
MK_CFLAGS += -fsanitize=undefined -g
MK_CXXFLAGS += -fsanitize=undefined -g
MK_LDFLAGS += -fsanitize=undefined -g
endif
ifdef LLAMA_SERVER_VERBOSE
MK_CPPFLAGS += -DSERVER_VERBOSE=$(LLAMA_SERVER_VERBOSE)
endif
ifdef LLAMA_SERVER_SSL
MK_CPPFLAGS += -DCPPHTTPLIB_OPENSSL_SUPPORT
MK_LDFLAGS += -lssl -lcrypto
endif
ifdef LLAMA_DISABLE_LOGS
MK_CPPFLAGS += -DLOG_DISABLE_LOGS
endif # LLAMA_DISABLE_LOGS
# warnings
WARN_FLAGS = \
-Wall \
-Wextra \
-Wpedantic \
-Wcast-qual \
-Wno-unused-function
MK_CFLAGS += \
$(WARN_FLAGS) \
-Wshadow \
-Wstrict-prototypes \
-Wpointer-arith \
-Wmissing-prototypes \
-Werror=implicit-int \
-Werror=implicit-function-declaration
MK_CXXFLAGS += \
$(WARN_FLAGS) \
-Wmissing-declarations \
-Wmissing-noreturn
ifeq ($(LLAMA_FATAL_WARNINGS),1)
MK_CFLAGS += -Werror
MK_CXXFLAGS += -Werror
endif
# this version of Apple ld64 is buggy
ifneq '' '$(findstring dyld-1015.7,$(shell $(CC) $(LDFLAGS) -Wl,-v 2>&1))'
MK_CPPFLAGS += -DHAVE_BUGGY_APPLE_LINKER
endif
# OS specific
# TODO: support Windows
ifneq '' '$(filter $(UNAME_S),Linux Darwin FreeBSD NetBSD OpenBSD Haiku)'
MK_CFLAGS += -pthread
MK_CXXFLAGS += -pthread
endif
# detect Windows
ifneq ($(findstring _NT,$(UNAME_S)),)
_WIN32 := 1
endif
# library name prefix
ifneq ($(_WIN32),1)
LIB_PRE := lib
endif
# Dynamic Shared Object extension
ifneq ($(_WIN32),1)
DSO_EXT := .so
else
DSO_EXT := .dll
endif
# Windows Sockets 2 (Winsock) for network-capable apps
ifeq ($(_WIN32),1)
LWINSOCK2 := -lws2_32
endif
ifdef LLAMA_GPROF
MK_CFLAGS += -pg
MK_CXXFLAGS += -pg
endif
# Architecture specific
# TODO: probably these flags need to be tweaked on some architectures
# feel free to update the Makefile for your architecture and send a pull request or issue
ifndef RISCV
ifeq ($(UNAME_M),$(filter $(UNAME_M),x86_64 i686 amd64))
# Use all CPU extensions that are available:
MK_CFLAGS += -march=native -mtune=native
HOST_CXXFLAGS += -march=native -mtune=native
# Usage AVX-only
#MK_CFLAGS += -mfma -mf16c -mavx
#MK_CXXFLAGS += -mfma -mf16c -mavx
# Usage SSSE3-only (Not is SSE3!)
#MK_CFLAGS += -mssse3
#MK_CXXFLAGS += -mssse3
endif
ifneq '' '$(findstring mingw,$(shell $(CC) -dumpmachine))'
# The stack is only 16-byte aligned on Windows, so don't let gcc emit aligned moves.
# https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54412
# https://github.com/ggerganov/llama.cpp/issues/2922
MK_CFLAGS += -Xassembler -muse-unaligned-vector-move
MK_CXXFLAGS += -Xassembler -muse-unaligned-vector-move
# Target Windows 8 for PrefetchVirtualMemory
MK_CPPFLAGS += -D_WIN32_WINNT=0x602
endif
ifneq ($(filter aarch64%,$(UNAME_M)),)
# Apple M1, M2, etc.
# Raspberry Pi 3, 4, Zero 2 (64-bit)
# Nvidia Jetson
MK_CFLAGS += -mcpu=native
MK_CXXFLAGS += -mcpu=native
JETSON_RELEASE_INFO = $(shell jetson_release)
ifdef JETSON_RELEASE_INFO
ifneq ($(filter TX2%,$(JETSON_RELEASE_INFO)),)
JETSON_EOL_MODULE_DETECT = 1
CC = aarch64-unknown-linux-gnu-gcc
cxx = aarch64-unknown-linux-gnu-g++
endif
endif
endif
ifneq ($(filter armv6%,$(UNAME_M)),)
# Raspberry Pi 1, Zero
MK_CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access
MK_CXXFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access
endif
ifneq ($(filter armv7%,$(UNAME_M)),)
# Raspberry Pi 2
MK_CFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access -funsafe-math-optimizations
MK_CXXFLAGS += -mfpu=neon-fp-armv8 -mfp16-format=ieee -mno-unaligned-access -funsafe-math-optimizations
endif
ifneq ($(filter armv8%,$(UNAME_M)),)
# Raspberry Pi 3, 4, Zero 2 (32-bit)
MK_CFLAGS += -mfp16-format=ieee -mno-unaligned-access
MK_CXXFLAGS += -mfp16-format=ieee -mno-unaligned-access
endif
ifneq ($(filter ppc64%,$(UNAME_M)),)
POWER9_M := $(shell grep "POWER9" /proc/cpuinfo)
ifneq (,$(findstring POWER9,$(POWER9_M)))
MK_CFLAGS += -mcpu=power9
MK_CXXFLAGS += -mcpu=power9
endif
endif
ifneq ($(filter ppc64le%,$(UNAME_M)),)
MK_CFLAGS += -mcpu=powerpc64le
MK_CXXFLAGS += -mcpu=powerpc64le
CUDA_POWER_ARCH = 1
endif
ifneq ($(filter loongarch64%,$(UNAME_M)),)
MK_CFLAGS += -mlasx
MK_CXXFLAGS += -mlasx
endif
else
MK_CFLAGS += -march=rv64gcv -mabi=lp64d
MK_CXXFLAGS += -march=rv64gcv -mabi=lp64d
endif
ifndef GGML_NO_ACCELERATE
# Mac OS - include Accelerate framework.
# `-framework Accelerate` works both with Apple Silicon and Mac Intel
ifeq ($(UNAME_S),Darwin)
MK_CPPFLAGS += -DGGML_USE_ACCELERATE -DGGML_USE_BLAS
MK_CPPFLAGS += -DACCELERATE_NEW_LAPACK
MK_CPPFLAGS += -DACCELERATE_LAPACK_ILP64
MK_LDFLAGS += -framework Accelerate
OBJ_GGML += ggml/src/ggml-blas.o
endif
endif # GGML_NO_ACCELERATE
ifdef GGML_MUSA
CC := clang
CXX := clang++
GGML_CUDA := 1
MK_CPPFLAGS += -DGGML_USE_MUSA
endif
ifndef GGML_NO_OPENMP
MK_CPPFLAGS += -DGGML_USE_OPENMP
MK_CFLAGS += -fopenmp
MK_CXXFLAGS += -fopenmp
ifdef GGML_MUSA
MK_CPPFLAGS += -I/usr/lib/llvm-10/include/openmp
MK_LDFLAGS += -L/usr/lib/llvm-10/lib
endif # GGML_MUSA
endif # GGML_NO_OPENMP
ifdef GGML_OPENBLAS
MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas)
MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas)
MK_LDFLAGS += $(shell pkg-config --libs openblas)
OBJ_GGML += ggml/src/ggml-blas.o
endif # GGML_OPENBLAS
ifdef GGML_OPENBLAS64
MK_CPPFLAGS += -DGGML_USE_BLAS $(shell pkg-config --cflags-only-I openblas64)
MK_CFLAGS += $(shell pkg-config --cflags-only-other openblas64)
MK_LDFLAGS += $(shell pkg-config --libs openblas64)
OBJ_GGML += ggml/src/ggml-blas.o
endif # GGML_OPENBLAS64
ifdef GGML_BLIS
MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_BLIS -I/usr/local/include/blis -I/usr/include/blis
MK_LDFLAGS += -lblis -L/usr/local/lib
OBJ_GGML += ggml/src/ggml-blas.o
endif # GGML_BLIS
ifdef GGML_NVPL
MK_CPPFLAGS += -DGGML_USE_BLAS -DGGML_BLAS_USE_NVPL -DNVPL_ILP64 -I/usr/local/include/nvpl_blas -I/usr/include/nvpl_blas
MK_LDFLAGS += -L/usr/local/lib -lnvpl_blas_core -lnvpl_blas_ilp64_gomp
OBJ_GGML += ggml/src/ggml-blas.o
endif # GGML_NVPL
OBJ_GGML += ggml/src/iqk/iqk_quantize.o
ifndef GGML_NO_IQKMULMAT
MK_CFLAGS += -DGGML_USE_IQK_MULMAT
MK_CPPFLAGS += -DGGML_USE_IQK_MULMAT
OBJ_GGML += ggml/src/iqk/iqk_mul_mat.o
endif
ifndef GGML_NO_LLAMAFILE
MK_CPPFLAGS += -DGGML_USE_LLAMAFILE
OBJ_GGML += ggml/src/llamafile/sgemm.o
endif
ifdef GGML_RPC
MK_CPPFLAGS += -DGGML_USE_RPC
OBJ_GGML += ggml/src/ggml-rpc.o
endif # GGML_RPC
OBJ_CUDA_TMPL = $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-wmma*.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/mmq*.cu))
ifdef GGML_CUDA_FA_ALL_QUANTS
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*.cu))
else
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*f16-f16.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*q8_0-iq4_nl.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*:iq4_nl-iq4_nl.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*:q6_0-q5_0.cu))
OBJ_CUDA_TMPL += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/template-instances/fattn-vec*:q8_0-q6_0.cu))
endif # GGML_CUDA_FA_ALL_QUANTS
ifdef GGML_CUDA
ifdef GGML_MUSA
ifneq ('', '$(wildcard /opt/musa)')
CUDA_PATH ?= /opt/musa
else
CUDA_PATH ?= /usr/local/musa
endif
MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include
MK_LDFLAGS += -lmusa -lmublas -lmusart -lpthread -ldl -lrt -L$(CUDA_PATH)/lib -L/usr/lib64
MK_NVCCFLAGS += -x musa -mtgpu --cuda-gpu-arch=mp_22
else
ifneq ('', '$(wildcard /opt/cuda)')
CUDA_PATH ?= /opt/cuda
else
CUDA_PATH ?= /usr/local/cuda
endif
MK_CPPFLAGS += -DGGML_USE_CUDA -I$(CUDA_PATH)/include -I$(CUDA_PATH)/targets/$(UNAME_M)-linux/include -DGGML_CUDA_USE_GRAPHS
MK_LDFLAGS += -lcuda -lcublas -lculibos -lcudart -lcublasLt -lpthread -ldl -lrt -L$(CUDA_PATH)/lib64 -L/usr/lib64 -L$(CUDA_PATH)/targets/$(UNAME_M)-linux/lib -L$(CUDA_PATH)/lib64/stubs -L/usr/lib/wsl/lib
MK_NVCCFLAGS += -use_fast_math
endif # GGML_MUSA
OBJ_GGML += ggml/src/ggml-cuda.o
OBJ_GGML += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu))
OBJ_GGML += $(OBJ_CUDA_TMPL)
ifdef LLAMA_FATAL_WARNINGS
MK_NVCCFLAGS += -Werror all-warnings
endif # LLAMA_FATAL_WARNINGS
ifndef GGML_MUSA
ifndef JETSON_EOL_MODULE_DETECT
MK_NVCCFLAGS += --forward-unknown-to-host-compiler
endif # JETSON_EOL_MODULE_DETECT
endif # GGML_MUSA
ifdef LLAMA_DEBUG
MK_NVCCFLAGS += -lineinfo
endif # LLAMA_DEBUG
ifdef GGML_CUDA_DEBUG
MK_NVCCFLAGS += --device-debug
endif # GGML_CUDA_DEBUG
ifdef GGML_CUDA_NVCC
NVCC = $(CCACHE) $(GGML_CUDA_NVCC)
else
ifdef GGML_MUSA
NVCC = $(CCACHE) mcc
else
NVCC = $(CCACHE) nvcc
endif # GGML_MUSA
endif # GGML_CUDA_NVCC
ifdef CUDA_DOCKER_ARCH
MK_NVCCFLAGS += -Wno-deprecated-gpu-targets -arch=$(CUDA_DOCKER_ARCH)
else ifndef CUDA_POWER_ARCH
MK_NVCCFLAGS += -arch=native
endif # CUDA_DOCKER_ARCH
ifdef GGML_CUDA_FORCE_DMMV
MK_NVCCFLAGS += -DGGML_CUDA_FORCE_DMMV
endif # GGML_CUDA_FORCE_DMMV
ifdef GGML_CUDA_FORCE_MMQ
MK_NVCCFLAGS += -DGGML_CUDA_FORCE_MMQ
endif # GGML_CUDA_FORCE_MMQ
ifdef GGML_CUDA_FORCE_CUBLAS
MK_NVCCFLAGS += -DGGML_CUDA_FORCE_CUBLAS
endif # GGML_CUDA_FORCE_CUBLAS
ifdef GGML_CUDA_DMMV_X
MK_NVCCFLAGS += -DGGML_CUDA_DMMV_X=$(GGML_CUDA_DMMV_X)
else
MK_NVCCFLAGS += -DGGML_CUDA_DMMV_X=32
endif # GGML_CUDA_DMMV_X
ifdef GGML_CUDA_MMV_Y
MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_MMV_Y)
else ifdef GGML_CUDA_DMMV_Y
MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_DMMV_Y) # for backwards compatibility
else
MK_NVCCFLAGS += -DGGML_CUDA_MMV_Y=1
endif # GGML_CUDA_MMV_Y
ifdef GGML_CUDA_F16
MK_NVCCFLAGS += -DGGML_CUDA_F16
endif # GGML_CUDA_F16
ifdef GGML_CUDA_DMMV_F16
MK_NVCCFLAGS += -DGGML_CUDA_F16
endif # GGML_CUDA_DMMV_F16
ifdef GGML_CUDA_KQUANTS_ITER
MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER)
else
MK_NVCCFLAGS += -DK_QUANTS_PER_ITERATION=2
endif
ifdef GGML_CUDA_PEER_MAX_BATCH_SIZE
MK_NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=$(GGML_CUDA_PEER_MAX_BATCH_SIZE)
else
MK_NVCCFLAGS += -DGGML_CUDA_PEER_MAX_BATCH_SIZE=128
endif # GGML_CUDA_PEER_MAX_BATCH_SIZE
ifdef GGML_CUDA_NO_PEER_COPY
MK_NVCCFLAGS += -DGGML_CUDA_NO_PEER_COPY
endif # GGML_CUDA_NO_PEER_COPY
ifdef GGML_CUDA_CCBIN
MK_NVCCFLAGS += -ccbin $(GGML_CUDA_CCBIN)
endif # GGML_CUDA_CCBIN
ifdef GGML_CUDA_FA_ALL_QUANTS
MK_NVCCFLAGS += -DGGML_CUDA_FA_ALL_QUANTS
endif # GGML_CUDA_FA_ALL_QUANTS
ifdef JETSON_EOL_MODULE_DETECT
define NVCC_COMPILE
$(NVCC) -I. -Icommon -D_XOPEN_SOURCE=600 -D_GNU_SOURCE -DNDEBUG -DGGML_USE_CUDA -I/usr/local/cuda/include -I/opt/cuda/include -I/usr/local/cuda/targets/aarch64-linux/include -std=c++11 -O3 $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@
endef # NVCC_COMPILE
else
ifdef GGML_MUSA
define NVCC_COMPILE
$(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -c $< -o $@
endef # NVCC_COMPILE
else
define NVCC_COMPILE
$(NVCC) $(NVCCFLAGS) $(CPPFLAGS) -Xcompiler "$(CUDA_CXXFLAGS)" -c $< -o $@
endef # NVCC_COMPILE
endif # GGML_MUSA
endif # JETSON_EOL_MODULE_DETECT
ggml/src/ggml-cuda/%.o: \
ggml/src/ggml-cuda/%.cu \
ggml/include/ggml.h \
ggml/src/ggml-common.h \
ggml/src/ggml-cuda/common.cuh
$(NVCC_COMPILE)
ggml/src/ggml-cuda.o: \
ggml/src/ggml-cuda.cu \
ggml/include/ggml-cuda.h \
ggml/include/ggml.h \
ggml/include/ggml-backend.h \
ggml/src/ggml-backend-impl.h \
ggml/src/ggml-common.h \
$(wildcard ggml/src/ggml-cuda/*.cuh)
$(NVCC_COMPILE)
endif # GGML_CUDA
ifdef GGML_VULKAN
MK_CPPFLAGS += -DGGML_USE_VULKAN
MK_LDFLAGS += $(shell pkg-config --libs vulkan)
OBJ_GGML += ggml/src/ggml-vulkan.o ggml/src/ggml-vulkan-shaders.o
ifdef GGML_VULKAN_CHECK_RESULTS
MK_CPPFLAGS += -DGGML_VULKAN_CHECK_RESULTS
endif
ifdef GGML_VULKAN_DEBUG
MK_CPPFLAGS += -DGGML_VULKAN_DEBUG
endif
ifdef GGML_VULKAN_MEMORY_DEBUG
MK_CPPFLAGS += -DGGML_VULKAN_MEMORY_DEBUG
endif
ifdef GGML_VULKAN_PERF
MK_CPPFLAGS += -DGGML_VULKAN_PERF
endif
ifdef GGML_VULKAN_VALIDATE
MK_CPPFLAGS += -DGGML_VULKAN_VALIDATE
endif
ifdef GGML_VULKAN_RUN_TESTS
MK_CPPFLAGS += -DGGML_VULKAN_RUN_TESTS
endif
GLSLC_CMD = glslc
_ggml_vk_genshaders_cmd = $(shell pwd)/vulkan-shaders-gen
_ggml_vk_header = ggml/src/ggml-vulkan-shaders.hpp
_ggml_vk_source = ggml/src/ggml-vulkan-shaders.cpp
_ggml_vk_input_dir = ggml/src/vulkan-shaders
_ggml_vk_shader_deps = $(echo $(_ggml_vk_input_dir)/*.comp)
ggml/src/ggml-vulkan.o: ggml/src/ggml-vulkan.cpp ggml/include/ggml-vulkan.h $(_ggml_vk_header) $(_ggml_vk_source)
$(CXX) $(CXXFLAGS) $(shell pkg-config --cflags vulkan) -c $< -o $@
$(_ggml_vk_header): $(_ggml_vk_source)
$(_ggml_vk_source): $(_ggml_vk_shader_deps) vulkan-shaders-gen
$(_ggml_vk_genshaders_cmd) \
--glslc $(GLSLC_CMD) \
--input-dir $(_ggml_vk_input_dir) \
--target-hpp $(_ggml_vk_header) \
--target-cpp $(_ggml_vk_source)
vulkan-shaders-gen: ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp
$(CXX) $(CXXFLAGS) -o $@ $(LDFLAGS) ggml/src/vulkan-shaders/vulkan-shaders-gen.cpp
endif # GGML_VULKAN
ifdef GGML_HIPBLAS
ifeq ($(wildcard /opt/rocm),)
ROCM_PATH ?= /usr
AMDGPU_TARGETS ?= $(shell $(shell which amdgpu-arch))
else
ROCM_PATH ?= /opt/rocm
AMDGPU_TARGETS ?= $(shell $(ROCM_PATH)/llvm/bin/amdgpu-arch)
endif
GGML_CUDA_DMMV_X ?= 32
GGML_CUDA_MMV_Y ?= 1
GGML_CUDA_KQUANTS_ITER ?= 2
MK_CPPFLAGS += -DGGML_USE_HIPBLAS -DGGML_USE_CUDA
ifdef GGML_HIP_UMA
MK_CPPFLAGS += -DGGML_HIP_UMA
endif # GGML_HIP_UMA
MK_LDFLAGS += -L$(ROCM_PATH)/lib -Wl,-rpath=$(ROCM_PATH)/lib
MK_LDFLAGS += -L$(ROCM_PATH)/lib64 -Wl,-rpath=$(ROCM_PATH)/lib64
MK_LDFLAGS += -lhipblas -lamdhip64 -lrocblas
HIPCC ?= $(CCACHE) $(ROCM_PATH)/bin/hipcc
HIPFLAGS += $(addprefix --offload-arch=,$(AMDGPU_TARGETS))
HIPFLAGS += -DGGML_CUDA_DMMV_X=$(GGML_CUDA_DMMV_X)
HIPFLAGS += -DGGML_CUDA_MMV_Y=$(GGML_CUDA_MMV_Y)
HIPFLAGS += -DK_QUANTS_PER_ITERATION=$(GGML_CUDA_KQUANTS_ITER)
ifdef GGML_CUDA_FORCE_DMMV
HIPFLAGS += -DGGML_CUDA_FORCE_DMMV
endif # GGML_CUDA_FORCE_DMMV
ifdef GGML_CUDA_FORCE_MMQ
HIPFLAGS += -DGGML_CUDA_FORCE_MMQ
endif # GGML_CUDA_FORCE_MMQ
ifdef GGML_CUDA_FORCE_CUBLAS
HIPFLAGS += -DGGML_CUDA_FORCE_CUBLAS
endif # GGML_CUDA_FORCE_CUBLAS
ifdef GGML_CUDA_NO_PEER_COPY
HIPFLAGS += -DGGML_CUDA_NO_PEER_COPY
endif # GGML_CUDA_NO_PEER_COPY
OBJ_GGML += ggml/src/ggml-cuda.o
OBJ_GGML += $(patsubst %.cu,%.o,$(wildcard ggml/src/ggml-cuda/*.cu))
OBJ_GGML += $(OBJ_CUDA_TMPL)
ggml/src/ggml-cuda.o: \
ggml/src/ggml-cuda.cu \
ggml/include/ggml-cuda.h \
ggml/include/ggml.h \
ggml/include/ggml-backend.h \
ggml/src/ggml-backend-impl.h \
ggml/src/ggml-common.h \
$(wildcard ggml/src/ggml-cuda/*.cuh)
$(HIPCC) $(CXXFLAGS) $(HIPFLAGS) -x hip -c -o $@ $<
ggml/src/ggml-cuda/%.o: \
ggml/src/ggml-cuda/%.cu \
ggml/include/ggml.h \
ggml/src/ggml-common.h \
ggml/src/ggml-cuda/common.cuh
$(HIPCC) $(CXXFLAGS) $(HIPFLAGS) -x hip -c -o $@ $<
endif # GGML_HIPBLAS
ifdef GGML_METAL
MK_CPPFLAGS += -DGGML_USE_METAL
MK_LDFLAGS += -framework Foundation -framework Metal -framework MetalKit
OBJ_GGML += ggml/src/ggml-metal.o
ifdef GGML_METAL_NDEBUG
MK_CPPFLAGS += -DGGML_METAL_NDEBUG
endif
ifdef GGML_METAL_EMBED_LIBRARY
MK_CPPFLAGS += -DGGML_METAL_EMBED_LIBRARY
OBJ_GGML += ggml/src/ggml-metal-embed.o
endif
endif # GGML_METAL
ifdef GGML_METAL
ggml/src/ggml-metal.o: \
ggml/src/ggml-metal.m \
ggml/include/ggml-metal.h \
ggml/include/ggml.h
$(CC) $(CFLAGS) -c $< -o $@
ifdef GGML_METAL_EMBED_LIBRARY
ggml/src/ggml-metal-embed.o: \
ggml/src/ggml-metal.metal \
ggml/src/ggml-common.h
@echo "Embedding Metal library"
@sed -e '/#include "ggml-common.h"/r ggml/src/ggml-common.h' -e '/#include "ggml-common.h"/d' < ggml/src/ggml-metal.metal > ggml/src/ggml-metal-embed.metal
$(eval TEMP_ASSEMBLY=$(shell mktemp -d))
@echo ".section __DATA, __ggml_metallib" > $(TEMP_ASSEMBLY)/ggml-metal-embed.s
@echo ".globl _ggml_metallib_start" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
@echo "_ggml_metallib_start:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
@echo ".incbin \"ggml/src/ggml-metal-embed.metal\"" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
@echo ".globl _ggml_metallib_end" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
@echo "_ggml_metallib_end:" >> $(TEMP_ASSEMBLY)/ggml-metal-embed.s
$(CC) $(CFLAGS) -c $(TEMP_ASSEMBLY)/ggml-metal-embed.s -o $@
@rm -f ${TEMP_ASSEMBLY}/ggml-metal-embed.s
@rmdir ${TEMP_ASSEMBLY}
endif
endif # GGML_METAL
OBJ_GGML += \
ggml/src/ggml.o \
ggml/src/ggml-alloc.o \
ggml/src/ggml-backend.o \
ggml/src/ggml-quants.o \
ggml/src/ggml-aarch64.o
OBJ_LLAMA = \
src/llama.o \
src/llama-vocab.o \
src/llama-grammar.o \
src/llama-sampling.o \
src/unicode.o \
src/unicode-data.o
OBJ_COMMON = \
common/common.o \
common/console.o \
common/ngram-cache.o \
common/sampling.o \
common/train.o \
common/grammar-parser.o \
common/build-info.o \
common/json-schema-to-grammar.o
OBJ_ALL = $(OBJ_GGML) $(OBJ_LLAMA) $(OBJ_COMMON)
LIB_GGML = $(LIB_PRE)ggml$(DSO_EXT)
LIB_GGML_S = $(LIB_PRE)ggml.a
LIB_LLAMA = $(LIB_PRE)llama$(DSO_EXT)
LIB_LLAMA_S = $(LIB_PRE)llama.a
LIB_COMMON = $(LIB_PRE)common$(DSO_EXT)
LIB_COMMON_S = $(LIB_PRE)common.a
LIB_ALL = $(LIB_GGML) $(LIB_LLAMA) $(LIB_COMMON)
LIB_ALL_S = $(LIB_GGML_S) $(LIB_LLAMA_S) $(LIB_COMMON_S)
GF_CC := $(CC)
include scripts/get-flags.mk
# combine build flags with cmdline overrides
override CPPFLAGS := $(MK_CPPFLAGS) $(CPPFLAGS)
override CFLAGS := $(CPPFLAGS) $(MK_CFLAGS) $(GF_CFLAGS) $(CFLAGS)
BASE_CXXFLAGS := $(MK_CXXFLAGS) $(CXXFLAGS)
override CXXFLAGS := $(BASE_CXXFLAGS) $(HOST_CXXFLAGS) $(GF_CXXFLAGS) $(CPPFLAGS)
override NVCCFLAGS := $(MK_NVCCFLAGS) $(NVCCFLAGS)
override LDFLAGS := $(MK_LDFLAGS) $(LDFLAGS)
# identify CUDA host compiler
ifdef GGML_CUDA
GF_CC := $(NVCC) $(NVCCFLAGS) 2>/dev/null .c -Xcompiler
include scripts/get-flags.mk
CUDA_CXXFLAGS := $(BASE_CXXFLAGS) $(GF_CXXFLAGS) -Wno-pedantic
endif
ifdef LLAMA_CURL
override CXXFLAGS := $(CXXFLAGS) -DLLAMA_USE_CURL
override LDFLAGS := $(LDFLAGS) -lcurl
endif
#
# Print build information
#
$(info I llama.cpp build info: )
$(info I UNAME_S: $(UNAME_S))
$(info I UNAME_P: $(UNAME_P))
$(info I UNAME_M: $(UNAME_M))
$(info I CFLAGS: $(CFLAGS))
$(info I CXXFLAGS: $(CXXFLAGS))
$(info I NVCCFLAGS: $(NVCCFLAGS))
$(info I LDFLAGS: $(LDFLAGS))
$(info I CC: $(shell $(CC) --version | head -n 1))
$(info I CXX: $(shell $(CXX) --version | head -n 1))
ifdef GGML_CUDA
$(info I NVCC: $(shell $(NVCC) --version | tail -n 1))
CUDA_VERSION := $(shell $(NVCC) --version | grep -oP 'release (\K[0-9]+\.[0-9])')
ifndef GGML_MUSA
ifeq ($(shell awk -v "v=$(CUDA_VERSION)" 'BEGIN { print (v < 11.7) }'),1)
ifndef CUDA_DOCKER_ARCH
ifndef CUDA_POWER_ARCH
$(error I ERROR: For CUDA versions < 11.7 a target CUDA architecture must be explicitly provided via environment variable CUDA_DOCKER_ARCH, e.g. by running "export CUDA_DOCKER_ARCH=compute_XX" on Unix-like systems, where XX is the minimum compute capability that the code needs to run on. A list with compute capabilities can be found here: https://developer.nvidia.com/cuda-gpus )
endif # CUDA_POWER_ARCH
endif # CUDA_DOCKER_ARCH
endif # eq ($(shell echo "$(CUDA_VERSION) < 11.7" | bc),1)
endif # GGML_MUSA
endif # GGML_CUDA
$(info )
ifdef DEPRECATE_WARNING
$(info !!! DEPRECATION WARNING !!!)
$(info The following LLAMA_ options are deprecated and will be removed in the future. Use the GGML_ prefix instead)
$(info - LLAMA_CUDA)
$(info - LLAMA_METAL)
$(info - LLAMA_METAL_EMBED_LIBRARY)
$(info - LLAMA_OPENMP)
$(info - LLAMA_RPC)
$(info - LLAMA_SYCL)
$(info - LLAMA_SYCL_F16)
$(info - LLAMA_OPENBLAS)
$(info - LLAMA_OPENBLAS64)
$(info - LLAMA_BLIS)
$(info - LLAMA_NO_LLAMAFILE)
$(info - LLAMA_NO_ACCELERATE)
$(info - LLAMA_NO_OPENMP)
$(info - LLAMA_NO_METAL)
$(info - LLAMA_NO_CCACHE)
$(info )
endif
#
# Build libraries
#
# ggml
ggml/src/ggml.o: \
ggml/src/ggml.c \
ggml/include/ggml.h
$(CC) $(CFLAGS) -c $< -o $@
ggml/src/ggml-alloc.o: \
ggml/src/ggml-alloc.c \
ggml/include/ggml.h \
ggml/include/ggml-alloc.h
$(CC) $(CFLAGS) -c $< -o $@
ggml/src/ggml-backend.o: \
ggml/src/ggml-backend.c \
ggml/include/ggml.h \
ggml/include/ggml-backend.h
$(CC) $(CFLAGS) -c $< -o $@
ggml/src/ggml-quants.o: \
ggml/src/ggml-quants.c \
ggml/include/ggml.h \
ggml/src/ggml-quants.h \
ggml/src/ggml-common.h
$(CC) $(CFLAGS) -c $< -o $@
ggml/src/ggml-aarch64.o: \
ggml/src/ggml-aarch64.c \
ggml/include/ggml.h \
ggml/src/ggml-aarch64.h \
ggml/src/ggml-common.h
$(CC) $(CFLAGS) -c $< -o $@
ggml/src/ggml-blas.o: \
ggml/src/ggml-blas.cpp \
ggml/include/ggml-blas.h
$(CXX) $(CXXFLAGS) -c $< -o $@
ggml/src/iqk/iqk_quantize.o: \
ggml/src/iqk/iqk_quantize.cpp \
ggml/src/iqk/iqk_quantize.h \
ggml/src/ggml-quants.h ggml/src/ggml-common.h ggml/include/ggml.h ggml/src/ggml-impl.h
ifndef GGML_NO_IQKMULMAT
ggml/src/iqk/iqk_mul_mat.o: \
ggml/src/iqk/iqk_mul_mat.cpp \
ggml/src/iqk/iqk_mul_mat.h \
ggml/src/iqk/iqk_quantize.h \
ggml/src/ggml-quants.h ggml/src/ggml-common.h ggml/include/ggml.h ggml/src/ggml-impl.h
$(CXX) $(CXXFLAGS) -c $< -o $@
endif # GGML_NO_IQKMULMAT
ifndef GGML_NO_LLAMAFILE
ggml/src/llamafile/sgemm.o: \
ggml/src/llamafile/sgemm.cpp \
ggml/src/llamafile/sgemm.h \
ggml/include/ggml.h
$(CXX) $(CXXFLAGS) -c $< -o $@
endif # GGML_NO_LLAMAFILE
ifdef GGML_RPC
ggml/src/ggml-rpc.o: \
ggml/src/ggml-rpc.cpp \
ggml/include/ggml-rpc.h
$(CXX) $(CXXFLAGS) -c $< -o $@
endif # GGML_RPC
$(LIB_GGML): \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS)
$(LIB_GGML_S): \
$(OBJ_GGML)
ar rcs $(LIB_GGML_S) $^
# llama
src/unicode.o: \
src/unicode.cpp \
src/unicode.h
$(CXX) $(CXXFLAGS) -c $< -o $@
src/unicode-data.o: \
src/unicode-data.cpp \
src/unicode-data.h
$(CXX) $(CXXFLAGS) -c $< -o $@
src/llama.o: \
src/llama.cpp \
src/llama-impl.h \
src/llama-vocab.h \
src/llama-grammar.h \
src/llama-sampling.h \
src/unicode.h \
include/llama.h \
ggml/include/ggml-cuda.h \
ggml/include/ggml-metal.h \
ggml/include/ggml.h \
ggml/include/ggml-alloc.h \
ggml/include/ggml-backend.h
$(CXX) $(CXXFLAGS) -c $< -o $@
src/llama-vocab.o: \
src/llama-vocab.cpp \
src/llama-vocab.h \
src/llama-impl.h \
include/llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@
src/llama-grammar.o: \
src/llama-grammar.cpp \
src/llama-grammar.h \
src/llama-impl.h \
src/llama-vocab.h \
src/llama-sampling.h \
include/llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@
src/llama-sampling.o: \
src/llama-sampling.cpp \
src/llama-sampling.h \
src/llama-impl.h \
include/llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@
$(LIB_LLAMA): \
$(OBJ_LLAMA) \
$(LIB_GGML)
$(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS)
$(LIB_LLAMA_S): \
$(OBJ_LLAMA)
ar rcs $(LIB_LLAMA_S) $^
# common
common/common.o: \
common/common.cpp \
common/common.h \
common/console.h \
common/sampling.h \
common/json.hpp \
common/json-schema-to-grammar.h \
include/llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/sampling.o: \
common/sampling.cpp \
common/sampling.h \
include/llama.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/console.o: \
common/console.cpp \
common/console.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/grammar-parser.o: \
common/grammar-parser.cpp \
common/grammar-parser.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/json-schema-to-grammar.o: \
common/json-schema-to-grammar.cpp \
common/json-schema-to-grammar.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/train.o: \
common/train.cpp \
common/train.h
$(CXX) $(CXXFLAGS) -c $< -o $@
common/ngram-cache.o: \
common/ngram-cache.cpp \
common/ngram-cache.h
$(CXX) $(CXXFLAGS) -c $< -o $@
$(LIB_COMMON): \
$(OBJ_COMMON) \
$(LIB_LLAMA) \
$(LIB_GGML)
$(CXX) $(CXXFLAGS) -shared -fPIC -o $@ $^ $(LDFLAGS)
$(LIB_COMMON_S): \
$(OBJ_COMMON)
ar rcs $(LIB_COMMON_S) $^
clean:
rm -vrf *.dot $(BUILD_TARGETS) $(TEST_TARGETS)
rm -rvf src/*.o
rm -rvf tests/*.o
rm -rvf examples/*.o
rm -rvf common/*.o
rm -rvf *.a
rm -rvf *.dll
rm -rvf *.so
rm -rvf *.dot
rm -rvf ggml/*.a
rm -rvf ggml/*.dll
rm -rvf ggml/*.so
rm -vrf ggml/src/*.o
rm -rvf ggml/src/llamafile/*.o
rm -rvf ggml/src/iqk/*.o
rm -rvf common/build-info.cpp
rm -vrf ggml/src/ggml-metal-embed.metal
rm -vrf ggml/src/ggml-cuda/*.o
rm -vrf ggml/src/ggml-cuda/template-instances/*.o
rm -rvf $(BUILD_TARGETS)
rm -rvf $(TEST_TARGETS)
rm -f vulkan-shaders-gen ggml/src/ggml-vulkan-shaders.hpp ggml/src/ggml-vulkan-shaders.cpp
rm -rvf $(LEGACY_TARGETS_CLEAN)
find examples pocs -type f -name "*.o" -delete
#
# Examples
#
# $< is the first prerequisite, i.e. the source file.
# Explicitly compile this to an object file so that it can be cached with ccache.
# The source file is then filtered out from $^ (the list of all prerequisites) and the object file is added instead.
# Helper function that replaces .c, .cpp, and .cu file endings with .o:
GET_OBJ_FILE = $(patsubst %.c,%.o,$(patsubst %.cpp,%.o,$(patsubst %.cu,%.o,$(1))))
llama-cli: examples/main/main.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
@echo
@echo '==== Run ./llama-cli -h for help. ===='
@echo
llama-infill: examples/infill/infill.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-simple: examples/simple/simple.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-tokenize: examples/tokenize/tokenize.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-batched: examples/batched/batched.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-batched-bench: examples/batched-bench/batched-bench.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-quantize: examples/quantize/quantize.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-quantize-stats: examples/quantize-stats/quantize-stats.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-perplexity: examples/perplexity/perplexity.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-imatrix: examples/imatrix/imatrix.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-embedding: examples/embedding/embedding.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-gritlm: examples/gritlm/gritlm.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-save-load-state: examples/save-load-state/save-load-state.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-gguf: examples/gguf/gguf.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
examples/gguf-hash/deps/sha1/sha1.o: \
examples/gguf-hash/deps/sha1/sha1.c
$(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@
examples/gguf-hash/deps/xxhash/xxhash.o: \
examples/gguf-hash/deps/xxhash/xxhash.c
$(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@
examples/gguf-hash/deps/sha256/sha256.o: \
examples/gguf-hash/deps/sha256/sha256.c
$(CC) $(CFLAGS) -Iexamples/gguf-hash/deps -c $< -o $@
llama-gguf-hash: examples/gguf-hash/gguf-hash.cpp examples/gguf-hash/deps/sha1/sha1.o examples/gguf-hash/deps/xxhash/xxhash.o examples/gguf-hash/deps/sha256/sha256.o\
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -Iexamples/gguf-hash/deps -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-gguf-split: examples/gguf-split/gguf-split.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-eval-callback: examples/eval-callback/eval-callback.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-cvector-generator: examples/cvector-generator/cvector-generator.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-convert-llama2c-to-ggml: examples/convert-llama2c-to-ggml/convert-llama2c-to-ggml.cpp \
$(OBJ_GGML) $(OBJ_LLAMA)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-bench: examples/llama-bench/llama-bench.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-baby-llama: examples/baby-llama/baby-llama.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-export-lora: examples/export-lora/export-lora.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-retrieval: examples/retrieval/retrieval.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-speculative: examples/speculative/speculative.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-parallel: examples/parallel/parallel.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-lookahead: examples/lookahead/lookahead.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-lookup: examples/lookup/lookup.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-lookup-create: examples/lookup/lookup-create.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-lookup-merge: examples/lookup/lookup-merge.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-lookup-stats: examples/lookup/lookup-stats.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-passkey: examples/passkey/passkey.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-gbnf-validator: examples/gbnf-validator/gbnf-validator.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
ifdef GGML_RPC
rpc-server: examples/rpc/rpc-server.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) $^ -o $@ $(LDFLAGS)
endif # GGML_RPC
llama-server: \
examples/server/server.cpp \
examples/server/utils.hpp \
examples/server/httplib.h \
examples/server/colorthemes.css.hpp \
examples/server/style.css.hpp \
examples/server/theme-beeninorder.css.hpp \
examples/server/theme-ketivah.css.hpp \
examples/server/theme-mangotango.css.hpp \
examples/server/theme-playground.css.hpp \
examples/server/theme-polarnight.css.hpp \
examples/server/theme-snowstorm.css.hpp \
examples/server/index.html.hpp \
examples/server/index-new.html.hpp \
examples/server/index.js.hpp \
examples/server/completion.js.hpp \
examples/server/system-prompts.js.hpp \
examples/server/prompt-formats.js.hpp \
examples/server/json-schema-to-grammar.mjs.hpp \
common/json.hpp \
common/stb_image.h \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h %.hpp $<,$^) -Iexamples/server $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS) $(LWINSOCK2)
# Portable equivalent of `cd examples/server/public && xxd -i $(notdir $<) ../$(notdir $<).hpp`:
examples/server/%.hpp: examples/server/public/% Makefile
@( export NAME=$(subst .,_,$(subst -,_,$(notdir $<))) && \
echo "unsigned char $${NAME}[] = {" && \
cat $< | od -v -t x1 -An | sed -E 's/([0-9a-fA-F]+)/0x\1, /g' && \
echo "};" && \
echo "unsigned int $${NAME}_len = $(shell cat $< | wc -c );" \
) > $@
libllava.a: examples/llava/llava.cpp \
examples/llava/llava.h \
examples/llava/clip.cpp \
examples/llava/clip.h \
common/stb_image.h \
common/base64.hpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -static -fPIC -c $< -o $@ -Wno-cast-qual
llama-llava-cli: examples/llava/llava-cli.cpp \
examples/llava/llava.cpp \
examples/llava/llava.h \
examples/llava/clip.cpp \
examples/llava/clip.h \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) $< $(filter-out %.h $<,$^) -o $@ $(LDFLAGS) -Wno-cast-qual
llama-minicpmv-cli: examples/llava/minicpmv-cli.cpp \
examples/llava/llava.cpp \
examples/llava/llava.h \
examples/llava/clip.cpp \
examples/llava/clip.h \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) $< $(filter-out %.h $<,$^) -o $@ $(LDFLAGS) -Wno-cast-qual
ifeq ($(UNAME_S),Darwin)
swift: examples/batched.swift
(cd examples/batched.swift; make build)
endif
common/build-info.cpp: $(wildcard .git/index) scripts/build-info.sh
@sh scripts/build-info.sh "$(CC)" > $@.tmp
@if ! cmp -s $@.tmp $@; then \
mv $@.tmp $@; \
else \
rm $@.tmp; \
fi
common/build-info.o: common/build-info.cpp
$(CXX) $(CXXFLAGS) -c $(filter-out %.h,$^) -o $@
#
# Tests
#
tests: $(TEST_TARGETS)
llama-benchmark-matmult: examples/benchmark/benchmark-matmult.cpp \
$(OBJ_GGML) common/build-info.o
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
run-benchmark-matmult: llama-benchmark-matmult
./$@
.PHONY: run-benchmark-matmult swift
tests/test-llama-grammar: tests/test-llama-grammar.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-grammar-parser: tests/test-grammar-parser.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-grammar-integration: tests/test-grammar-integration.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-double-float: tests/test-double-float.cpp
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-json-schema-to-grammar: tests/test-json-schema-to-grammar.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -Iexamples/server -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-grad0: tests/test-grad0.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-opt: tests/test-opt.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-quantize-fns: tests/test-quantize-fns.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-quantize-perf: tests/test-quantize-perf.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-sampling: tests/test-sampling.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-tokenizer-0: tests/test-tokenizer-0.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-tokenizer-1-bpe: tests/test-tokenizer-1-bpe.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-tokenizer-1-spm: tests/test-tokenizer-1-spm.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-rope: tests/test-rope.cpp ggml/src/ggml.o \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-c.o: tests/test-c.c include/llama.h
$(CC) $(CFLAGS) -c $(filter-out %.h,$^) -o $@
tests/test-backend-ops: tests/test-backend-ops.cpp \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-model-load-cancel: tests/test-model-load-cancel.cpp tests/get-model.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-autorelease: tests/test-autorelease.cpp tests/get-model.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
tests/test-chat-template: tests/test-chat-template.cpp \
$(OBJ_ALL)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out %.h $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
#
# PoCs
#
llama-vdot: pocs/vdot/vdot.cpp ggml/src/ggml.o \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
llama-q8dot: pocs/vdot/q8dot.cpp ggml/src/ggml.o \
$(OBJ_GGML)
$(CXX) $(CXXFLAGS) -c $< -o $(call GET_OBJ_FILE, $<)
$(CXX) $(CXXFLAGS) $(filter-out $<,$^) $(call GET_OBJ_FILE, $<) -o $@ $(LDFLAGS)
#
# Deprecated binaries that we want to keep around long enough for people to migrate to the new filenames, then these can be removed.
#
# Mark legacy binary targets as .PHONY so that they are always checked.
.PHONY: main quantize perplexity embedding server
# Define the object file target
examples/deprecation-warning/deprecation-warning.o: examples/deprecation-warning/deprecation-warning.cpp
$(CXX) $(CXXFLAGS) -c $< -o $@
# NOTE: We currently will always build the deprecation-warning `main` and `server` binaries to help users migrate.
# Eventually we will want to remove these target from building all the time.
main: examples/deprecation-warning/deprecation-warning.o
$(CXX) $(CXXFLAGS) $< -o $@ $(LDFLAGS)
@echo "NOTICE: The 'main' binary is deprecated. Please use 'llama-cli' instead."
server: examples/deprecation-warning/deprecation-warning.o
$(CXX) $(CXXFLAGS) $< -o $@ $(LDFLAGS)
@echo "NOTICE: The 'server' binary is deprecated. Please use 'llama-server' instead."
quantize: examples/deprecation-warning/deprecation-warning.o
ifneq (,$(wildcard quantize))
$(CXX) $(CXXFLAGS) $< -o $@ $(LDFLAGS)
@echo "#########"
@echo "WARNING: The 'quantize' binary is deprecated. Please use 'llama-quantize' instead."
@echo " Remove the 'quantize' binary to remove this warning."
@echo "#########"
endif
perplexity: examples/deprecation-warning/deprecation-warning.o
ifneq (,$(wildcard perplexity))
$(CXX) $(CXXFLAGS) $< -o $@ $(LDFLAGS)
@echo "#########"
@echo "WARNING: The 'perplexity' binary is deprecated. Please use 'llama-perplexity' instead."
@echo " Remove the 'perplexity' binary to remove this warning."
@echo "#########"
endif
embedding: examples/deprecation-warning/deprecation-warning.o
ifneq (,$(wildcard embedding))
$(CXX) $(CXXFLAGS) $< -o $@ $(LDFLAGS)
@echo "#########"
@echo "WARNING: The 'embedding' binary is deprecated. Please use 'llama-embedding' instead."
@echo " Remove the 'embedding' binary to remove this warning."
@echo "#########"
endif
Reference in New Issue View Git Blame Copy Permalink