Commit Graph

1988 Commits

Author SHA1 Message Date
Georgi Gerganov
48f421de23 cmake : enable warnings in llama (llama/10474)
* cmake : enable warnings in llama

ggml-ci

* cmake : add llama_get_flags and respect LLAMA_FATAL_WARNINGS

* cmake : get_flags -> ggml_get_flags

* speculative-simple : fix warnings

* cmake : reuse ggml_get_flags

ggml-ci

* speculative-simple : fix compile warning

ggml-ci
2024-12-08 20:14:35 +02:00
Charles Xu
e7afb2b991 ggml-cpu: cmake add arm64 cpu feature check for macos (llama/10487)
* ggml-cpu: cmake add arm64 cpu feature check for macos

* use vmmlaq_s32 for compile option i8mm check
2024-12-08 20:14:35 +02:00
Shanshan Shen
9a5ef7b169 CANN: Improve the Inferencing Performance for Ascend NPU Device (llama/10454)
* improve inferencing performance for ascend npu.

Co-authored-by: Frank Mai <thxCode@thxcode0824@gmail.com>

* some modification after review

* some modifications after review

* restore some modifications

* restore some modifications

---------

Co-authored-by: shanshan shen <shanshanshen333@gmail.com>
Co-authored-by: Frank Mai <thxCode@thxcode0824@gmail.com>
2024-12-08 20:14:35 +02:00
Chenguang Li
453cc0fcf1 CANN: RoPE and CANCAT operator optimization (llama/10488)
Co-authored-by: noemotiovon <noemotiovon@gmail.com>
2024-12-08 20:14:35 +02:00
Junil Kim
78dfec6bc5 vulkan: Fix a vulkan-shaders-gen arugment parsing error (llama/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.
2024-12-08 20:14:35 +02:00
Georgi Gerganov
f6d518fc4c metal : enable mat-vec kernels for bs <= 4 (llama/10491) 2024-12-08 20:14:35 +02:00
Diego Devesa
ac33379a35 llama : accept a list of devices to use to offload a model (llama/10497)
* llama : accept a list of devices to use to offload a model

* accept `--dev none` to completely disable offloading

* fix dev list with dl backends

* rename env parameter to LLAMA_ARG_DEVICE for consistency
2024-12-08 20:14:35 +02:00
Diego Devesa
77e3e4a090 ggml : add support for dynamic loading of backends (llama/10469)
* ggml : add support for dynamic loading of backends

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2024-12-08 20:14:35 +02:00
Georgi Gerganov
b840bb09be metal : minor code formatting 2024-12-08 20:14:35 +02:00
Diego Devesa
8b1c1c30a7 ggml : do not use ARM features not included in the build (llama/10457) 2024-12-08 20:14:35 +02:00
leo-pony
4b81335f75 CANN: Support Ascend310P to accelerate F32 and F16 Model (llama/10216)
* CANN Support Ascend310P to accelerate F32 and F16 Model

* Add compile option soc type macro ASCEND_310P to ggml-cann lib

* Remove unused code

* Remove the ascend soc_type hard code compile option in CMakelist.txt
2024-12-08 20:14:35 +02:00
Diego Devesa
2a4b5c9d7e cuda : optimize argmax (llama/10441)
* cuda : optimize argmax

* remove unused parameter

ggml-ci

* fixup : use full warps

ggml-ci

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* fix ub

* ggml : check ne00 <= INT32_MAX in argmax and argsort

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
2024-12-08 20:14:35 +02:00
Jeff Bolz
04662748aa vulkan: predicate max operation in soft_max shaders/soft_max (llama/10437)
Fixes #10434
2024-12-08 20:14:35 +02:00
Jeff Bolz
a117279e13 vulkan: copy iq4_nl LUT into shared memory (llama/10409) 2024-12-08 20:14:35 +02:00
Jeff Bolz
bbb292ed38 vulkan: further optimize mul_mat_vec using larger loads (llama/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.
2024-12-08 20:14:35 +02:00
haopeng
95e8901e71 add cmake rvv support (llama/10411) 2024-12-08 20:14:35 +02:00
mahorozte
4af9626702 CUDA: remove unnecessary warp reduce in FA (ggml/1032)
* kqmax_new_j in every thread within warp is same after operate at line 199,this reduce can be omit

* same problem in vec32

---------

Co-authored-by: ZhaoXiaoYu <zhao.xiaoyu@zte.com.cn>
2024-12-08 20:14:35 +02:00
PAB
c52d1035de feat: add GGML_UNARY_OP_ARGMAX Metal kernel (ggml/1019)
* implemented argmax kernel

* tpig -> tgpig

* change to strides

* contiguous assertions

* kernel working and tested

* argmax simd parallel implementation

* added 2 new tests for argmax in test-backend-ops

* cosmit

* added 3 tests cases for perf eval

* add test_argmax in make_test_cases_perf

* Update test-backend-ops.cpp

Co-authored-by: Diego Devesa <slarengh@gmail.com>

---------

Co-authored-by: Diego Devesa <slarengh@gmail.com>
2024-12-08 20:14:35 +02:00
PAB
5773a14980 metal : add GGML_OP_CONV_TRANSPOSE_1D kernels (ggml/1026)
* wip

* wip implementation f32

* kernel conv transpose 1d f32 working

* initial commit
2024-12-08 20:14:35 +02:00
Frankie Robertson
6939147c47 Do not include arm_neon.h when compiling CUDA code (ggml/1028) 2024-12-08 20:14:35 +02:00
Johannes Gäßler
98f9916c9f ggml-opt: fix data corruption (ggml/1022) 2024-12-08 20:14:35 +02:00
KITAITI Makoto
021eef1000
ruby : Add low-level methods to transcribe (#2585)
* Add tests for Whisper::Context#full

* Add Whisper::Context#full

* Add tests for Whisper::Error

* Add document of Whisper::Context#full [skip ci]

* Add additional signature for Whisper::Context#full

* Add description to Whisper::Context#full

* Add test for Whisper::Context#full_parallel

* Add Whisper::Context#full_parallel

* Hide Whisper's instance methods from Ruby code

* Add class to test MemoryView

* Build test class before running test

* Add test for MemoryView

* Make Whisper::Context#full and #full_parallel accept MemoryView

* Use Ruby 3.1 on CI

* Add comment on samples data type

* Update README

* Update README

* Remove unused code
2024-11-28 10:33:07 +02:00
Michael Rienstra
a9d06ce151
models : add q8_0 models to download-ggml-model.sh (#2589) 2024-11-28 10:31:54 +02:00
KITAITI Makoto
8c6a9b8bb6
ruby : Follow source tree change (#2580)
* Follow whisper.cpp source tree change

* Update whispercpp.gemspec

* Follow whisper.cpp log level change

* Fix paths in GitHub workflow for Ruby bindings

* Use GitHub workflow setting for dependency definition

* Use ternary operator
2024-11-21 17:04:29 +02:00
Georgi Gerganov
37c88027e1 whisper : use backend registry (#0) 2024-11-20 21:00:08 +02:00
slaren
9db070a3c5 ggml/sched : do not skip views in pre-assignments 2024-11-20 21:00:08 +02:00
Georgi Gerganov
7fd8d9c220 whisper : adapt to new ggml (wip) 2024-11-20 21:00:08 +02:00
Georgi Gerganov
06e059b8f8 talk-llama : sync llama.cpp 2024-11-20 21:00:08 +02:00
Georgi Gerganov
c9f49d5f9d sync : ggml 2024-11-20 21:00:08 +02:00
Georgi Gerganov
f4c1d7df39 ggml : sync resolve (skip) (#0) 2024-11-20 21:00:08 +02:00
bandoti
339b8e559c Add required ggml-base and backend libs to cmake pkg (llama/10407) 2024-11-20 21:00:08 +02:00
Diego Devesa
5f6d6919b4 cuda : fix CUDA_FLAGS not being applied (llama/10403) 2024-11-20 21:00:08 +02:00
Romain Biessy
8ee767732f sycl : Add option to set the SYCL architecture for all targets (llama/10266)
* Add option to set the SYCL architecture for all targets
* Convert GGML_SYCL_HIP_TARGET to the more generic GGML_SYCL_ARCH option
* Document that setting GGML_SYCL_ARCH can improve the performance
2024-11-20 21:00:08 +02:00
Jeff Bolz
45f1f9144f vulkan: Optimize soft_max (llama/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.
2024-11-20 21:00:08 +02:00
Alberto Cabrera Pérez
53589c8f12 sycl: Revert MUL_MAT_OP support changes (llama/10385) 2024-11-20 21:00:08 +02:00
Diego Devesa
7ac2f17fac cuda : only use native when supported by cmake (llama/10389) 2024-11-20 21:00:08 +02:00
Jeff Bolz
48862c7b27 vulkan: remove use of null initializer (llama/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?
2024-11-20 21:00:08 +02:00
Plamen Minev
44f7d9f4e3 metal : fox offset integer overflows in im2col (ggml/1015)
-- While running StableDiffusion.cpp locally with Metal some offsets overflow and results in incorrect calculations
2024-11-20 21:00:08 +02:00
0cc4m
fd12302587 Vulkan: Fix device info output format specifiers (llama/10366)
* Vulkan: Fix device info output format specifiers

* Vulkan: Use zu printf specifier for size_t instead of ld
2024-11-20 21:00:08 +02:00
PAB
f80bef4630 metal : add GGML_UNARY_OP_ELU kernel (ggml/1018) 2024-11-20 21:00:08 +02:00
Johannes Gäßler
161b443514 CUDA: fix MMV kernel being used for FP16 src1 (llama/10357) 2024-11-20 21:00:08 +02:00
Johannes Gäßler
ef7fbe1c66 CMake: fix typo in comment [no ci] (llama/10360) 2024-11-20 21:00:08 +02:00
Diego Devesa
0879d3599e llama : only use default buffer types for the KV cache (llama/10358) 2024-11-20 21:00:08 +02:00
Georgi Gerganov
2a444dc5bd metal : refactor kernel args into structs (llama/10238)
* metal : add kernel arg structs (wip)

* metal : fattn args

ggml-ci

* metal : cont + avoid potential int overflow [no ci]

* metal : mul mat struct (wip)

* cont : mul mat vec

* cont : pass by reference

* cont : args is first argument

* cont : use char ptr

* cont : shmem style

* cont : thread counters style

* cont : mul mm id

ggml-ci

* cont : int safety + register optimizations

ggml-ci

* metal : GGML_OP_CONCAT

ggml-ci

* metal : GGML_OP_ADD, GGML_OP_SUB, GGML_OP_MUL, GGML_OP_DIV

* metal : GGML_OP_REPEAT

* metal : GGML_OP_CPY

* metal : GGML_OP_RMS_NORM

* metal : GGML_OP_NORM

* metal : add TODOs for rest of ops

* ggml : add ggml-metal-impl.h

ggml-ci
2024-11-20 21:00:08 +02:00
FirstTimeEZ
45cf1634dc ggml : fix undefined reference to 'getcpu' (llama/10354)
https://github.com/ggerganov/llama.cpp/issues/10352
2024-11-20 21:00:08 +02:00
Johannes Gäßler
dcb2922d1d CUDA: remove DMMV, consolidate F16 mult mat vec (llama/10318) 2024-11-20 21:00:08 +02:00
Johannes Gäßler
3c5c751174 CMake: default to -arch=native for CUDA build (llama/10320) 2024-11-20 21:00:08 +02:00
Diego Devesa
24ad19d0e9 ggml : fix possible buffer use after free in sched reserve (llama/9930) 2024-11-20 21:00:08 +02:00
Georgi Gerganov
bd574b05af ggml : inttypes.h -> cinttypes (llama/0)
ggml-ci
2024-11-20 21:00:08 +02:00
Georgi Gerganov
7e0eafcb1e ggml : adapt AMX to tensor->grad removal (llama/0)
ggml-ci
2024-11-20 21:00:08 +02:00