Jeff Bolz [Wed, 9 Apr 2025 05:25:08 +0000 (00:25 -0500)]
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.
server : webui : Improve Chat Input with Auto-Sizing Textarea (#12785)
* Update ChatScreen.tsx
* useAutosizeTextarea.ts
useAutosizeTextarea to encapsulate the logic.
* Implement responsive auto-sizing chat textarea
Replaces the manual textarea resizing with an automatic height adjustment based on content.
- `useChatTextarea` hook to manage textarea state and auto-sizing logic via refs, preserving the optimization
- Textarea now grows vertically up to a maximum height (`lg:max-h-48`) on large screens (lg breakpoint and up).
- Disables auto-sizing and enables manual vertical resizing (`resize-vertical`) on smaller screens for better mobile usability.
- Aligns the "Send" button to the bottom of the textarea (`items-end`) for consistent positioning during resize.
* -update compressed index.html.gz after npm run build
-refactor: replace OptimizedTextareaValue with AutosizeTextareaApi in VSCode context hook
* chore: normalize line endings to LF
refactor: AutosizeTextareaApi -> chatTextareaApi
Jeff Bolz [Sun, 6 Apr 2025 08:47:13 +0000 (03:47 -0500)]
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.
Add support for custom huggingface endpoints via HF_ENDPOINT environment variable
You can now specify a custom huggingface endpoint using the HF_ENDPOINT environment variable when using the --hf-repo flag, which works similarly to huggingface-cli's endpoint configuration.
Example usage:
HF_ENDPOINT=https://hf-mirror.com/ ./bin/llama-cli --hf-repo Qwen/Qwen1.5-0.5B-Chat-GGUF --hf-file qwen1_5-0_5b-chat-q2_k.gguf -p "The meaning to life and the universe is"
The trailing slash in the URL is optional:
HF_ENDPOINT=https://hf-mirror.com ./bin/llama-cli --hf-repo Qwen/Qwen1.5-0.5B-Chat-GGUF --hf-file qwen1_5-0_5b-chat-q2_k.gguf -p "The meaning to life and the universe is"
* Update common/arg.cpp
readability Improvement
Co-authored-by: Xuan-Son Nguyen <redacted>
* Apply suggestions from code review
Jeff Bolz [Fri, 4 Apr 2025 05:54:35 +0000 (00:54 -0500)]
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.
CUDA: Prefer vector flash decoding kernel for Gemma models (#12738)
* Prefer vector flash decoding kernel for Gemma models
Vector flash decoding kernel was not being picked for models with head dimension 256. Gemma models are in this category.
Removing this limit improves e2e performance by upto 12% in gen phase throughput for Gemm models.
* Update ggml/src/ggml-cuda/fattn.cu
Co-authored-by: Johannes Gäßler <redacted>
---------
Alan Gray [Thu, 3 Apr 2025 01:31:15 +0000 (02:31 +0100)]
Simplify and improve CUDA graphs through use of indirect copy pointers (#9017)
* CUDA: Simplify and improve CUDA graphs through use of indirect copy pointers
Previously there was complexity in the CUDA graphs implementation due
frequently changing parameters to copy kernels associated with K and V
cache pointers. This patch simplifies by using indirection to avoid
such parameters frequently changing, avoiding the need for frequent
graph updates.
Jeff Bolz [Wed, 2 Apr 2025 19:25:08 +0000 (14:25 -0500)]
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.
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.
Daniel Bevenius [Mon, 31 Mar 2025 16:40:56 +0000 (18:40 +0200)]
vocab : add special infill tokens for CodeLlama (#11850)
* vocab : add special infill tokens for CodeLlama
The commit adds the following special tokens for CodeLlama infill:
- `▁<PRE>`
- `▁<SUF>`
- `▁<MID>`
The motivation for this is that currently the infill example uses
CodeLlama as a suggested model. But when using this model the following
error is generated:
```console
/llama.cpp-debug/examples/infill/infill.cpp:165: GGML_ASSERT(llama_vocab_fim_pre(vocab) >= 0) failed
Could not attach to process. If your uid matches the uid of the target
process, check the setting of /proc/sys/kernel/yama/ptrace_scope, or try
again as the root user. For more details, see /etc/sysctl.d/10-ptrace.conf
ptrace: Operation not permitted.
No stack.
The program is not being run.
305251 Aborted (core dumped)
./build/bin/llama-infill -t 10 -ngl 0 -m models/codellama-13b.Q5_K_S.gguf \
-c 4096 --temp 0.7 --repeat_penalty 1.1 -n 20 \
--in-prefix "def helloworld():\n print(\"hell" \
--in-suffix "\n print(\"goodbye world\")\n "
```
* squash! vocab : add special infill tokens for CodeLlama
marcoStocchi [Mon, 31 Mar 2025 08:20:30 +0000 (10:20 +0200)]
tts : remove printfs (#12640)
* tts.cpp : llama tokens console output is done using LOG_INF instead of printf(). Therefore the options '--log-disable' and '--log-file' have now uniform impact on all output.
Daniel Bevenius [Mon, 24 Mar 2025 08:53:38 +0000 (09:53 +0100)]
ggml : add logging for native build options/vars (whisper/2935)
This commit adds debug level logging for the native build options and
variables to ggml/CMakeLists.txt.
The motivation for this is that it can be useful to see the effective
result of `GGML_NATIVE`, `GGML_NATIVE_DEFAULT`, and `INS_ENB` for a
cmake build. I've found myself adding similar logging a few times now,
so I thought it might be a good idea to add this.
Example output, specifying `-DCMAKE_MESSAGE_LOG_LEVEL=DEBUG` when
running cmake produces the following output:
```console
-- GGML_NATIVE : OFF
-- GGML_NATIVE_DEFAULT : OFF
-- INS_ENB : OFF
```
overview / pkg / ggml / sources / llama.cpp / log