ci : disable mmap

.github/ISSUE_TEMPLATE/010-bug-compilation.yml

@@ -8,8 +8,7 @@ body:
       value: >
         Thanks for taking the time to fill out this bug report!
         This issue template is intended for bug reports where the compilation of llama.cpp fails.
-        Before opening an issue, please confirm that the compilation still fails
-        after recreating the CMake build directory and with `-DGGML_CCACHE=OFF`.
+        Before opening an issue, please confirm that the compilation still fails with `-DGGML_CCACHE=OFF`.
         If the compilation succeeds with ccache disabled you should be able to permanently fix the issue
         by clearing `~/.cache/ccache` (on Linux).
   - type: textarea

.github/ISSUE_TEMPLATE/011-bug-results.yml

@@ -98,18 +98,7 @@ body:
       label: Relevant log output
       description: >
           Please copy and paste any relevant log output, including the command that you entered and any generated text.
-          For very long logs (thousands of lines), preferably upload them as files instead.
-          On Linux you can redirect console output into a file by appending ` > llama.log 2>&1` to your command.
-      value: |
-        <details>
-        <summary>Logs</summary>
-        <!-- Copy-pasted short logs go into the "console" area here -->
-
-        ```console
-
-        ```
-        </details>
-
-        <!-- Long logs that you upload as files go here, outside the "console" area -->
+          This will be automatically formatted into code, so no need for backticks.
+      render: shell
     validations:
       required: true

.github/ISSUE_TEMPLATE/019-bug-misc.yml

@@ -85,19 +85,8 @@ body:
       label: Relevant log output
       description: >
           If applicable, please copy and paste any relevant log output, including any generated text.
+          This will be automatically formatted into code, so no need for backticks.
           If you are encountering problems specifically with the `llama_params_fit` module, always upload `--verbose` logs as well.
-          For very long logs (thousands of lines), please upload them as files instead.
-          On Linux you can redirect console output into a file by appending ` > llama.log 2>&1` to your command.
-      value: |
-        <details>
-        <summary>Logs</summary>
-        <!-- Copy-pasted short logs go into the "console" area here -->
-
-        ```console
-
-        ```
-        </details>
-
-        <!-- Long logs that you upload as files go here, outside the "console" area -->
+      render: shell
     validations:
       required: false

ci/run.sh

@@ -47,6 +47,8 @@ SRC=`pwd`
 
 CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=${LLAMA_FATAL_WARNINGS:-ON} -DLLAMA_CURL=ON -DGGML_SCHED_NO_REALLOC=ON"
 
+export LLAMA_ARG_NO_MMAP=1
+
 if [ ! -z ${GG_BUILD_METAL} ]; then
     CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON"
 fi

ggml/CMakeLists.txt

@@ -430,22 +430,10 @@ if (MSVC)
     configure_msvc_target(ggml-cpu-x64)
     configure_msvc_target(ggml-cpu-sse42)
     configure_msvc_target(ggml-cpu-sandybridge)
-    # __FMA__ and __F16C__ are not defined in MSVC, however they are implied with AVX2/AVX512
-    # skipping            ggml-cpu-ivybridge
-    # skipping            ggml-cpu-piledriver
     configure_msvc_target(ggml-cpu-haswell)
     configure_msvc_target(ggml-cpu-skylakex)
-    configure_msvc_target(ggml-cpu-cannonlake)
-    configure_msvc_target(ggml-cpu-cascadelake)
     configure_msvc_target(ggml-cpu-icelake)
-    # MSVC 2022 doesn't support BF16 intrinsics without `/arch:AVX10.1` ?!
-    # https://learn.microsoft.com/en-us/cpp/intrinsics/x64-amd64-intrinsics-list?view=msvc-170
-    # https://learn.microsoft.com/en-us/cpp/build/reference/arch-x64?view=msvc-170
-    # skipping            ggml-cpu-cooperlake
-    # skipping            ggml-cpu-zen4
     configure_msvc_target(ggml-cpu-alderlake)
-    # MSVC doesn't support AMX
-    # skipping            ggml-cpu-sapphirerapids
 
     if (GGML_BUILD_EXAMPLES)
         configure_msvc_target(common-ggml)

ggml/src/CMakeLists.txt

@@ -357,29 +357,15 @@ if (GGML_CPU_ALL_VARIANTS)
     endif()
     if (GGML_SYSTEM_ARCH STREQUAL "x86")
         ggml_add_cpu_backend_variant(x64)
-        ggml_add_cpu_backend_variant(sse42              SSE42)
-        ggml_add_cpu_backend_variant(sandybridge        SSE42 AVX)
-        if (NOT MSVC)
-            # __FMA__ and __F16C__ are not defined in MSVC, however they are implied with AVX2/AVX512
-            ggml_add_cpu_backend_variant(ivybridge      SSE42 AVX F16C)
-            ggml_add_cpu_backend_variant(piledriver     SSE42 AVX F16C FMA)
-        endif()
-        ggml_add_cpu_backend_variant(haswell            SSE42 AVX F16C FMA AVX2 BMI2)
-        ggml_add_cpu_backend_variant(skylakex           SSE42 AVX F16C FMA AVX2 BMI2 AVX512)
-        ggml_add_cpu_backend_variant(cannonlake         SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI)
-        ggml_add_cpu_backend_variant(cascadelake        SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VNNI)
-        ggml_add_cpu_backend_variant(icelake            SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI)
-        if (NOT MSVC)
-            # MSVC 2022 doesn't support BF16 intrinsics without `/arch:AVX10.1` ?!
-            # https://learn.microsoft.com/en-us/cpp/intrinsics/x64-amd64-intrinsics-list?view=msvc-170
-            # https://learn.microsoft.com/en-us/cpp/build/reference/arch-x64?view=msvc-170
-            ggml_add_cpu_backend_variant(cooperlake     SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VNNI AVX512_BF16)
-            ggml_add_cpu_backend_variant(zen4           SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16)
-        endif()
-        ggml_add_cpu_backend_variant(alderlake          SSE42 AVX F16C FMA AVX2 BMI2 AVX_VNNI)
+        ggml_add_cpu_backend_variant(sse42        SSE42)
+        ggml_add_cpu_backend_variant(sandybridge  SSE42 AVX)
+        ggml_add_cpu_backend_variant(haswell      SSE42 AVX F16C AVX2 BMI2 FMA)
+        ggml_add_cpu_backend_variant(skylakex     SSE42 AVX F16C AVX2 BMI2 FMA AVX512)
+        ggml_add_cpu_backend_variant(icelake      SSE42 AVX F16C AVX2 BMI2 FMA AVX512 AVX512_VBMI AVX512_VNNI)
+        ggml_add_cpu_backend_variant(alderlake    SSE42 AVX F16C AVX2 BMI2 FMA AVX_VNNI)
         if (NOT MSVC)
             # MSVC doesn't support AMX
-            ggml_add_cpu_backend_variant(sapphirerapids SSE42 AVX F16C FMA AVX2 BMI2 AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16 AMX_TILE AMX_INT8)
+            ggml_add_cpu_backend_variant(sapphirerapids SSE42 AVX F16C AVX2 BMI2 FMA AVX512 AVX512_VBMI AVX512_VNNI AVX512_BF16 AMX_TILE AMX_INT8)
         endif()
     elseif(GGML_SYSTEM_ARCH STREQUAL "ARM")
         if (CMAKE_SYSTEM_NAME MATCHES "Linux")

ggml/src/ggml-cpu/ggml-cpu-impl.h

@@ -328,7 +328,7 @@ inline static int32x4_t ggml_vdotq_s32(int32x4_t acc, int8x16_t a, int8x16_t b)
 
 #if defined(_MSC_VER) || defined(__MINGW32__)
 #include <intrin.h>
-#elif defined(__SSE__) || defined(__SSE3__) || defined(__SSSE3__) || defined(__AVX__) || defined(__F16C__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__AVX512BF16__)
+#elif defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) || defined(__SSE3__) || defined(__SSE__)
 #include <immintrin.h>
 #endif
 

ggml/src/ggml-cpu/simd-mappings.h

@@ -14,6 +14,10 @@
 #include <arm_neon.h>
 #endif
 
+#if defined(__F16C__)
+#include <immintrin.h>
+#endif
+
 #if defined(__riscv_v_intrinsic)
 #include <riscv_vector.h>
 #endif

ggml/src/ggml-impl.h

@@ -24,6 +24,10 @@
 #include <arm_neon.h>
 #endif
 
+#if defined(__F16C__)
+#include <immintrin.h>
+#endif
+
 #ifdef __cplusplus
 extern "C" {
 #endif

src/llama.cpp

@@ -512,9 +512,6 @@ static void llama_params_fit_impl(
             if (mem_high[id] > targets[id]) {
                 assert(ngl_per_device_high[id].n_layer > ngl_per_device[id].n_layer);
                 uint32_t delta = ngl_per_device_high[id].n_layer - ngl_per_device[id].n_layer;
-                if (hp_nex > 0 && size_t(id) == nd - 1) {
-                    delta--;
-                }
                 LLAMA_LOG_DEBUG("%s: start filling device %" PRIu32 ", delta=%" PRIu32 "\n", __func__, id, delta);
                 while (delta > 1) {
                     uint32_t step_size = int64_t(delta) * (targets[id] - mem[id]) / (mem_high[id] - mem[id]);

tools/mtmd/models/models.h

@@ -2,11 +2,6 @@
 
 #include "../clip-graph.h"
 
-/*
- * IMPORTANT: The mtmd module does NOT accept pull requests that are fully or predominantly AI-generated.
- * We encourage human contributors to ensure the quality and reliability of the codebase.
- */
-
 struct clip_graph_siglip : clip_graph {
     clip_graph_siglip(clip_ctx * ctx, const clip_image_f32 & img) : clip_graph(ctx, img) {}
     ggml_cgraph * build() override;

tools/mtmd/mtmd.h

@@ -27,9 +27,6 @@
  * - Make sure the C API is aligned with the libllama C API (as in llama.h)
  * - Do not include model name (e.g., qwen, gemma) in the API, use generic terms instead
  * - Keep the API minimal, do not expose internal details unless necessary
- *
- * IMPORTANT: The mtmd module does NOT accept pull requests that are fully or predominantly AI-generated.
- * We encourage human contributors to ensure the quality and reliability of the codebase.
  */
 
 #ifdef LLAMA_SHARED