Fix: Compile failure due to Microsoft STL breaking change (#11836)

* Bug fix for clamp_f32

When using tensors larger than 1d clamp operation does not work due to the restriction of returning if ith is not 0.

* Bug fix for clamp_f32

* Bug fix for clamp_f32

common/common.h

@@ -424,13 +424,13 @@ bool set_process_priority(enum ggml_sched_priority prio);
 //
 
 #ifdef __GNUC__
-#ifdef __MINGW32__
-#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
+#    if defined(__MINGW32__) && !defined(__clang__)
+#        define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
+#    else
+#        define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
+#    endif
 #else
-#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
-#endif
-#else
-#define LLAMA_COMMON_ATTRIBUTE_FORMAT(...)
+#    define LLAMA_COMMON_ATTRIBUTE_FORMAT(...)
 #endif
 
 LLAMA_COMMON_ATTRIBUTE_FORMAT(1, 2)

common/log.cpp

@@ -1,5 +1,6 @@
 #include "log.h"
 
+#include <chrono>
 #include <condition_variable>
 #include <cstdarg>
 #include <cstdio>

common/log.h

@@ -15,7 +15,7 @@
 
 #ifndef __GNUC__
 #    define LOG_ATTRIBUTE_FORMAT(...)
-#elif defined(__MINGW32__)
+#elif defined(__MINGW32__) && !defined(__clang__)
 #    define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
 #else
 #    define LOG_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))

examples/imatrix/imatrix.cpp

@@ -3,6 +3,7 @@
 #include "log.h"
 #include "llama.h"
 
+#include <chrono>
 #include <cmath>
 #include <cstdio>
 #include <cstring>

examples/perplexity/perplexity.cpp

@@ -3,6 +3,7 @@
 #include "log.h"
 #include "llama.h"
 
+#include <chrono>
 #include <algorithm>
 #include <array>
 #include <atomic>

ggml/include/ggml.h

@@ -198,7 +198,7 @@
 
 #ifndef __GNUC__
 #    define GGML_ATTRIBUTE_FORMAT(...)
-#elif defined(__MINGW32__)
+#elif defined(__MINGW32__) && !defined(__clang__)
 #    define GGML_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
 #else
 #    define GGML_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))

ggml/src/ggml-cpu/ggml-cpu.c

@@ -9074,10 +9074,6 @@ static void ggml_compute_forward_clamp_f32(
 
     const struct ggml_tensor * src0 = dst->src[0];
 
-    if (params->ith != 0) {
-        return;
-    }
-
     float min;
     float max;
     memcpy(&min, (float *) dst->op_params + 0, sizeof(float));

ggml/src/ggml-cpu/ggml-cpu.cpp

@@ -534,9 +534,6 @@ static ggml_backend_feature * ggml_backend_cpu_get_features(ggml_backend_reg_t r
         if (ggml_cpu_has_dotprod()) {
             features.push_back({ "DOTPROD", "1" });
         }
-        if (ggml_cpu_has_matmul_int8()) {
-            features.push_back({ "MATMUL_INT8", "1" });
-        }
         if (ggml_cpu_get_sve_cnt() > 0) {
             static std::string sve_cnt = std::to_string(ggml_cpu_get_sve_cnt());
             features.push_back({ "SVE_CNT", sve_cnt.c_str() });

ggml/src/ggml-cuda/ggml-cuda.cu

@@ -178,11 +178,11 @@ static ggml_cuda_device_info ggml_cuda_init() {
         int major_version = 0;
         size_t version_length = 0;
         if (rocblas_get_version_string_size(&version_length) == rocblas_status_success) {
-            std::string version(version_length, '\0');
+            std::vector<char> version(version_length+1, '\0');
             if (rocblas_get_version_string(version.data(), version.size()) == rocblas_status_success) {
-                version.resize(::strlen(version.c_str()));
+                version.resize(::strlen(version.data()));
                 int parsed_value = 0;
-                if (std::from_chars(version.c_str(), version.c_str() + version.length(), parsed_value).ec == std::errc()) {
+                if (std::from_chars(version.data(), version.data() + version.size(), parsed_value).ec == std::errc()) {
                     major_version = parsed_value;
                 }
             }

scripts/sync-ggml.last

@@ -1 +1 @@
-08b538031f7f944e84f472483ef5d26bf5190ead
+98a61a0d0b43cba06c3ac1c603813639552a0701

src/llama-impl.h

@@ -6,13 +6,13 @@
 #include <vector>
 
 #ifdef __GNUC__
-#ifdef __MINGW32__
-#define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
+#    if defined(__MINGW32__) && !defined(__clang__)
+#        define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(gnu_printf, __VA_ARGS__)))
+#    else
+#        define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
+#    endif
 #else
-#define LLAMA_ATTRIBUTE_FORMAT(...) __attribute__((format(printf, __VA_ARGS__)))
-#endif
-#else
-#define LLAMA_ATTRIBUTE_FORMAT(...)
+#    define LLAMA_ATTRIBUTE_FORMAT(...)
 #endif
 
 //

tests/test-gguf.cpp

@@ -697,8 +697,8 @@ static std::pair<int, int> test_handcrafted_file(const unsigned int seed) {
 
 #ifdef _WIN32
         if (!file) {
-            printf("%s: failed to create tmpfile(), needs elevated privileges on Windows");
-            printf("%s: skipping tests");
+            printf("failed to create tmpfile(), needs elevated privileges on Windows");
+            printf("skipping tests");
             continue;
         }
 #else
@@ -1086,8 +1086,8 @@ static std::pair<int, int> test_roundtrip(ggml_backend_dev_t dev, const unsigned
 
 #ifdef _WIN32
     if (!file) {
-        printf("%s: failed to create tmpfile(), needs elevated privileges on Windows");
-        printf("%s: skipping tests");
+        printf("failed to create tmpfile(), needs elevated privileges on Windows");
+        printf("skipping tests");
         return std::make_pair(0, 0);
     }
 #else