try to add ios-xcode-build

ggerganov · Oct 26, 2024 · 60ece70 · 60ece70
1 parent 05c55e5
commit 60ece70
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 21 deletions.
diff --git a/Package.swift b/Package.swift
@@ -14,6 +14,7 @@ var sources = [
     "ggml/src/ggml-backend.cpp",
     "ggml/src/ggml-quants.c",
     "ggml/src/ggml-aarch64.c",
+    "ggml/src/ggml-fp8.cpp",
 ]
 
 var resources: [Resource] = []
@@ -75,5 +76,5 @@ let package = Package(
             linkerSettings: linkerSettings
         )
     ],
-    cxxLanguageStandard: .cxx11
+    cxxLanguageStandard: .cxx17
 )
diff --git a/ggml/src/ggml-common.h b/ggml/src/ggml-common.h
@@ -6,7 +6,8 @@
 typedef uint16_t ggml_half;
 typedef uint32_t ggml_half2;
 
-#define GGML_COMMON_AGGR
+#define GGML_COMMON_AGGR_U
+#define GGML_COMMON_AGGR_S
 
 #define GGML_COMMON_DECL
 #elif defined(GGML_COMMON_DECL_CPP)
@@ -15,7 +16,8 @@ typedef uint32_t ggml_half2;
 typedef uint16_t ggml_half;
 typedef uint32_t ggml_half2;
 
-#define GGML_COMMON_AGGR data
+#define GGML_COMMON_AGGR_U data
+#define GGML_COMMON_AGGR_S data
 
 #define GGML_COMMON_DECL
 #elif defined(GGML_COMMON_DECL_METAL)
@@ -24,7 +26,8 @@ typedef uint32_t ggml_half2;
 typedef half  ggml_half;
 typedef half2 ggml_half2;
 
-#define GGML_COMMON_AGGR
+#define GGML_COMMON_AGGR_U
+#define GGML_COMMON_AGGR_S
 
 #define GGML_COMMON_DECL
 #elif defined(GGML_COMMON_DECL_CUDA)
@@ -38,7 +41,8 @@ typedef half2 ggml_half2;
 typedef half  ggml_half;
 typedef half2 ggml_half2;
 
-#define GGML_COMMON_AGGR data
+#define GGML_COMMON_AGGR_U
+#define GGML_COMMON_AGGR_S data
 
 #define GGML_COMMON_DECL
 #elif defined(GGML_COMMON_DECL_HIP)
@@ -48,7 +52,8 @@ typedef half2 ggml_half2;
 typedef half  ggml_half;
 typedef half2 ggml_half2;
 
-#define GGML_COMMON_AGGR data
+#define GGML_COMMON_AGGR_U
+#define GGML_COMMON_AGGR_S data
 
 #define GGML_COMMON_DECL
 #elif defined(GGML_COMMON_DECL_SYCL)
@@ -58,7 +63,8 @@ typedef half2 ggml_half2;
 typedef sycl::half  ggml_half;
 typedef sycl::half2 ggml_half2;
 
-#define GGML_COMMON_AGGR data
+#define GGML_COMMON_AGGR_U
+#define GGML_COMMON_AGGR_S data
 
 #define GGML_COMMON_DECL
 #endif
@@ -163,9 +169,9 @@ typedef struct {
         struct {
             ggml_half d; // delta
             ggml_half m; // min
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 dm;
-    };
+    } GGML_COMMON_AGGR_U;
     uint8_t qs[QK4_1 / 2]; // nibbles / quants
 } block_q4_1;
 static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_half) + QK4_1 / 2, "wrong q4_1 block size/padding");
@@ -184,9 +190,9 @@ typedef struct {
         struct {
             ggml_half d; // delta
             ggml_half m; // min
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 dm;
-    };
+    } GGML_COMMON_AGGR_U;
     uint8_t qh[4];         // 5-th bit of quants
     uint8_t qs[QK5_1 / 2]; // nibbles / quants
 } block_q5_1;
@@ -205,9 +211,9 @@ typedef struct {
         struct {
             ggml_half d; // delta
             ggml_half s; // d * sum(qs[i])
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 ds;
-    };
+    } GGML_COMMON_AGGR_U;
     int8_t qs[QK8_1]; // quants
 } block_q8_1;
 static_assert(sizeof(block_q8_1) == 2*sizeof(ggml_half) + QK8_1, "wrong q8_1 block size/padding");
@@ -270,9 +276,9 @@ typedef struct {
         struct {
             ggml_half d;    // super-block scale for quantized scales
             ggml_half dmin; // super-block scale for quantized mins
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 dm;
-    };
+    } GGML_COMMON_AGGR_U;
 } block_q2_K;
 static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_half) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding");
 
@@ -297,9 +303,9 @@ typedef struct {
         struct {
             ggml_half d;    // super-block scale for quantized scales
             ggml_half dmin; // super-block scale for quantized mins
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 dm;
-    };
+    } GGML_COMMON_AGGR_U;
     uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits
     uint8_t qs[QK_K/2];           // 4--bit quants
 } block_q4_K;
@@ -314,9 +320,9 @@ typedef struct {
         struct {
             ggml_half d;    // super-block scale for quantized scales
             ggml_half dmin; // super-block scale for quantized mins
-        } GGML_COMMON_AGGR;
+        } GGML_COMMON_AGGR_S;
         ggml_half2 dm;
-    };
+    } GGML_COMMON_AGGR_U;
     uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits
     uint8_t qh[QK_K/8];           // quants, high bit
     uint8_t qs[QK_K/2];           // quants, low 4 bits

diff --git a/ggml/src/ggml-fp8.cpp b/ggml/src/ggml-fp8.cpp
@@ -151,9 +151,9 @@ static inline void conv(const float* x, bloc_fp8<E, QK>* y, int64_t size) {
     const auto qk_size = size / QK;
     for (int64_t q=0; q<qk_size; ++q) {
         float m = 0;
-        // @ voir si c'est lui qui pose probleme et si c'est sur toutes les target
 #ifdef GGML_USE_OPENMP_SIMD
-        #pragma omp simd reduction(max:m)
+        // not work on macos and warn.
+        // #pragma omp simd reduction(max:m)
 #endif
         for (int64_t i=0; i<QK; i++) {
             m = std::max(std::abs(x[q*QK+i]),m);