llamafile : improve moe prompt eval speed on cpu

This change introduces a llamafile_mixmul() API that allows tinyBLAS to
speed up "Mixture of Expert" models. On my Threadripper, Mixtral's 8x7b
F16 weights now process prompts 2x faster. I'm also seeing a 60 percent
improvement with Mixtral 8x22b Q4_0. The same applies to Q8_0, which is
also supported by tinyBLAS. MoE models spend the majority of their time
inside MUL_MAT_ID rather than MUL_MAT, which is why llamafile_sgemm was
not able to help them before. llamafile_mixmul works by decomposing the
mixmul operation into sgemm calls.

common/common.cpp

@@ -74,7 +74,7 @@
 using json = nlohmann::ordered_json;
 
 int32_t get_num_physical_cores() {
-#ifdef __linux__
+#if defined(__linux__) || defined(__COSMOPOLITAN__)
     // enumerate the set of thread siblings, num entries is num cores
     std::unordered_set<std::string> siblings;
     for (uint32_t cpu=0; cpu < UINT32_MAX; ++cpu) {
@@ -109,7 +109,7 @@ int32_t get_num_physical_cores() {
     return n_threads > 0 ? (n_threads <= 4 ? n_threads : n_threads / 2) : 4;
 }
 
-#if defined(__x86_64__) && defined(__linux__) && !defined(__ANDROID__)
+#if defined(__x86_64__) && (defined(__linux__) || defined(__COSMOPOLITAN__)) && !defined(__ANDROID__)
 #include <pthread.h>
 
 static void cpuid(unsigned leaf, unsigned subleaf,
@@ -163,7 +163,7 @@ static int count_math_cpus(int cpu_count) {
  * Returns number of CPUs on system that are useful for math.
  */
 int get_math_cpu_count() {
-#if defined(__x86_64__) && defined(__linux__) && !defined(__ANDROID__)
+#if defined(__x86_64__) && (defined(__linux__) || defined(__COSMOPOLITAN__)) && !defined(__ANDROID__)
     int cpu_count = sysconf(_SC_NPROCESSORS_ONLN);
     if (cpu_count < 1) {
         return get_num_physical_cores();

ggml-impl.h

@@ -17,6 +17,9 @@
 #define MIN(a, b) ((a) < (b) ? (a) : (b))
 #define MAX(a, b) ((a) > (b) ? (a) : (b))
 
+// some compilers don't provide _mm256_set_m128i, e.g. gcc 7
+#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1)
+
 #if defined(_WIN32)
 
 #define m512bh(p) p

ggml-quants.c

@@ -28,9 +28,6 @@
 
 #define UNUSED GGML_UNUSED
 
-// some compilers don't provide _mm256_set_m128i, e.g. gcc 7
-#define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1)
-
 #if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__)
 // multiply int8_t, add results pairwise twice
 static inline __m128i mul_sum_i8_pairs(const __m128i x, const __m128i y) {

ggml.c

@@ -12687,11 +12687,16 @@ static void ggml_compute_forward_mul_mat_id(
     const struct ggml_tensor * src1 = dst->src[1];
     const struct ggml_tensor * ids = dst->src[2];
 
-    GGML_TENSOR_BINARY_OP_LOCALS
+#if GGML_USE_LLAMAFILE
+    if (llamafile_mixmul(params, src0, src1, ids, dst))
+        return;
+#endif
 
     const int ith = params->ith;
     const int nth = params->nth;
 
+    GGML_TENSOR_BINARY_OP_LOCALS
+
     const enum ggml_type type = src0->type;
 
     const bool src1_cont = ggml_is_contiguous(src1);
@@ -20167,6 +20172,9 @@ struct ggml_cplan ggml_graph_plan(const struct ggml_cgraph * cgraph, int n_threa
                     cur = 0;
                     const struct ggml_tensor * src0 = node->src[0];
                     const struct ggml_tensor * src1 = node->src[1];
+#if GGML_USE_LLAMAFILE
+                    const struct ggml_tensor * src2 = node->src[2];
+#endif
                     const enum ggml_type vec_dot_type = type_traits[src0->type].vec_dot_type;
                     if (src1->type != vec_dot_type) {
                         cur += ggml_row_size(vec_dot_type, ggml_nelements(src1));
@@ -20175,6 +20183,10 @@ struct ggml_cplan ggml_graph_plan(const struct ggml_cgraph * cgraph, int n_threa
                     cur += GGML_PAD(cur, sizeof(int64_t));       // align
                     cur += n_as * sizeof(int64_t);               // matrix_row_counts
                     cur += n_as * src1->ne[2] * sizeof(int64_t); // matrix_rows
+#if GGML_USE_LLAMAFILE
+                    size_t cur2 = llamafile_mixmul_needs(src0, src1, src2);
+                    cur = cur > cur2 ? cur : cur2;
+#endif
                 } break;
             case GGML_OP_OUT_PROD:
                 {