metal : bug-fix when enable ggml-alloc (#2757)

* metal: better memory alloc w/ concurrency dispatch

The ggml-alloc should only free tensors at memory barriers.

* ggml-alloc: avoid return silently

In certain cases, the allocate_node() function may silently return
without performing any memory allocation.

convert.py

@@ -106,6 +106,9 @@ class Params:
 
     ftype: Optional[GGMLFileType] = None
 
+    # path to the directory containing the model files
+    path_model: Optional['Path'] = None
+
     @staticmethod
     def find_n_mult(n_ff: int, n_embd: int) -> int:
         # hardcoded magic range
@@ -191,7 +194,7 @@ class Params:
     def loadOriginalParamsJson(model: 'LazyModel', config_path: 'Path') -> 'Params':
         config = json.load(open(config_path))
 
-        n_vocab    = config["vocab_size"]
+        n_vocab    = config["vocab_size"] if "vocab_size" in config else -1
         n_embd     = config["dim"]
         n_layer    = config["n_layers"]
         n_mult     = config["multiple_of"]
@@ -231,6 +234,8 @@ class Params:
         else:
             params = Params.guessed(model_plus.model)
 
+        params.path_model = model_plus.paths[0].parent
+
         return params
 
 
@@ -733,11 +738,13 @@ class OutputFile:
         self.gguf = gguf.GGUFWriter(fname_out, gguf.MODEL_ARCH_NAMES[ARCH])
 
     def add_meta_arch(self, params: Params) -> None:
-        ver = None
+        name = "LLaMA"
         if (params.n_ctx == 4096):
-            ver = "v2"
+            name = "LLaMA v2"
+            if params.path_model:
+                name = str(params.path_model.parent).split('/')[-1]
 
-        self.gguf.add_name                ("LLaMA" if ver == None else "LLaMA " + ver)
+        self.gguf.add_name                (name)
         self.gguf.add_context_length      (params.n_ctx)
         self.gguf.add_embedding_length    (params.n_embd)
         self.gguf.add_block_count         (params.n_layer)

examples/main/main.cpp

@@ -798,7 +798,8 @@ int main(int argc, char ** argv) {
         }
 
         // In interactive mode, respect the maximum number of tokens and drop back to user input when reached.
-        if (params.interactive && n_remain <= 0 && params.n_predict != -1) {
+        // We skip this logic when n_predict == -1 (infinite) or -2 (stop at context size).
+        if (params.interactive && n_remain <= 0 && params.n_predict >= 0) {
             n_remain = params.n_predict;
             is_interacting = true;
         }

ggml-alloc.c

@@ -68,7 +68,7 @@ struct ggml_allocr {
     size_t max_size;
     bool measure;
     int parse_seq[GGML_MAX_NODES];
-    bool has_parse_seq;
+    int parse_seq_len;
 
 #ifdef GGML_ALLOCATOR_DEBUG
     struct ggml_tensor * allocated_tensors[1024];
@@ -239,14 +239,10 @@ static void ggml_allocator_free_tensor(struct ggml_allocr * alloc, struct ggml_t
 }
 
 void ggml_allocr_set_parse_seq(struct ggml_allocr * alloc, const int * list, int n) {
-    int pos = 0;
     for (int i = 0; i < n; i++) {
-        if (list[i] != -1) {
-            alloc->parse_seq[pos] = list[i];
-            pos++;
-        }
+        alloc->parse_seq[i] = list[i];
     }
-    alloc->has_parse_seq = true;
+    alloc->parse_seq_len = n;
 }
 
 void ggml_allocr_reset(struct ggml_allocr * alloc) {
@@ -269,7 +265,7 @@ struct ggml_allocr * ggml_allocr_new(void * data, size_t size, size_t alignment)
         /*.max_size      = */ 0,
         /*.measure       = */ false,
         /*.parse_seq     = */ {0},
-        /*.has_parse_seq = */ false,
+        /*.parse_seq_len = */ 0,
 #ifdef GGML_ALLOCATOR_DEBUG
         /*.allocated_tensors = */ = {0},
 #endif
@@ -298,7 +294,7 @@ struct ggml_allocr * ggml_allocr_new_measure(size_t alignment) {
         /*.max_size      = */ 0,
         /*.measure       = */ true,
         /*.parse_seq     = */ {0},
-        /*.has_parse_seq = */ false,
+        /*.parse_seq_len = */ 0,
 #ifdef GGML_ALLOCATOR_DEBUG
         /*.allocated_tensors = */ = {0},
 #endif
@@ -445,8 +441,8 @@ static void allocate_node(struct ggml_allocr * alloc, struct ggml_tensor * node)
                         else {
                             AT_PRINTF("reusing parent %s for %s\n", parent->name, node->name);
                             node->data = parent->data;
+                            return;
                         }
-                        return;
                     }
                 }
             }
@@ -497,69 +493,86 @@ static size_t ggml_allocator_alloc_graph_tensors_n(
                 allocate_node(alloc, input);
             }
         }
-        for (int ind = 0; ind < gf->n_nodes; ind++) {
-            int i;
-            if (alloc->has_parse_seq) {
-                i = alloc->parse_seq[ind];
-            } else {
-                i = ind;
-            }
-            struct ggml_tensor * node = gf->nodes[i];
+        // if we have parse_seq then we allocate nodes following the list, and we only free nodes at barriers
+        int last_barrier_pos = 0;
+        int n_nodes = alloc->parse_seq_len ? alloc->parse_seq_len : gf->n_nodes;
 
-            // allocate parents (leafs)
-            for (int j = 0; j < GGML_MAX_SRC; j++) {
-                struct ggml_tensor * parent = node->src[j];
-                if (parent == NULL) {
-                    break;
+        for (int ind = 0; ind < n_nodes; ind++) {
+            // allocate a node if there is no parse_seq or this is not a barrier
+            if ((alloc->parse_seq_len==0) || alloc->parse_seq[ind] != -1) {
+                int i = alloc->parse_seq_len ? alloc->parse_seq[ind] : ind;
+                struct ggml_tensor * node = gf->nodes[i];
+
+                // allocate parents (leafs)
+                for (int j = 0; j < GGML_MAX_SRC; j++) {
+                    struct ggml_tensor * parent = node->src[j];
+                    if (parent == NULL) {
+                        break;
+                    }
+                    allocate_node(alloc, parent);
                 }
-                allocate_node(alloc, parent);
+
+                // allocate node
+                allocate_node(alloc, node);
+
+                AT_PRINTF("exec: %s (%s) <= ", ggml_op_name(node->op), node->name);
+                for (int j = 0; j < GGML_MAX_SRC; j++) {
+                    struct ggml_tensor * parent = node->src[j];
+                    if (parent == NULL) {
+                        break;
+                    }
+                    AT_PRINTF("%s", parent->name);
+                    if (j < GGML_MAX_SRC - 1 && node->src[j + 1] != NULL) {
+                        AT_PRINTF(", ");
+                    }
+                }
+                AT_PRINTF("\n");
             }
 
-            // allocate node
-            allocate_node(alloc, node);
-
-            AT_PRINTF("exec: %s (%s) <= ", ggml_op_name(node->op), node->name);
-            for (int j = 0; j < GGML_MAX_SRC; j++) {
-                struct ggml_tensor * parent = node->src[j];
-                if (parent == NULL) {
-                    break;
-                }
-                AT_PRINTF("%s", parent->name);
-                if (j < GGML_MAX_SRC - 1 && node->src[j + 1] != NULL) {
-                    AT_PRINTF(", ");
-                }
-            }
-            AT_PRINTF("\n");
 
             // update parents
-            for (int j = 0; j < GGML_MAX_SRC; j++) {
-                struct ggml_tensor * parent = node->src[j];
-                if (parent == NULL) {
-                    break;
+            // update immediately if there is no parse_seq
+            // update only at barriers if there is parse_seq
+            if ((alloc->parse_seq_len==0) || alloc->parse_seq[ind] == -1) {
+                int update_start = alloc->parse_seq_len ? last_barrier_pos : ind;
+                int update_end   = alloc->parse_seq_len ? ind              : ind + 1;
+                for (int i = update_start; i < update_end; i++) {
+                    int node_i = alloc->parse_seq_len ? alloc->parse_seq[i] : i;
+                    struct ggml_tensor * node = gf->nodes[node_i];
+
+                    for (int j = 0; j < GGML_MAX_SRC; j++) {
+                        struct ggml_tensor * parent = node->src[j];
+                        if (parent == NULL) {
+                            break;
+                        }
+                        struct hash_node * p_hn = hash_get(ht, parent);
+                        p_hn->n_children -= 1;
+
+                        //AT_PRINTF("parent %s: %d children, %d views\n", parent->name, parent->n_children, parent->n_views);
+
+                        if (p_hn->n_children == 0 && p_hn->n_views == 0) {
+                            if (ggml_is_view(parent)) {
+                                struct ggml_tensor * view_src = get_view_source(parent);
+                                struct hash_node * view_src_hn = hash_get(ht, view_src);
+                                view_src_hn->n_views -= 1;
+                                AT_PRINTF("view_src %s\n", view_src->name);
+                                if (view_src_hn->n_views == 0 && view_src_hn->n_children == 0 && view_src->data != node->data) {
+                                    ggml_allocator_free_tensor(alloc, view_src);
+                                }
+                            }
+                            else {
+                                if (parent->data != node->data) {
+                                    ggml_allocator_free_tensor(alloc, parent);
+                                }
+                            }
+                        }
+                    }
                 }
-                struct hash_node * p_hn = hash_get(ht, parent);
-                p_hn->n_children -= 1;
-
-                //AT_PRINTF("parent %s: %d children, %d views\n", parent->name, parent->n_children, parent->n_views);
-
-                if (p_hn->n_children == 0 && p_hn->n_views == 0) {
-                    if (ggml_is_view(parent)) {
-                        struct ggml_tensor * view_src = get_view_source(parent);
-                        struct hash_node * view_src_hn = hash_get(ht, view_src);
-                        view_src_hn->n_views -= 1;
-                        AT_PRINTF("view_src %s\n", view_src->name);
-                        if (view_src_hn->n_views == 0 && view_src_hn->n_children == 0 && view_src->data != node->data) {
-                            ggml_allocator_free_tensor(alloc, view_src);
-                        }
-                    }
-                    else {
-                        if (parent->data != node->data) {
-                            ggml_allocator_free_tensor(alloc, parent);
-                        }
-                    }
+                AT_PRINTF("\n");
+                if (alloc->parse_seq_len) {
+                    last_barrier_pos = ind + 1;
                 }
             }
-            AT_PRINTF("\n");
         }
         // free graph outputs here that wouldn't be freed otherwise because they have no children
         if (outputs != NULL && outputs[g] != NULL) {

llama.cpp

@@ -827,6 +827,7 @@ enum e_model {
     MODEL_7B,
     MODEL_13B,
     MODEL_30B,
+    MODEL_34B,
     MODEL_40B,
     MODEL_65B,
     MODEL_70B,
@@ -1518,6 +1519,7 @@ static const char * llama_model_type_name(e_model type) {
         case MODEL_7B:  return "7B";
         case MODEL_13B: return "13B";
         case MODEL_30B: return "30B";
+        case MODEL_34B: return "34B";
         case MODEL_40B: return "40B";
         case MODEL_65B: return "65B";
         case MODEL_70B: return "70B";
@@ -1590,6 +1592,7 @@ static void llm_load_hparams(
                     case 26: model.type = e_model::MODEL_3B; break;
                     case 32: model.type = e_model::MODEL_7B; break;
                     case 40: model.type = e_model::MODEL_13B; break;
+                    case 48: model.type = e_model::MODEL_34B; break;
                     case 60: model.type = e_model::MODEL_30B; break;
                     case 80: model.type = hparams.n_head == hparams.n_head_kv ? e_model::MODEL_65B : e_model::MODEL_70B; break;
                     default: model.type = e_model::MODEL_UNKNOWN;
@@ -2704,11 +2707,6 @@ static struct ggml_cgraph * llm_build_falcon(
             struct ggml_tensor * inpFF = attn_norm;
 
             cur = ggml_mul_mat(ctx0, model.layers[il].w3, inpFF);
-
-            // TODO: this is temporary needed to introduce artificial dependency between FF and ATTN
-            //       adding this, because there seems to be a bug in the Metal concurrency optimization
-            //       without this line, the results are non-deterministic and wrong
-            cur->src[2] = attn_out;
             offload_func(cur);
 
             cur = ggml_gelu(ctx0, cur);

scripts/perf-run-all.sh

@@ -1,93 +0,0 @@
-#!/bin/bash
-#
-# Measure the performance (time per token) of the various quantization techniques
-#
-
-QUANTIZE=0
-if [ "$1" != "" ]; then
-    echo "Quantizing"
-    QUANTIZE=1
-fi
-
-if [ "$QUANTIZE" != "0" ]; then
-    #
-    # quantize
-    #
-
-    # 7B
-    time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q4_0.bin q4_0 2>&1 | tee ../qnt-7b-q4_0.txt
-    time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q4_1.bin q4_1 2>&1 | tee ../qnt-7b-q4_1.txt
-    time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q5_0.bin q5_0 2>&1 | tee ../qnt-7b-q5_0.txt
-    time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q5_1.bin q5_1 2>&1 | tee ../qnt-7b-q5_1.txt
-    time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q8_0.bin q8_0 2>&1 | tee ../qnt-7b-q8_0.txt
-
-    # 13B
-    time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q4_0.bin q4_0 2>&1 | tee ../qnt-13b-q4_0.txt
-    time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q4_1.bin q4_1 2>&1 | tee ../qnt-13b-q4_1.txt
-    time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q5_0.bin q5_0 2>&1 | tee ../qnt-13b-q5_0.txt
-    time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q5_1.bin q5_1 2>&1 | tee ../qnt-13b-q5_1.txt
-    time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q8_0.bin q8_0 2>&1 | tee ../qnt-13b-q8_0.txt
-fi
-
-#
-# perf
-# run each command twice
-#
-
-set -x
-
-# 7B - 4 threads
-     ./bin/main -m ../models/7B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-f16.txt  | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-q4_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-q4_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-q5_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-q5_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-7b-q8_0.txt | grep llama_print_timings
-
-# 7B - 8 threads
-     ./bin/main -m ../models/7B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-f16.txt  | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-q4_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-q4_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-q5_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-q5_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/7B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/7B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-7b-q8_0.txt | grep llama_print_timings
-
-# 13B - 4 threads
-     ./bin/main -m ../models/13B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-f16.txt  | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-q4_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-q4_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-q5_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-q5_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 4 2>&1 | tee ../perf-13b-q8_0.txt | grep llama_print_timings
-
-# 13B - 8 threads
-     ./bin/main -m ../models/13B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-f16.bin  -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-f16.txt  | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q4_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-q4_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q4_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-q4_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q5_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-q5_0.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q5_1.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-q5_1.txt | grep llama_print_timings
-     ./bin/main -m ../models/13B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | grep "I believe"
-time ./bin/main -m ../models/13B/ggml-model-q8_0.bin -p "I believe the meaning of life is" --no-mmap -c 2048 --ignore-eos -s 1 -n 64 -t 8 2>&1 | tee ../perf-13b-q8_0.txt | grep llama_print_timings

scripts/ppl-run-all.sh

@@ -1,39 +0,0 @@
-#!/bin/bash
-
-#
-# quantize
-#
-
-# 7B
-time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q4_0.bin q4_0 2>&1 | tee ../qnt-7b-q4_0.txt
-time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q4_1.bin q4_1 2>&1 | tee ../qnt-7b-q4_1.txt
-time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q5_0.bin q5_0 2>&1 | tee ../qnt-7b-q5_0.txt
-time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q5_1.bin q5_1 2>&1 | tee ../qnt-7b-q5_1.txt
-time ./bin/quantize ../models/7B/ggml-model-f16.bin ../models/7B/ggml-model-q8_0.bin q8_0 2>&1 | tee ../qnt-7b-q8_0.txt
-
-# 13B
-time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q4_0.bin q4_0 2>&1 | tee ../qnt-13b-q4_0.txt
-time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q4_1.bin q4_1 2>&1 | tee ../qnt-13b-q4_1.txt
-time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q5_0.bin q5_0 2>&1 | tee ../qnt-13b-q5_0.txt
-time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q5_1.bin q5_1 2>&1 | tee ../qnt-13b-q5_1.txt
-time ./bin/quantize ../models/13B/ggml-model-f16.bin ../models/13B/ggml-model-q8_0.bin q8_0 2>&1 | tee ../qnt-13b-q8_0.txt
-
-#
-# perplexity
-#
-
-# 7B
-time ./bin/perplexity -m ../models/7B/ggml-model-f16.bin  -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-f16.txt
-time ./bin/perplexity -m ../models/7B/ggml-model-q4_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-q4_0.txt
-time ./bin/perplexity -m ../models/7B/ggml-model-q4_1.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-q4_1.txt
-time ./bin/perplexity -m ../models/7B/ggml-model-q5_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-q5_0.txt
-time ./bin/perplexity -m ../models/7B/ggml-model-q5_1.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-q5_1.txt
-time ./bin/perplexity -m ../models/7B/ggml-model-q8_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-7b-q8_0.txt
-
-# 13B
-time ./bin/perplexity -m ../models/13B/ggml-model-f16.bin  -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-f16.txt
-time ./bin/perplexity -m ../models/13B/ggml-model-q4_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-q4_0.txt
-time ./bin/perplexity -m ../models/13B/ggml-model-q4_1.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-q4_1.txt
-time ./bin/perplexity -m ../models/13B/ggml-model-q5_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-q5_0.txt
-time ./bin/perplexity -m ../models/13B/ggml-model-q5_1.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-q5_1.txt
-time ./bin/perplexity -m ../models/13B/ggml-model-q8_0.bin -f ./wiki.test.raw --no-mmap -t 12 2>&1 | tee ../ppl-13b-q8_0.txt

scripts/qnt-all.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+
+qnt=(q8_0 q6_k q5_k q5_1 q5_0 q4_k q4_1 q4_0 q3_k q2_k)
+args=""
+
+if [ -z "$1" ]; then
+    echo "usage: $0 <model> [qnt] [args]"
+    echo "default: $0 <model> \"${qnt[@]}\" \"${args}\""
+    exit 1
+fi
+
+if [ ! -z "$2" ]; then
+    qnt=($2)
+fi
+
+if [ ! -z "$3" ]; then
+    args="$3"
+fi
+
+model="$1"
+out="../tmp/results-${model}"
+
+mkdir -p ${out}
+
+for q in ${qnt[@]}; do
+    time ./bin/quantize ../models/${model}/ggml-model-f16.gguf ../models/${model}/ggml-model-${q}.gguf ${q} 2>&1 ${args} | tee ${out}/qnt-${q}.txt
+done

scripts/run-all-perf.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+
+qnt=(f16 q8_0 q6_k q5_k q5_1 q5_0 q4_k q4_1 q4_0 q3_k q2_k)
+args="-ngl 999 -n 64 -p 512"
+
+if [ -z "$1" ]; then
+    echo "usage: $0 <model> [qnt] [args]"
+    echo "default: $0 <model> \"${qnt[@]}\" \"${args}\""
+    exit 1
+fi
+
+if [ ! -z "$2" ]; then
+    qnt=($2)
+fi
+
+if [ ! -z "$3" ]; then
+    args="$3"
+fi
+
+model="$1"
+out="../tmp/results-${model}"
+
+mkdir -p ${out}
+
+mstr=""
+
+for q in ${qnt[@]}; do
+    mstr="${mstr} -m ../models/${model}/ggml-model-${q}.gguf"
+done
+
+./bin/llama-bench ${mstr} ${args} 2> /dev/null

scripts/run-all-ppl.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+
+qnt=(f16 q8_0 q6_k q5_k q5_1 q5_0 q4_k q4_1 q4_0 q3_k q2_k)
+args="--no-mmap -ngl 999 -t 8"
+
+if [ -z "$1" ]; then
+    echo "usage: $0 <model> [qnt] [args]"
+    echo "default: $0 <model> \"${qnt[@]}\" \"${args}\""
+    exit 1
+fi
+
+if [ ! -z "$2" ]; then
+    qnt=($2)
+fi
+
+if [ ! -z "$3" ]; then
+    args="$3"
+fi
+
+model="$1"
+out="../tmp/results-${model}"
+
+mkdir -p ${out}
+
+for q in ${qnt[@]}; do
+    time ./bin/perplexity -m ../models/${model}/ggml-model-f16.gguf -f ./wiki.test.raw ${args} 2>&1 | tee ${out}/ppl-${q}.txt
+done