cont : kv-cells cp/set for non-cont slots

ggml-ci

Author: ggerganov

src/llama-kv-cache-unified.cpp

@@ -388,7 +388,8 @@ llama_kv_cache_unified::slot_info_vec_t llama_kv_cache_unified::prepare(const st
 
     struct state {
         uint32_t head_old; // old position of the head, before placing the ubatch
-        uint32_t head_new; // new position of the head, after placing the ubatch
+
+        slot_info sinfo; // slot info for the ubatch
 
         llama_kv_cells_unified cells; // copy of the old cells, before placing the ubatch
     };
@@ -409,21 +410,16 @@ llama_kv_cache_unified::slot_info_vec_t llama_kv_cache_unified::prepare(const st
         // remeber the position that we found
         res.push_back(sinfo_new);
 
-        // TODO: temporary
-        if (supports_set_rows) {
-            GGML_ASSERT(sinfo_new.is_cont());
-        }
-
         // store the old state of the cells in the recovery stack
-        states.push_back({head, sinfo_new.head(), cells.cp(sinfo_new.head(), ubatch.n_tokens)});
+        states.push_back({head, sinfo_new, cells.cp(sinfo_new.idxs)});
 
         // now emplace the ubatch
         apply_ubatch(sinfo_new, ubatch);
     }
 
     // iterate backwards and restore the cells to their original state
     for (auto it = states.rbegin(); it != states.rend(); ++it) {
-        cells.set(it->head_new, it->cells);
+        cells.set(it->sinfo.idxs, it->cells);
         head = it->head_old;
     }
 

src/llama-kv-cache-unified.h

@@ -49,20 +49,6 @@ class llama_kv_cache_unified : public llama_memory_i {
             return idxs.empty();
         }
 
-        // TODO: tmp until kv cells support non-cont slots
-        bool is_cont() const {
-            bool res = true;
-
-            for (uint32_t i = 1; i < idxs.size(); ++i) {
-                if (idxs[i] != idxs[i - 1] + 1) {
-                    res = false;
-                    break;
-                }
-            }
-
-            return res;
-        }
-
         // TODO: implement
         //std::vector<idx_vec_t> seq_idxs;
     };

src/llama-kv-cells.h

@@ -105,10 +105,29 @@ class llama_kv_cells_unified {
         res.resize(n);
 
         for (uint32_t j = 0; j < n; ++j) {
-            res.pos[j] = pos[i + j];
-            res.seq[j] = seq[i + j];
+            const auto idx = i + j;
 
-            assert(shift[i + j] == 0);
+            res.pos[j] = pos[idx];
+            res.seq[j] = seq[idx];
+
+            assert(shift[idx] == 0);
+        }
+
+        return res;
+    }
+
+    llama_kv_cells_unified cp(const std::vector<uint32_t> & idxs) const {
+        llama_kv_cells_unified res;
+
+        res.resize(idxs.size());
+
+        for (uint32_t j = 0; j < idxs.size(); ++j) {
+            const auto idx = idxs[j];
+
+            res.pos[j] = pos[idx];
+            res.seq[j] = seq[idx];
+
+            assert(shift[idx] == 0);
         }
 
         return res;
@@ -119,26 +138,57 @@ class llama_kv_cells_unified {
         assert(i + other.pos.size() <= pos.size());
 
         for (uint32_t j = 0; j < other.pos.size(); ++j) {
-            if (pos[i + j] == -1 && other.pos[j] != -1) {
+            const auto idx = i + j;
+
+            if (pos[idx] == -1 && other.pos[j] != -1) {
                 used.insert(i + j);
             }
 
-            if (pos[i + j] != -1 && other.pos[j] == -1) {
+            if (pos[idx] != -1 && other.pos[j] == -1) {
                 used.erase(i + j);
             }
 
-            if (pos[i + j] != -1) {
+            if (pos[idx] != -1) {
                 seq_pos_rm(i + j);
             }
 
-            pos[i + j] = other.pos[j];
-            seq[i + j] = other.seq[j];
+            pos[idx] = other.pos[j];
+            seq[idx] = other.seq[j];
 
-            if (pos[i + j] != -1) {
+            if (pos[idx] != -1) {
                 seq_pos_add(i + j);
             }
 
-            assert(shift[i + j] == 0);
+            assert(shift[idx] == 0);
+        }
+    }
+
+    void set(const std::vector<uint32_t> & idxs, const llama_kv_cells_unified & other) {
+        assert(idxs.size() == other.pos.size());
+
+        for (uint32_t j = 0; j < other.pos.size(); ++j) {
+            const auto idx = idxs[j];
+
+            if (pos[idx] == -1 && other.pos[j] != -1) {
+                used.insert(idx);
+            }
+
+            if (pos[idx] != -1 && other.pos[j] == -1) {
+                used.erase(idx);
+            }
+
+            if (pos[idx] != -1) {
+                seq_pos_rm(idx);
+            }
+
+            pos[idx] = other.pos[j];
+            seq[idx] = other.seq[j];
+
+            if (pos[idx] != -1) {
+                seq_pos_add(idx);
+            }
+
+            assert(shift[idx] == 0);
         }
     }