mpi : trying to move more MPI stuff into ggml-mpi (#2099)

examples/embd-input/embd-input-lib.cpp

@@ -34,7 +34,7 @@ struct MyModel* create_mymodel(int argc, char ** argv) {
     }
     fprintf(stderr, "%s: seed  = %d\n", __func__, params.seed);
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     llama_model * model;
     llama_context * ctx;

examples/embedding/embedding.cpp

@@ -35,7 +35,7 @@ int main(int argc, char ** argv) {
         params.prompt = gpt_random_prompt(rng);
     }
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     llama_model * model;
     llama_context * ctx;
@@ -93,5 +93,7 @@ int main(int argc, char ** argv) {
     llama_free(ctx);
     llama_free_model(model);
 
+    llama_backend_free();
+
     return 0;
 }

examples/main/main.cpp

@@ -105,7 +105,7 @@ int main(int argc, char ** argv) {
         params.prompt = gpt_random_prompt(rng);
     }
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     llama_model * model;
     llama_context * ctx;
@@ -671,7 +671,7 @@ int main(int argc, char ** argv) {
     llama_free(ctx);
     llama_free_model(model);
 
-    llama_finalize_backend();
+    llama_backend_free();
 
     return 0;
 }

examples/perplexity/perplexity.cpp

@@ -147,7 +147,7 @@ int main(int argc, char ** argv) {
         params.prompt = gpt_random_prompt(rng);
     }
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     llama_model * model;
     llama_context * ctx;
@@ -172,7 +172,7 @@ int main(int argc, char ** argv) {
     llama_free(ctx);
     llama_free_model(model);
 
-    llama_finalize_backend();
+    llama_backend_free();
 
     return 0;
 }

examples/quantize/quantize.cpp

@@ -180,7 +180,7 @@ int main(int argc, char ** argv) {
         usage(argv[0]);
     }
 
-    llama_init_backend(false);
+    llama_backend_init(false);
 
     // parse command line arguments
     const std::string fname_inp = argv[arg_idx];
@@ -257,5 +257,7 @@ int main(int argc, char ** argv) {
         printf("%s:    total time = %8.2f ms\n", __func__, (t_main_end_us - t_main_start_us)/1000.0);
     }
 
+    llama_backend_free();
+
     return 0;
 }

examples/server/server.cpp

@@ -1079,7 +1079,7 @@ int main(int argc, char **argv)
         params.model_alias = params.model;
     }
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     LOG_INFO("build info", {{"build", BUILD_NUMBER},
                             {"commit", BUILD_COMMIT}});
@@ -1309,5 +1309,7 @@ int main(int argc, char **argv)
         return 1;
     }
 
+    llama_backend_free();
+
     return 0;
 }

examples/simple/simple.cpp

@@ -66,7 +66,7 @@ int main(int argc, char ** argv)
     // Init LLM :
     //---------------------------------
 
-    llama_init_backend(params.numa);
+    llama_backend_init(params.numa);
 
     llama_model * model;
     llama_context * ctx;
@@ -173,7 +173,7 @@ int main(int argc, char ** argv)
     llama_free( ctx );
     llama_free_model( model );
 
-    llama_finalize_backend();
+    llama_backend_free();
 
     return 0;
 }

ggml-mpi.c

@@ -2,80 +2,211 @@
 
 #include "ggml.h"
 
+#include <mpi.h>
+
 #include <stdio.h>
 #include <stdlib.h>
-#include <mpi.h>
+
+#define MIN(a, b) ((a) < (b) ? (a) : (b))
+
 #define UNUSED GGML_UNUSED
 
-struct ggml_mpi_tensor_info {
+struct ggml_mpi_context {
     int rank;
+    int size;
 };
 
-// ggml_compute_forward_send
+void ggml_mpi_backend_init(void) {
+    MPI_Init(NULL, NULL);
+}
 
-static void ggml_mpi_compute_forward_send(
-        struct ggml_tensor * src,
-        const struct ggml_tensor * orig) {
-    UNUSED(orig);
-    GGML_ASSERT(src->type == GGML_TYPE_F32);
+void ggml_mpi_backend_free(void) {
+    MPI_Finalize();
+}
+
+struct ggml_mpi_context * ggml_mpi_init(void) {
+    struct ggml_mpi_context * ctx = calloc(1, sizeof(struct ggml_mpi_context));
 
-    int my_rank;
-    MPI_Comm_rank(MPI_COMM_WORLD, &my_rank);
+    MPI_Comm_rank(MPI_COMM_WORLD, &ctx->rank);
+    MPI_Comm_size(MPI_COMM_WORLD, &ctx->size);
 
-    int dst_rank = ((struct ggml_mpi_tensor_info *)src->extra)->rank;
-    // fprintf(stderr, "(%d) Sending to (%d)\n", my_rank, (int)dst->extra);
-    int retval = MPI_Send(src->data, ggml_nelements(src), MPI_FLOAT, dst_rank, 0, MPI_COMM_WORLD);
-    // fprintf(stderr, "(%d) Sent to (%d)\n", my_rank, (int)dst->extra);
-    GGML_ASSERT(retval == MPI_SUCCESS);
+    return ctx;
 }
 
-// ggml_compute_forward_recv
-
-static void ggml_mpi_compute_forward_recv(
-        struct ggml_tensor * dst,
-        const struct ggml_tensor * orig,
-        const struct ggml_tensor * parent) {
-    UNUSED(parent);
-    UNUSED(orig);
-    GGML_ASSERT(dst->type == GGML_TYPE_F32);
-    MPI_Status status;
-
-    int my_rank;
-    MPI_Comm_rank(MPI_COMM_WORLD, &my_rank);
-
-    int src_rank = ((struct ggml_mpi_tensor_info *)dst->extra)->rank;
-    // fprintf(stderr, "(%d) Receiving from (%d)\n", my_rank, src_extra);
-    int retval = MPI_Recv(dst->data, ggml_nelements(dst), MPI_FLOAT, src_rank, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
-    // fprintf(stderr, "(%d) Received from (%d)\n", my_rank, src_extra);
-    GGML_ASSERT(retval == MPI_SUCCESS);
+void ggml_mpi_free(struct ggml_mpi_context * ctx) {
+    free(ctx);
 }
 
-struct ggml_tensor * ggml_mpi_send_tensor(
-        struct ggml_context * ctx,
-        struct ggml_tensor *src,
-        int dst_rank) {
+int ggml_mpi_rank(struct ggml_mpi_context * ctx) {
+    return ctx->rank;
+}
 
-    struct ggml_tensor * result = ggml_map_custom1_inplace_f32(ctx, src, ggml_mpi_compute_forward_send);
+void ggml_mpi_eval_init(
+        struct ggml_mpi_context * ctx_mpi,
+                            int * n_tokens,
+                            int * n_past,
+                            int * n_threads) {
+    UNUSED(ctx_mpi);
 
-    // TODO how/when to free this struct?
-    struct ggml_mpi_tensor_info *info = calloc(1, sizeof(struct ggml_mpi_tensor_info));
-    info->rank = dst_rank;
-    result->extra = info;
+    // synchronize the worker node parameters with the root node
+    MPI_Barrier(MPI_COMM_WORLD);
+
+    MPI_Bcast(n_tokens,  1, MPI_INT, 0, MPI_COMM_WORLD);
+    MPI_Bcast(n_past,    1, MPI_INT, 0, MPI_COMM_WORLD);
+    MPI_Bcast(n_threads, 1, MPI_INT, 0, MPI_COMM_WORLD);
+}
 
-    return result;
+int ggml_graph_get_node_idx(struct ggml_cgraph * gf, const char * name) {
+    struct ggml_tensor * t = ggml_graph_get_tensor(gf, name);
+    if (t == NULL) {
+        fprintf(stderr, "%s: tensor %s not found\n", __func__, name);
+        return -1;
+    }
+
+    for (int i = 0; i < gf->n_nodes; i++) {
+        if (gf->nodes[i] == t) {
+            return i;
+        }
+    }
+
+    fprintf(stderr, "%s: tensor %s not found in graph (should not happen)\n", __func__, name);
+    return -1;
 }
 
-struct ggml_tensor * ggml_mpi_recv_tensor(
-        struct ggml_context * ctx,
-        struct ggml_tensor *parent,
-        struct ggml_tensor *dst,
-        int src_rank) {
-    struct ggml_tensor * result = ggml_map_custom2_inplace_f32(ctx, dst, parent, ggml_mpi_compute_forward_recv);
+// TODO: there are many improvements that can be done to this implementation
+void ggml_mpi_graph_compute(
+        struct ggml_mpi_context * ctx_mpi,
+        struct ggml_context     * ctx,
+             struct ggml_cgraph * gf,
+                            int   n_layers) {
+    const int mpi_rank = ctx_mpi->rank;
+    const int mpi_size = ctx_mpi->size;
+
+    struct ggml_tensor * inp_tokens = ggml_graph_get_tensor(gf, "inp_tokens");
+    if (inp_tokens == NULL) {
+        fprintf(stderr, "%s: tensor 'inp_tokens' not found\n", __func__);
+        return;
+    }
+
+    struct ggml_tensor * inp0 = ggml_graph_get_tensor(gf, "layer_inp_0");
+    if (inp0 == NULL) {
+        fprintf(stderr, "%s: tensor 'inp0' not found\n", __func__);
+        return;
+    }
+
+    GGML_ASSERT(inp0 == gf->nodes[0]);
+
+    // distribute the compute graph into slices across the MPI nodes
+    //
+    // the main node (0) processes the last layers + the remainder of the compute graph
+    // and is responsible to pass the input tokens to the first node (1)
+    //
+    // node 1:   [(  0) * n_per_node, (  1) * n_per_node)
+    // node 2:   [(  1) * n_per_node, (  2) * n_per_node)
+    // ...
+    // node n-1: [(n-2) * n_per_node, (n-1) * n_per_node)
+    // node 0:   [(n-1) * n_per_node,            n_nodes)
+    //
+    if (mpi_rank > 0) {
+        if (mpi_rank == 1) { // the first node receives the input tokens from the main node
+            MPI_Status status; UNUSED(status);
+
+            const int mpi_rank_src = mpi_rank - 1;
+
+            const int retval = MPI_Recv(inp_tokens->data, ggml_nelements(inp_tokens), MPI_INT, mpi_rank_src, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
+            GGML_ASSERT(retval == MPI_SUCCESS);
+        } else { // recv input data for each node into the "inp0" tensor (i.e. the first node in the compute graph)
+            MPI_Status status; UNUSED(status);
+
+            const int mpi_rank_src = mpi_rank - 1;
+
+            //printf("%s: node %d: waiting for %d elements from %d\n", __func__, mpi_rank, (int) ggml_nelements(inp0), mpi_rank_src);
+            const int retval = MPI_Recv(inp0->data, ggml_nelements(inp0), MPI_FLOAT, mpi_rank_src, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
+            GGML_ASSERT(retval == MPI_SUCCESS);
+        }
+    } else if (mpi_size > 1) {
+        // node 0 sends the input tokens to node 1
+        {
+            const int mpi_rank_dst = mpi_rank + 1;
+
+            const int retval = MPI_Send(inp_tokens->data, ggml_nelements(inp_tokens), MPI_INT, mpi_rank_dst, 0, MPI_COMM_WORLD);
+            GGML_ASSERT(retval == MPI_SUCCESS);
+        }
+
+        // recv the output data from the last node
+        {
+            MPI_Status status; UNUSED(status);
+
+            const int mpi_rank_src = mpi_size - 1;
+
+            //fprintf(stderr, "%s: node %d: waiting for %d elements from %d\n", __func__, mpi_rank, (int) ggml_nelements(inp0), mpi_rank_src);
+            const int retval = MPI_Recv(inp0->data, ggml_nelements(inp0), MPI_FLOAT, mpi_rank_src, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
+            GGML_ASSERT(retval == MPI_SUCCESS);
+        }
+    }
+
+    {
+        const int n_per_node = (n_layers + (mpi_size - 1)) / mpi_size;
+
+        const int mpi_idx = mpi_rank > 0 ? mpi_rank - 1 : mpi_size - 1;
+
+        const int il0 =               (mpi_idx + 0) * n_per_node;
+        const int il1 = MIN(n_layers, (mpi_idx + 1) * n_per_node);
+
+        char name_l0[GGML_MAX_NAME];
+        char name_l1[GGML_MAX_NAME];
+
+        snprintf(name_l0, sizeof(name_l0), "layer_inp_%d", il0);
+        snprintf(name_l1, sizeof(name_l1), "layer_inp_%d", il1);
+
+        const int idx_l0 =                ggml_graph_get_node_idx(gf, name_l0);
+        const int idx_l1 = mpi_rank > 0 ? ggml_graph_get_node_idx(gf, name_l1) + 1 : gf->n_nodes;
+
+        if (idx_l0 < 0 || idx_l1 < 0) {
+            fprintf(stderr, "%s: layer input nodes not found\n", __func__);
+            return;
+        }
+
+        // attach the input data to all nodes that need it
+        // TODO: not great - should be able to do this without modifying the compute graph (see next TODO below)
+        for (int i = idx_l0; i < idx_l1; i++) {
+            if (gf->nodes[i]->src0 == gf->nodes[idx_l0]) {
+                gf->nodes[i]->src0 =  inp0;
+            }
+            if (gf->nodes[i]->src1 == gf->nodes[idx_l0]) {
+                gf->nodes[i]->src1 =  inp0;
+            }
+        }
+
+        // TODO: instead of rearranging the nodes, we should be able to execute a subset of the compute graph
+        for (int i = 1; i < idx_l1 - idx_l0; i++) {
+            gf->nodes[i] = gf->nodes[idx_l0 + i];
+            gf->grads[i] = gf->grads[idx_l0 + i];
+        }
+
+        // the first node performs the "get_rows" operation, the rest of the nodes get the data from the previous node
+        if (mpi_idx != 0) {
+            gf->nodes[0]->op = GGML_OP_NONE;
+        }
+
+        gf->n_nodes = idx_l1 - idx_l0;
+
+        //fprintf(stderr, "%s: node %d: processing %d nodes [%d, %d)\n", __func__, mpi_rank, gf->n_nodes, il0, il1);
+    }
+
+    ggml_graph_compute(ctx, gf);
+
+    //fprintf(stderr, "%s: node %d: done\n", __func__, mpi_rank);
+
+    // send the output data to the next node
+    if (mpi_rank > 0) {
+        struct ggml_tensor * output = gf->nodes[gf->n_nodes - 1];
+
+        const int mpi_rank_dst = (mpi_rank + 1) % mpi_size;
 
-    // TODO how/when to free this struct?
-    struct ggml_mpi_tensor_info *info = calloc(1, sizeof(struct ggml_mpi_tensor_info));
-    info->rank = src_rank;
-    result->extra = info;
+        //fprintf(stderr, "%s: node %d: sending %d elements to node %d\n", __func__, mpi_rank, ggml_nelements(output), mpi_rank_dst);
 
-    return result;
+        const int retval = MPI_Send(output->data, ggml_nelements(output), MPI_FLOAT, mpi_rank_dst, 0, MPI_COMM_WORLD);
+        GGML_ASSERT(retval == MPI_SUCCESS);
+    }
 }