From: Georgi Gerganov Date: Thu, 28 Aug 2025 09:27:02 +0000 (+0300) Subject: kv-cache : remove LLAMA_SET_ROWS checks (llama/15505) X-Git-Tag: v0.9.1~135 X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=92b6797ab8069d92febf975d7f61101fb16f7d5f;p=pkg%2Fggml%2Fsources%2Fggml kv-cache : remove LLAMA_SET_ROWS checks (llama/15505) ggml-ci --- diff --git a/src/ggml-cann/common.h b/src/ggml-cann/common.h index 33794062..88cc3f48 100755 --- a/src/ggml-cann/common.h +++ b/src/ggml-cann/common.h @@ -374,7 +374,6 @@ struct ggml_backend_cann_context { #endif cann_task_queue task_queue; bool async_mode; - bool support_set_rows; // Rope Cache void* rope_init_ptr = nullptr; void* rope_sin_ptr = nullptr; @@ -400,14 +399,6 @@ struct ggml_backend_cann_context { async_mode = parse_bool(get_env("GGML_CANN_ASYNC_MODE").value_or("")); GGML_LOG_INFO("%s: device %d async operator submission is %s\n", __func__, device, async_mode ? "ON" : "OFF"); - - support_set_rows = parse_bool(get_env("LLAMA_SET_ROWS").value_or("")); - GGML_LOG_INFO("%s: LLAMA_SET_ROWS is %s\n", __func__, support_set_rows ? "ON" : "OFF"); - - if (!support_set_rows) { - GGML_LOG_INFO("%s: CANN Graph currently only supports execution when LLAMA_SET_ROWS is ON. " - "Falling back to eager mode.\n", __func__); - } } /** diff --git a/src/ggml-cann/ggml-cann.cpp b/src/ggml-cann/ggml-cann.cpp index 81215425..558121df 100755 --- a/src/ggml-cann/ggml-cann.cpp +++ b/src/ggml-cann/ggml-cann.cpp @@ -2251,11 +2251,6 @@ static enum ggml_status ggml_backend_cann_graph_compute( bool use_cann_graph = true; bool cann_graph_update_required = false; - // check environment LLAMA_SET_ROWS - if (!cann_ctx->support_set_rows) { - use_cann_graph = false; - } - if (use_cann_graph) { if (cann_ctx->cann_graph == nullptr) { cann_ctx->cann_graph.reset(new ggml_cann_graph());