From 12cade118e7f620e9daace519c8ad8d18be41b74 Mon Sep 17 00:00:00 2001 From: Neo Zhang Jianyu Date: Tue, 8 Apr 2025 15:03:21 +0800 Subject: [PATCH] Revert "sycl:remove redundant memcopy in function ggml_backend_sycl_buffer_set_tensor" (llama/12812) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit * Revert "sycl: remove redundant memcopy in function ggml_backend_sycl_buffer_s…" This reverts commit 518a01480eb3a7c80a4951b430db9dee55428310. * Update ggml/src/ggml-sycl/ggml-sycl.cpp * Update ggml/src/ggml-sycl/ggml-sycl.cpp * rm tail space --- ggml/src/ggml-sycl/ggml-sycl.cpp | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/ggml/src/ggml-sycl/ggml-sycl.cpp b/ggml/src/ggml-sycl/ggml-sycl.cpp index f6f288d9..89715eae 100644 --- a/ggml/src/ggml-sycl/ggml-sycl.cpp +++ b/ggml/src/ggml-sycl/ggml-sycl.cpp @@ -372,9 +372,14 @@ static void ggml_backend_sycl_buffer_set_tensor(ggml_backend_buffer_t buffer, auto stream = &(dpct::dev_mgr::instance().get_device(ctx->device).default_queue()); SYCL_CHECK( CHECK_TRY_ERROR(dpct::dev_mgr::instance().get_device(ctx->device).queues_wait_and_throw())); + // Note: Use host buffer to save the data from mmap(), then copy to device. It's workaround for mmap() issue on PVC GPU. + // This function will be called during load model from disk. Use memory buffer replace dynamic won't save more time and brings potential memory leak risk here. + char* host_buf = (char*)malloc(size); + memcpy(host_buf, data, size); SYCL_CHECK( - CHECK_TRY_ERROR((*stream).memcpy((char *)tensor->data + offset, data, size) + CHECK_TRY_ERROR((*stream).memcpy((char *)tensor->data + offset, host_buf, size) .wait())); + free(host_buf); } catch (sycl::exception const &exc) { std::cerr << exc.what() << "Exception caught at file:" << __FILE__