Mercurial > hg > truffle
diff src/gpu/ptx/vm/gpu_ptx.cpp @ 10879:d55f24eac4b1
PTX support for Linux
author | Morris Meyer <morris.meyer@oracle.com> |
---|---|
date | Thu, 25 Jul 2013 22:15:30 -0400 |
parents | |
children | 6a2d65cb5d7d |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/src/gpu/ptx/vm/gpu_ptx.cpp Thu Jul 25 22:15:30 2013 -0400 @@ -0,0 +1,316 @@ +/* + * Copyright (c) 2013, Oracle and/or its affiliates. All rights reserved. + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. + * + * This code is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 only, as + * published by the Free Software Foundation. + * + * This code is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License + * version 2 for more details (a copy is included in the LICENSE file that + * accompanied this code). + * + * You should have received a copy of the GNU General Public License version + * 2 along with this work; if not, write to the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. + * + * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA + * or visit www.oracle.com if you need additional information or have any + * questions. + * + */ + +#include "precompiled.hpp" +#include "runtime/gpu.hpp" +#include "utilities/globalDefinitions.hpp" +#include "utilities/ostream.hpp" + +void * gpu::Ptx::_device_context; + +gpu::Ptx::cuda_cu_init_func_t gpu::Ptx::_cuda_cu_init; +gpu::Ptx::cuda_cu_ctx_create_func_t gpu::Ptx::_cuda_cu_ctx_create; +gpu::Ptx::cuda_cu_ctx_detach_func_t gpu::Ptx::_cuda_cu_ctx_detach; +gpu::Ptx::cuda_cu_ctx_synchronize_func_t gpu::Ptx::_cuda_cu_ctx_synchronize; +gpu::Ptx::cuda_cu_device_get_count_func_t gpu::Ptx::_cuda_cu_device_get_count; +gpu::Ptx::cuda_cu_device_get_name_func_t gpu::Ptx::_cuda_cu_device_get_name; +gpu::Ptx::cuda_cu_device_get_func_t gpu::Ptx::_cuda_cu_device_get; +gpu::Ptx::cuda_cu_device_compute_capability_func_t gpu::Ptx::_cuda_cu_device_compute_capability; +gpu::Ptx::cuda_cu_device_get_attribute_func_t gpu::Ptx::_cuda_cu_device_get_attribute; +gpu::Ptx::cuda_cu_launch_kernel_func_t gpu::Ptx::_cuda_cu_launch_kernel; +gpu::Ptx::cuda_cu_module_get_function_func_t gpu::Ptx::_cuda_cu_module_get_function; +gpu::Ptx::cuda_cu_module_load_data_ex_func_t gpu::Ptx::_cuda_cu_module_load_data_ex; + +void gpu::probe_linkage() { +#if defined(__APPLE__) || defined(LINUX) + set_gpu_linkage(gpu::Ptx::probe_linkage()); +#else + set_gpu_linkage(false); +#endif +} + +void gpu::initialize_gpu() { + if (gpu::has_gpu_linkage()) { + set_initialized(gpu::Ptx::initialize_gpu()); + } +} + +void * gpu::generate_kernel(unsigned char *code, int code_len, const char *name) { + if (gpu::has_gpu_linkage()) { + return (gpu::Ptx::generate_kernel(code, code_len, name)); + } else { + return NULL; + } +} + +bool gpu::execute_kernel(address kernel) { + if (gpu::has_gpu_linkage()) { + return (gpu::Ptx::execute_kernel(kernel)); + } else { + return false; + } +} + +bool gpu::Ptx::initialize_gpu() { + + /* Initialize CUDA driver API */ + int status = _cuda_cu_init(0); + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("Failed to initialize CUDA device"); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("CUDA driver initialization: Success"); + } + + /* Get the number of compute-capable device count */ + int device_count = 0; + status = _cuda_cu_device_get_count(&device_count); + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get compute-capable device count"); + return false; + } + + if (device_count == 0) { + tty->print_cr("[CUDA] Found no device supporting CUDA"); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Number of compute-capable devices found: %d", device_count); + } + + /* Get the handle to the first compute device */ + int device_id = 0; + /* Compute-capable device handle */ + int cu_device = 0; + status = _cuda_cu_device_get(&cu_device, device_id); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get handle of first compute-capable device i.e., the one at ordinal: %d", device_id); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Got the handle of first compute-device"); + } + + /* Get device attributes */ + int minor, major; + status = _cuda_cu_device_get_attribute(&minor, GRAAL_CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MINOR, cu_device); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get minor attribute of device: %d", cu_device); + return false; + } + + status = _cuda_cu_device_get_attribute(&major, GRAAL_CU_DEVICE_ATTRIBUTE_COMPUTE_CAPABILITY_MAJOR, cu_device); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get major attribute of device: %d", cu_device); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Compatibility version of device %d: %d.%d", cu_device, major, minor); + } + + /* Get device name */ + char device_name[256]; + status = _cuda_cu_device_get_name(device_name, 256, cu_device); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get name of device: %d", cu_device); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Using %s", device_name); + } + + /* Create CUDA context */ + status = _cuda_cu_ctx_create(&_device_context, 0, cu_device); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to create CUDA context for device: %d", cu_device); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Success: Created context for device: %d", cu_device); + } + + return true; +} + +void *gpu::Ptx::generate_kernel(unsigned char *code, int code_len, const char *name) { + + struct CUmod_st * cu_module; + // Use three JIT compiler options + const unsigned int jit_num_options = 3; + int *jit_options = NEW_C_HEAP_ARRAY(int, jit_num_options, mtCompiler); + void **jit_option_values = NEW_C_HEAP_ARRAY(void *, jit_num_options, mtCompiler); + + // Set up PTX JIT compiler options + // 1. set size of compilation log buffer + int jit_log_buffer_size = 1024; + jit_options[0] = GRAAL_CU_JIT_INFO_LOG_BUFFER_SIZE_BYTES; + jit_option_values[0] = (void *)(size_t)jit_log_buffer_size; + + // 2. set pointer to compilation log buffer + char *jit_log_buffer = NEW_C_HEAP_ARRAY(char, jit_log_buffer_size, mtCompiler); + jit_options[1] = GRAAL_CU_JIT_INFO_LOG_BUFFER; + jit_option_values[1] = jit_log_buffer; + + // 3. set pointer to set the Maximum # of registers (32) for the kernel + int jit_register_count = 32; + jit_options[2] = GRAAL_CU_JIT_MAX_REGISTERS; + jit_option_values[2] = (void *)(size_t)jit_register_count; + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] PTX Kernel\n%s", code); + tty->print_cr("[CUDA] Function name : %s", name); + + } + + /* Load module's data with compiler options */ + int status = _cuda_cu_module_load_data_ex(&cu_module, code, jit_num_options, + jit_options, (void **)jit_option_values); + if (status != GRAAL_CUDA_SUCCESS) { + if (status == GRAAL_CUDA_ERROR_NO_BINARY_FOR_GPU) { + tty->print_cr("[CUDA] Check for malformed PTX kernel or incorrect PTX compilation options"); + } + tty->print_cr("[CUDA] *** Error (%d) Failed to load module data with online compiler options for method %s", + status, name); + return NULL; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Loaded data for PTX Kernel"); + } + + struct CUfunc_st * cu_function; + + status = _cuda_cu_module_get_function(&cu_function, cu_module, name); + + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to get function %s", name); + return NULL; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Got function handle for %s", name); + } + return cu_function; +} + +bool gpu::Ptx::execute_kernel(address kernel) { + // grid dimensionality + unsigned int gridX = 1; + unsigned int gridY = 1; + unsigned int gridZ = 1; + + // thread dimensionality + unsigned int blockX = 1; + unsigned int blockY = 1; + unsigned int blockZ = 1; + + int *cu_function = (int *)kernel; + + if (kernel == NULL) { + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] launching kernel"); + } + int status = _cuda_cu_launch_kernel(cu_function, + gridX, gridY, gridZ, + blockX, blockY, blockZ, + 0, NULL, NULL, NULL); + if (status != GRAAL_CUDA_SUCCESS) { + tty->print_cr("[CUDA] Failed to launch kernel"); + return false; + } + + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Success: Kernel Launch"); + } + return status == 0; // GRAAL_CUDA_SUCCESS +} + +#if defined(LINUX) +static const char cuda_library_name[] = "libcuda.so"; +#elif defined(__APPLE__) +static char const cuda_library_name[] = "/usr/local/cuda/lib/libcuda.dylib"; +#else +static char const cuda_library_name[] = ""; +#endif + +bool gpu::Ptx::probe_linkage() { + if (cuda_library_name != NULL) { + void *handle = dlopen(cuda_library_name, RTLD_LAZY); + if (handle != NULL) { + _cuda_cu_init = + CAST_TO_FN_PTR(cuda_cu_init_func_t, dlsym(handle, "cuInit")); + _cuda_cu_ctx_create = + CAST_TO_FN_PTR(cuda_cu_ctx_create_func_t, dlsym(handle, "cuCtxCreate")); + _cuda_cu_ctx_detach = + CAST_TO_FN_PTR(cuda_cu_ctx_detach_func_t, dlsym(handle, "cuCtxDetach")); + _cuda_cu_ctx_synchronize = + CAST_TO_FN_PTR(cuda_cu_ctx_synchronize_func_t, dlsym(handle, "cuCtxSynchronize")); + _cuda_cu_device_get_count = + CAST_TO_FN_PTR(cuda_cu_device_get_count_func_t, dlsym(handle, "cuDeviceGetCount")); + _cuda_cu_device_get_name = + CAST_TO_FN_PTR(cuda_cu_device_get_name_func_t, dlsym(handle, "cuDeviceGetName")); + _cuda_cu_device_get = + CAST_TO_FN_PTR(cuda_cu_device_get_func_t, dlsym(handle, "cuDeviceGet")); + _cuda_cu_device_compute_capability = + CAST_TO_FN_PTR(cuda_cu_device_compute_capability_func_t, dlsym(handle, "cuDeviceComputeCapability")); + _cuda_cu_device_get_attribute = + CAST_TO_FN_PTR(cuda_cu_device_get_attribute_func_t, dlsym(handle, "cuDeviceGetAttribute")); + _cuda_cu_module_get_function = + CAST_TO_FN_PTR(cuda_cu_module_get_function_func_t, dlsym(handle, "cuModuleGetFunction")); + _cuda_cu_module_load_data_ex = + CAST_TO_FN_PTR(cuda_cu_module_load_data_ex_func_t, dlsym(handle, "cuModuleLoadDataEx")); + _cuda_cu_launch_kernel = + CAST_TO_FN_PTR(cuda_cu_launch_kernel_func_t, dlsym(handle, "cuLaunchKernel")); + if (TraceGPUInteraction) { + tty->print_cr("[CUDA] Success: library linkage"); + } + return true; + } else { + // Unable to dlopen libcuda + tty->print_cr("Use LD_LIBRARY_PATH (or other means) to specify installed location of CUDA library"); + return false; + } + } else { + tty->print_cr("Unsupported CUDA platform"); + return false; + } + tty->print_cr("Failed to find CUDA linkage"); + return false; +} +