mirror of
https://gitlab.freedesktop.org/mesa/mesa.git
synced 2026-05-26 12:28:12 +02:00
2c64e12462
Add optional OA performance counter collection around each execute() call. Examples: ``` # List all profiles and counters, with descriptions. $ executor --oa list # Collect all counters from a profile. $ executor --oa ComputeBasic file.lua # Collect a subset of counters from a profile, separated by comma. $ executor --oa ComputeBasic:GpuTime,AvgGpuCoreFrequency file.lua # By default use ComputeBasic profile, so counter names only also work. $ executor --oa GpuTime file.lua ``` The selected counters are printed to stdout after the script finishes, or written to a file specified by --oa-csv FILENAME. Assisted-by: Pi coding agent (GPT-5.5) Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/41610>
108 lines
2.1 KiB
C
108 lines
2.1 KiB
C
/*
|
|
* Copyright © 2024 Intel Corporation
|
|
* SPDX-License-Identifier: MIT
|
|
*/
|
|
|
|
#ifndef EXECUTOR_H
|
|
#define EXECUTOR_H
|
|
|
|
#include <stdbool.h>
|
|
#include <stdint.h>
|
|
|
|
#include "intel/dev/intel_device_info.h"
|
|
#include "intel/isl/isl.h"
|
|
#include "perf/intel_perf.h"
|
|
#include "perf/intel_perf_query.h"
|
|
|
|
typedef struct {
|
|
uint32_t size;
|
|
uint32_t handle;
|
|
void *map;
|
|
void *cursor;
|
|
uint64_t addr;
|
|
} executor_bo;
|
|
|
|
typedef struct {
|
|
void *mem_ctx;
|
|
|
|
struct intel_device_info *devinfo;
|
|
struct isl_device *isl_dev;
|
|
int fd;
|
|
|
|
struct {
|
|
uint32_t ctx_id;
|
|
} i915;
|
|
|
|
struct {
|
|
uint32_t vm_id;
|
|
uint32_t queue_id;
|
|
} xe;
|
|
|
|
struct {
|
|
executor_bo batch;
|
|
executor_bo extra;
|
|
executor_bo data;
|
|
executor_bo perf;
|
|
} bo;
|
|
|
|
bool perf_enabled;
|
|
|
|
struct {
|
|
struct intel_perf_context *ctx;
|
|
struct intel_perf_query_object *obj;
|
|
} perf_query;
|
|
|
|
uint64_t batch_start;
|
|
} executor_context;
|
|
|
|
typedef struct {
|
|
const char *original_src;
|
|
|
|
void *kernel_bin;
|
|
uint32_t kernel_size;
|
|
} executor_params;
|
|
|
|
typedef struct {
|
|
uint64_t offset;
|
|
} executor_address;
|
|
|
|
__attribute__((unused)) static uint64_t
|
|
executor_combine_address(void *data, void *location,
|
|
executor_address address, uint32_t delta)
|
|
{
|
|
return address.offset + delta;
|
|
}
|
|
|
|
executor_address executor_address_of_ptr(executor_bo *bo, void *ptr);
|
|
|
|
void *executor_alloc_bytes(executor_bo *bo, uint32_t size);
|
|
void *executor_alloc_bytes_aligned(executor_bo *bo, uint32_t size, uint32_t alignment);
|
|
|
|
void failf(const char *fmt, ...) PRINTFLIKE(1, 2);
|
|
|
|
const char *executor_apply_macros(executor_context *ec, const char *original_src);
|
|
|
|
#ifdef genX
|
|
# include "executor_genx.h"
|
|
#else
|
|
# define genX(x) gfx9_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
# define genX(x) gfx11_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
# define genX(x) gfx12_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
# define genX(x) gfx125_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
# define genX(x) gfx20_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
# define genX(x) gfx30_##x
|
|
# include "executor_genx.h"
|
|
# undef genX
|
|
#endif
|
|
|
|
#endif /* EXECUTOR_H */
|