AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
cd4b7cea858bb1169f1aa3b8bcdeca1ad0022c2e
rocm-systems/rocminfo.cc
T
Matt Arsenault cd4b7cea85 Change check for /dev/kfd permissions 
On newer kernels, it seems the group was changed from "video" to
"render". The check for video group misled me for quite some time, so
relax this check. Rather than specifically checking for video group
ownership, first see if /dev/kfd can be open as read-write. Then
diagnose whether the user belongs to the group that owns it, rather
than hardcoding the video group.

Change-Id: I9e65427363e9a5cdba802e09cee2f40fb80520ed
2020-05-01 10:14:21 -04:00

1172 lines
39 KiB
C++
Executable File
Raw Blame History

/*
* =============================================================================
* ROC Runtime Conformance Release License
* =============================================================================
* The University of Illinois/NCSA
* Open Source License (NCSA)
*
* Copyright (c) 2017, Advanced Micro Devices, Inc.
* All rights reserved.
*
* Developed by:
*
* AMD Research and AMD ROC Software Development
*
* Advanced Micro Devices, Inc.
*
* www.amd.com
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal with the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* - Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimers.
* - Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimers in
* the documentation and/or other materials provided with the distribution.
* - Neither the names of <Name of Development Group, Name of Institution>,
* nor the names of its contributors may be used to endorse or promote
* products derived from this Software without specific prior written
* permission.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE CONTRIBUTORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
* OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
* ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS WITH THE SOFTWARE.
*
*/
#include <stdio.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <fcntl.h>
#include <grp.h>
#include <unistd.h>
#include <pwd.h>
#include <vector>
#include <string>
#include <sstream>
#include "hsa/hsa.h"
#include "hsa/hsa_ext_amd.h"
#define COL_BLU "\x1B[34m"
#define COL_KCYN "\x1B[36m"
#define COL_GRN "\x1B[32m"
#define COL_NRM "\x1B[0m"
#define COL_RED "\x1B[31m"
#define COL_MAG "\x1B[35m"
#define COL_WHT "\x1B[37m"
#define COL_YEL "\x1B[33m"
#define COL_RESET "\033[0m"
#define RET_IF_HSA_ERR(err) { \
if ((err) != HSA_STATUS_SUCCESS) { \
char err_val[12]; \
char* err_str = NULL; \
if (hsa_status_string(err, \
(const char**)&err_str) != HSA_STATUS_SUCCESS) { \
snprintf(&(err_val[0]), sizeof(err_val), "%#x", (uint32_t)err); \
err_str = &(err_val[0]); \
} \
printf("%shsa api call failure at: %s:%d\n", \
COL_RED, __FILE__, __LINE__); \
printf("%sCall returned %s\n", COL_RED, err_str); \
printf("%s", COL_RESET); \
return (err); \
} \
}
// This structure holds system information acquired through hsa info related
// calls, and is later used for reference when displaying the information.
typedef struct {
uint16_t major, minor;
uint64_t timestamp_frequency = 0;
uint64_t max_wait = 0;
hsa_endianness_t endianness;
hsa_machine_model_t machine_model;
} system_info_t;
// This structure holds agent information acquired through hsa info related
// calls, and is later used for reference when displaying the information.
typedef struct {
char name[64];
char uuid[24];
char vendor_name[64];
char device_mkt_name[64];
hsa_agent_feature_t agent_feature;
hsa_profile_t agent_profile;
hsa_default_float_rounding_mode_t float_rounding_mode;
uint32_t max_queue;
uint32_t queue_min_size;
uint32_t queue_max_size;
hsa_queue_type_t queue_type;
uint32_t node;
hsa_device_type_t device_type;
uint32_t cache_size[4];
uint32_t chip_id;
uint32_t cacheline_size;
uint32_t max_clock_freq;
uint32_t internal_node_id;
uint32_t max_addr_watch_pts;
// HSA_AMD_AGENT_INFO_MEMORY_WIDTH is deprecated, so exclude
// uint32_t mem_max_freq; Not supported by get_info
uint32_t compute_unit;
uint32_t wavefront_size;
uint32_t workgroup_max_size;
uint32_t grid_max_size;
uint32_t fbarrier_max_size;
uint32_t max_waves_per_cu;
uint32_t simds_per_cu;
uint32_t shader_engs;
uint32_t shader_arrs_per_sh_eng;
hsa_isa_t agent_isa;
hsa_dim3_t grid_max_dim;
uint16_t workgroup_max_dim[3];
uint16_t bdf_id;
bool fast_f16;
} agent_info_t;
// This structure holds memory pool information acquired through hsa info
// related calls, and is later used for reference when displaying the
// information.
typedef struct {
uint32_t segment;
size_t pool_size;
bool alloc_allowed;
size_t alloc_granule;
size_t pool_alloc_alignment;
bool pl_access;
uint32_t global_flag;
} pool_info_t;
// This structure holds ISA information acquired through hsa info
// related calls, and is later used for reference when displaying the
// information.
typedef struct {
char *name_str;
uint32_t workgroup_max_size;
hsa_dim3_t grid_max_dim;
uint64_t grid_max_size;
uint32_t fbarrier_max_size;
uint16_t workgroup_max_dim[3];
bool def_rounding_modes[3];
bool base_rounding_modes[3];
bool mach_models[2];
bool profiles[2];
bool fast_f16;
} isa_info_t;
// This structure holds cache information acquired through hsa info
// related calls, and is later used for reference when displaying the
// information.
typedef struct {
char *name_str;
uint8_t level;
uint32_t size;
} cache_info_t;
static const uint32_t kLabelFieldSize = 25;
static const uint32_t kValueFieldSize = 35;
static const uint32_t kIndentSize = 2;
enum rocmi_int_format {
ROCMI_INT_FORMAT_DEC = 1,
ROCMI_INT_FORMAT_HEX = 2,
};
// Make the most common format the default
std::string int_to_string(uint32_t i,
uint32_t fmt = ROCMI_INT_FORMAT_DEC|ROCMI_INT_FORMAT_HEX) {
std::stringstream sd;
std::string ret = "";
bool need_parens = false;
if (fmt & ROCMI_INT_FORMAT_DEC) {
if (need_parens) {
sd << "(";
}
sd << i;
if (need_parens) {
sd << ") ";
}
need_parens = true;
}
if (fmt & ROCMI_INT_FORMAT_HEX) {
if (need_parens) {
sd << "(0x";
}
sd << std::hex << i;
if (need_parens) {
sd << ") ";
}
need_parens = true;
}
return sd.str();
}
static void printLabelInt(char const *l, int d, uint32_t indent_lvl = 0) {
std::string ind(kIndentSize * indent_lvl, ' ');
printf("%s%-*s%-*u\n", ind.c_str(), kLabelFieldSize, l, kValueFieldSize, d);
}
static void printLabelStr(char const *l, char const *s,
uint32_t indent_lvl = 0) {
std::string ind(kIndentSize * indent_lvl, ' ');
printf("%s%-*s%-*s\n", ind.c_str(), kLabelFieldSize, l, kValueFieldSize, s);
}
static void printLabelStr(char const *l, std::string const &s,
uint32_t indent_lvl = 0) {
std::string ind(kIndentSize * indent_lvl, ' ');
printf("%s%-*s%-*s\n", ind.c_str(), kLabelFieldSize, l, kValueFieldSize,
s.c_str());
}
static void printLabel(char const *l, bool newline = false,
uint32_t indent_lvl = 0) {
std::string ind(kIndentSize * indent_lvl, ' ');
printf("%s%-*s", ind.c_str(), kLabelFieldSize, l);
if (newline) {
printf("\n");
}
}
static void printValueStr(char const *s, bool newline = true) {
printf("%-*s\n", kValueFieldSize, s);
}
// Acquire system information
static hsa_status_t AcquireSystemInfo(system_info_t *sys_info) {
hsa_status_t err;
// Get Major and Minor version of runtime
err = hsa_system_get_info(HSA_SYSTEM_INFO_VERSION_MAJOR, &sys_info->major);
RET_IF_HSA_ERR(err);
err = hsa_system_get_info(HSA_SYSTEM_INFO_VERSION_MINOR, &sys_info->minor);
RET_IF_HSA_ERR(err);
// Get timestamp frequency
err = hsa_system_get_info(HSA_SYSTEM_INFO_TIMESTAMP_FREQUENCY,
&sys_info->timestamp_frequency);
RET_IF_HSA_ERR(err);
// Get maximum duration of a signal wait operation
err = hsa_system_get_info(HSA_SYSTEM_INFO_SIGNAL_MAX_WAIT,
&sys_info->max_wait);
RET_IF_HSA_ERR(err);
// Get Endianness of the system
err = hsa_system_get_info(HSA_SYSTEM_INFO_ENDIANNESS, &sys_info->endianness);
RET_IF_HSA_ERR(err);
// Get machine model info
err = hsa_system_get_info(HSA_SYSTEM_INFO_MACHINE_MODEL,
&sys_info->machine_model);
RET_IF_HSA_ERR(err);
return err;
}
static void DisplaySystemInfo(system_info_t const *sys_info) {
printLabel("Runtime Version:");
printf("%d.%d\n", sys_info->major, sys_info->minor);
printLabel("System Timestamp Freq.:");
printf("%fMHz\n", sys_info->timestamp_frequency / 1e6);
printLabel("Sig. Max Wait Duration:");
printf("%lu (0x%lX) (timestamp count)\n", sys_info->max_wait,
sys_info->max_wait);
printLabel("Machine Model:");
if (HSA_MACHINE_MODEL_SMALL == sys_info->machine_model) {
printValueStr("SMALL");
} else if (HSA_MACHINE_MODEL_LARGE == sys_info->machine_model) {
printValueStr("LARGE");
}
printLabel("System Endianness:");
if (HSA_ENDIANNESS_LITTLE == sys_info->endianness) {
printValueStr("LITTLE");
} else if (HSA_ENDIANNESS_BIG == sys_info->endianness) {
printValueStr("BIG");
}
printf("\n");
}
static hsa_status_t
AcquireAgentInfo(hsa_agent_t agent, agent_info_t *agent_i) {
hsa_status_t err;
// Get agent name and vendor
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_NAME, agent_i->name);
RET_IF_HSA_ERR(err);
// Get UUID, an Ascii string, of a ROCm device
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_UUID,
&agent_i->uuid);
// Get device's vendor name
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_VENDOR_NAME,
&agent_i->vendor_name);
RET_IF_HSA_ERR(err);
// Get device marketing name
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_PRODUCT_NAME,
&agent_i->device_mkt_name);
RET_IF_HSA_ERR(err);
// Get agent feature
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_FEATURE,
&agent_i->agent_feature);
RET_IF_HSA_ERR(err);
// Get profile supported by the agent
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_PROFILE,
&agent_i->agent_profile);
RET_IF_HSA_ERR(err);
// Get floating-point rounding mode
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_DEFAULT_FLOAT_ROUNDING_MODE,
&agent_i->float_rounding_mode);
RET_IF_HSA_ERR(err);
// Get max number of queue
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_QUEUES_MAX,
&agent_i->max_queue);
RET_IF_HSA_ERR(err);
// Get queue min size
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_QUEUE_MIN_SIZE,
&agent_i->queue_min_size);
RET_IF_HSA_ERR(err);
// Get queue max size
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_QUEUE_MAX_SIZE,
&agent_i->queue_max_size);
RET_IF_HSA_ERR(err);
// Get queue type
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_QUEUE_TYPE,
&agent_i->queue_type);
RET_IF_HSA_ERR(err);
// Get agent node
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_NODE, &agent_i->node);
RET_IF_HSA_ERR(err);
// Get device type
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_DEVICE,
&agent_i->device_type);
RET_IF_HSA_ERR(err);
if (HSA_DEVICE_TYPE_GPU == agent_i->device_type) {
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_ISA, &agent_i->agent_isa);
RET_IF_HSA_ERR(err);
}
// Get cache size
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_CACHE_SIZE,
agent_i->cache_size);
RET_IF_HSA_ERR(err);
// Get chip id
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_CHIP_ID,
&agent_i->chip_id);
RET_IF_HSA_ERR(err);
// Get cacheline size
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_CACHELINE_SIZE,
&agent_i->cacheline_size);
RET_IF_HSA_ERR(err);
// Get Max clock frequency
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_MAX_CLOCK_FREQUENCY,
&agent_i->max_clock_freq);
RET_IF_HSA_ERR(err);
// Internal Driver node ID
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_DRIVER_NODE_ID,
&agent_i->internal_node_id);
RET_IF_HSA_ERR(err);
// Max number of watch points on mem. addr. ranges to generate exeception
// events
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_MAX_ADDRESS_WATCH_POINTS,
&agent_i->max_addr_watch_pts);
RET_IF_HSA_ERR(err);
// Get Agent BDFID
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_BDFID, &agent_i->bdf_id);
RET_IF_HSA_ERR(err);
// Get Max Memory Clock
// Not supported by hsa_agent_get_info
// err = hsa_agent_get_info(agent,d
// (hsa_agent_info_t)HSA_AMD_AGENT_INFO_MEMORY_MAX_FREQUENCY,
// &agent_i->mem_max_freq);
// RET_IF_HSA_ERR(err);
// Get Num SIMDs per CU
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_NUM_SIMDS_PER_CU,
&agent_i->simds_per_cu);
RET_IF_HSA_ERR(err);
// Get Num Shader Engines
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_NUM_SHADER_ENGINES,
&agent_i->shader_engs);
RET_IF_HSA_ERR(err);
// Get Num Shader Arrays per Shader engine
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_NUM_SHADER_ARRAYS_PER_SE,
&agent_i->shader_arrs_per_sh_eng);
RET_IF_HSA_ERR(err);
// Get number of Compute Unit
err = hsa_agent_get_info(agent,
(hsa_agent_info_t) HSA_AMD_AGENT_INFO_COMPUTE_UNIT_COUNT,
&agent_i->compute_unit);
RET_IF_HSA_ERR(err);
// Check if the agent is kernel agent
if (agent_i->agent_feature & HSA_AGENT_FEATURE_KERNEL_DISPATCH) {
// Get flaf of fast_f16 operation
err = hsa_agent_get_info(agent,
HSA_AGENT_INFO_FAST_F16_OPERATION, &agent_i->fast_f16);
RET_IF_HSA_ERR(err);
// Get wavefront size
err = hsa_agent_get_info(agent,
HSA_AGENT_INFO_WAVEFRONT_SIZE, &agent_i->wavefront_size);
RET_IF_HSA_ERR(err);
// Get max total number of work-items in a workgroup
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_WORKGROUP_MAX_SIZE,
&agent_i->workgroup_max_size);
RET_IF_HSA_ERR(err);
// Get max number of work-items of each dimension of a work-group
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_WORKGROUP_MAX_DIM,
&agent_i->workgroup_max_dim);
RET_IF_HSA_ERR(err);
// Get max number of a grid per dimension
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_GRID_MAX_DIM,
&agent_i->grid_max_dim);
RET_IF_HSA_ERR(err);
// Get max total number of work-items in a grid
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_GRID_MAX_SIZE,
&agent_i->grid_max_size);
RET_IF_HSA_ERR(err);
// Get max number of fbarriers per work group
err = hsa_agent_get_info(agent, HSA_AGENT_INFO_FBARRIER_MAX_SIZE,
&agent_i->fbarrier_max_size);
RET_IF_HSA_ERR(err);
err = hsa_agent_get_info(agent,
(hsa_agent_info_t)HSA_AMD_AGENT_INFO_MAX_WAVES_PER_CU,
&agent_i->max_waves_per_cu);
RET_IF_HSA_ERR(err);
}
return err;
}
static void DisplayAgentInfo(agent_info_t *agent_i) {
printLabelStr("Name:", agent_i->name, 1);
printLabelStr("Uuid:", agent_i->uuid, 1);
printLabelStr("Marketing Name:", agent_i->device_mkt_name, 1);
printLabelStr("Vendor Name:", agent_i->vendor_name, 1);
printLabel("Feature:", false, 1);
if (agent_i->agent_feature & HSA_AGENT_FEATURE_KERNEL_DISPATCH
&& agent_i->agent_feature & HSA_AGENT_FEATURE_AGENT_DISPATCH) {
printValueStr("KERNEL_DISPATCH & AGENT_DISPATCH");
} else if (agent_i->agent_feature & HSA_AGENT_FEATURE_KERNEL_DISPATCH) {
printValueStr("KERNEL_DISPATCH");
} else if (agent_i->agent_feature & HSA_AGENT_FEATURE_AGENT_DISPATCH) {
printValueStr("AGENT_DISPATCH");
} else {
printValueStr("None specified");
}
printLabel("Profile:", false, 1);
if (HSA_PROFILE_BASE == agent_i->agent_profile) {
printValueStr("BASE_PROFILE");
} else if (HSA_PROFILE_FULL == agent_i->agent_profile) {
printValueStr("FULL_PROFILE");
} else {
printValueStr("Unknown");
}
printLabel("Float Round Mode:", false, 1);
if (HSA_DEFAULT_FLOAT_ROUNDING_MODE_ZERO == agent_i->float_rounding_mode) {
printValueStr("ZERO");
} else if (HSA_DEFAULT_FLOAT_ROUNDING_MODE_NEAR ==
agent_i->float_rounding_mode) {
printValueStr("NEAR");
} else {
printValueStr("Not Supported");
}
printLabelStr("Max Queue Number:", int_to_string(agent_i->max_queue), 1);
printLabelStr("Queue Min Size:", int_to_string(agent_i->queue_min_size), 1);
printLabelStr("Queue Max Size:", int_to_string(agent_i->queue_max_size), 1);
if (HSA_QUEUE_TYPE_MULTI == agent_i->queue_type) {
printLabelStr("Queue Type:", "MULTI", 1);
} else if (HSA_QUEUE_TYPE_SINGLE == agent_i->queue_type) {
printLabelStr("Queue Type:", "SINGLE", 1);
} else {
printLabelStr("Queue Type:", "Unknown", 1);
}
printLabelInt("Node:", agent_i->node, 1);
printLabel("Device Type:", false, 1);
if (HSA_DEVICE_TYPE_CPU == agent_i->device_type) {
printValueStr("CPU");
} else if (HSA_DEVICE_TYPE_GPU == agent_i->device_type) {
printValueStr("GPU");
} else {
printValueStr("DSP");
}
printLabel("Cache Info:", true, 1);
for (int i = 0; i < 4; i++) {
if (agent_i->cache_size[i]) {
std::string tmp_str("L");
tmp_str += std::to_string(i+1);
tmp_str += ":";
printLabel(tmp_str.c_str(), false, 2);
// tmp_str = std::to_string(agent_i->cache_size[i]/1024);
tmp_str = int_to_string(agent_i->cache_size[i]/1024);
tmp_str += "KB";
printValueStr(tmp_str.c_str());
}
}
printLabelStr("Chip ID:", int_to_string(agent_i->chip_id), 1);
printLabelStr("Cacheline Size:", int_to_string(agent_i->cacheline_size), 1);
printLabelInt("Max Clock Freq. (MHz):", agent_i->max_clock_freq, 1);
printLabelInt("BDFID:", agent_i->bdf_id, 1);
printLabelInt("Internal Node ID:", agent_i->internal_node_id, 1);
printLabelInt("Compute Unit:", agent_i->compute_unit, 1);
printLabelInt("SIMDs per CU:", agent_i->simds_per_cu, 1);
printLabelInt("Shader Engines:", agent_i->shader_engs, 1);
printLabelInt("Shader Arrs. per Eng.:", agent_i->shader_arrs_per_sh_eng, 1);
printLabelInt("WatchPts on Addr. Ranges:", agent_i->max_addr_watch_pts, 1);
printLabel("Features:", false, 1);
if (agent_i->agent_feature & HSA_AGENT_FEATURE_KERNEL_DISPATCH) {
printf("%s", "KERNEL_DISPATCH ");
}
if (agent_i->agent_feature & HSA_AGENT_FEATURE_AGENT_DISPATCH) {
printf("%s", "AGENT_DISPATCH");
}
if (agent_i->agent_feature == 0) {
printf("None");
}
printf("\n");
if (agent_i->agent_feature & HSA_AGENT_FEATURE_KERNEL_DISPATCH) {
printLabelStr("Fast F16 Operation:",
agent_i->fast_f16 ? "TRUE":"FALSE", 1);
printLabelStr("Wavefront Size:",
int_to_string(agent_i->wavefront_size), 1);
printLabelStr("Workgroup Max Size:",
int_to_string(agent_i->workgroup_max_size), 1);
printLabel("Workgroup Max Size per Dimension:", true, 1);
printLabelStr("x",
int_to_string(static_cast<uint32_t>(agent_i->workgroup_max_dim[0])), 2);
printLabelStr("y",
int_to_string(static_cast<uint32_t>(agent_i->workgroup_max_dim[1])), 2);
printLabelStr("z",
int_to_string(static_cast<uint32_t>(agent_i->workgroup_max_dim[2])), 2);
printLabelStr("Max Waves Per CU:",
int_to_string(agent_i->max_waves_per_cu), 1);
printLabelStr("Max Work-item Per CU:",
int_to_string(agent_i->wavefront_size*agent_i->max_waves_per_cu), 1);
printLabelStr("Grid Max Size:", int_to_string(agent_i->grid_max_size), 1);
printLabel("Grid Max Size per Dimension:", true, 1);
printLabelStr("x", int_to_string(agent_i->grid_max_dim.x), 2);
printLabelStr("y", int_to_string(agent_i->grid_max_dim.y), 2);
printLabelStr("z", int_to_string(agent_i->grid_max_dim.z), 2);
printLabelInt("Max fbarriers/Workgrp:", agent_i->fbarrier_max_size, 1);
}
}
static hsa_status_t AcquirePoolInfo(hsa_amd_memory_pool_t pool,
pool_info_t *pool_i) {
hsa_status_t err;
err = hsa_amd_memory_pool_get_info(pool,
HSA_AMD_MEMORY_POOL_INFO_GLOBAL_FLAGS, &pool_i->global_flag);
RET_IF_HSA_ERR(err);
err = hsa_amd_memory_pool_get_info(pool, HSA_AMD_MEMORY_POOL_INFO_SEGMENT,
&pool_i->segment);
RET_IF_HSA_ERR(err);
// Get the size of the POOL
err = hsa_amd_memory_pool_get_info(pool, HSA_AMD_MEMORY_POOL_INFO_SIZE,
&pool_i->pool_size);
RET_IF_HSA_ERR(err);
err = hsa_amd_memory_pool_get_info(pool,
HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_ALLOWED,
&pool_i->alloc_allowed);
RET_IF_HSA_ERR(err);
err = hsa_amd_memory_pool_get_info(pool,
HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_GRANULE,
&pool_i->alloc_granule);
RET_IF_HSA_ERR(err);
err = hsa_amd_memory_pool_get_info(pool,
HSA_AMD_MEMORY_POOL_INFO_RUNTIME_ALLOC_ALIGNMENT,
&pool_i->pool_alloc_alignment);
RET_IF_HSA_ERR(err);
err = hsa_amd_memory_pool_get_info(pool,
HSA_AMD_MEMORY_POOL_INFO_ACCESSIBLE_BY_ALL,
&pool_i->pl_access);
RET_IF_HSA_ERR(err);
return HSA_STATUS_SUCCESS;
}
static void MakeGlobalFlagsString(uint32_t global_flag, std::string* out_str) {
*out_str = "";
std::vector<std::string> flags;
if (HSA_AMD_MEMORY_POOL_GLOBAL_FLAG_KERNARG_INIT & global_flag) {
flags.push_back("KERNARG");
}
if (HSA_AMD_MEMORY_POOL_GLOBAL_FLAG_FINE_GRAINED & global_flag) {
flags.push_back("FINE GRAINED");
}
if (HSA_AMD_MEMORY_POOL_GLOBAL_FLAG_COARSE_GRAINED & global_flag) {
flags.push_back("COARSE GRAINED");
}
if (flags.size() > 0) {
*out_str += flags[0];
}
for (size_t i = 1; i < flags.size(); i++) {
*out_str += ", " + flags[i];
}
}
static void DumpSegment(pool_info_t *pool_i, uint32_t ind_lvl) {
std::string seg_str;
std::string tmp_str;
printLabel("Segment:", false, ind_lvl);
switch (pool_i->segment) {
case HSA_AMD_SEGMENT_GLOBAL:
MakeGlobalFlagsString(pool_i->global_flag, &tmp_str);
seg_str += "GLOBAL; FLAGS: " + tmp_str;
break;
case HSA_AMD_SEGMENT_READONLY:
seg_str += "READONLY";
break;
case HSA_AMD_SEGMENT_PRIVATE:
seg_str += "PRIVATE";
break;
case HSA_AMD_SEGMENT_GROUP:
seg_str += "GROUP";
break;
default:
printf("Not Supported\n");
break;
}
printValueStr(seg_str.c_str());
}
static void DisplayPoolInfo(pool_info_t *pool_i, uint32_t indent) {
DumpSegment(pool_i, indent);
size_t sz = pool_i->pool_size/1024;
printLabelStr("Size:", int_to_string(sz) + "KB", indent);
printLabelStr("Allocatable:", (pool_i->alloc_allowed ? "TRUE" : "FALSE"),
indent);
std::string gr_str = std::to_string(pool_i->alloc_granule/1024)+"KB";
printLabelStr("Alloc Granule:", gr_str.c_str(), indent);
std::string al_str = std::to_string(pool_i->pool_alloc_alignment/1024)+"KB";
printLabelStr("Alloc Alignment:", al_str.c_str(), indent);
printLabelStr("Acessible by all:", (pool_i->pl_access ? "TRUE" : "FALSE"),
indent);
}
static hsa_status_t
AcquireAndDisplayMemPoolInfo(const hsa_amd_memory_pool_t pool,
uint32_t indent) {
hsa_status_t err;
pool_info_t pool_i;
err = AcquirePoolInfo(pool, &pool_i);
RET_IF_HSA_ERR(err);
DisplayPoolInfo(&pool_i, 3);
return err;
}
static hsa_status_t get_pool_info(hsa_amd_memory_pool_t pool, void* data) {
hsa_status_t err;
int* p_int = reinterpret_cast<int*>(data);
(*p_int)++;
std::string pool_str("Pool ");
pool_str += std::to_string(*p_int);
printLabel(pool_str.c_str(), true, 2);
err = AcquireAndDisplayMemPoolInfo(pool, 3);
RET_IF_HSA_ERR(err);
return err;
}
static hsa_status_t AcquireISAInfo(hsa_isa_t isa, isa_info_t *isa_i) {
hsa_status_t err;
uint32_t name_len;
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_NAME_LENGTH, &name_len);
RET_IF_HSA_ERR(err);
isa_i->name_str = new char[name_len];
if (isa_i->name_str == nullptr) {
return HSA_STATUS_ERROR_OUT_OF_RESOURCES;
}
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_NAME, isa_i->name_str);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_MACHINE_MODELS,
isa_i->mach_models);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_PROFILES, isa_i->profiles);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_DEFAULT_FLOAT_ROUNDING_MODES,
isa_i->def_rounding_modes);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa,
HSA_ISA_INFO_BASE_PROFILE_DEFAULT_FLOAT_ROUNDING_MODES,
isa_i->base_rounding_modes);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_FAST_F16_OPERATION,
&isa_i->fast_f16);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_WORKGROUP_MAX_DIM,
&isa_i->workgroup_max_dim);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_WORKGROUP_MAX_SIZE,
&isa_i->workgroup_max_size);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_GRID_MAX_DIM,
&isa_i->grid_max_dim);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_GRID_MAX_SIZE,
&isa_i->grid_max_size);
RET_IF_HSA_ERR(err);
err = hsa_isa_get_info_alt(isa, HSA_ISA_INFO_FBARRIER_MAX_SIZE,
&isa_i->fbarrier_max_size);
RET_IF_HSA_ERR(err);
return err;
}
static void DisplayISAInfo(isa_info_t *isa_i, uint32_t indent) {
printLabelStr("Name:", isa_i->name_str, indent);
std::string models("");
if (isa_i->mach_models[HSA_MACHINE_MODEL_SMALL]) {
models = "HSA_MACHINE_MODEL_SMALL ";
}
if (isa_i->mach_models[HSA_MACHINE_MODEL_LARGE]) {
models += "HSA_MACHINE_MODEL_LARGE";
}
printLabelStr("Machine Models:", models.c_str(), indent);
std::string profiles("");
if (isa_i->profiles[HSA_PROFILE_BASE]) {
profiles = "HSA_PROFILE_BASE ";
}
if (isa_i->profiles[HSA_PROFILE_FULL]) {
profiles += "HSA_PROFILE_FULL";
}
printLabelStr("Profiles:", profiles.c_str(), indent);
std::string rounding_modes("");
if (isa_i->def_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_DEFAULT]) {
rounding_modes = "DEFAULT ";
}
if (isa_i->def_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_ZERO]) {
rounding_modes += "ZERO ";
}
if (isa_i->def_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_NEAR]) {
rounding_modes += "NEAR";
}
printLabelStr("Default Rounding Mode:", rounding_modes.c_str(), indent);
rounding_modes = "";
if (isa_i->base_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_DEFAULT]) {
rounding_modes = "DEFAULT ";
}
if (isa_i->base_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_ZERO]) {
rounding_modes += "ZERO ";
}
if (isa_i->base_rounding_modes[HSA_DEFAULT_FLOAT_ROUNDING_MODE_NEAR]) {
rounding_modes += "NEAR";
}
printLabelStr("Default Rounding Mode:", rounding_modes.c_str(), indent);
printLabelStr("Fast f16:", (isa_i->fast_f16 ? "TRUE" : "FALSE"), indent);
printLabelStr("Workgroup Max Size:",
int_to_string(isa_i->workgroup_max_size), indent);
printLabel("Workgroup Max Size per Dimension:", true, indent);
printLabelStr("x", int_to_string(
static_cast<uint32_t>(isa_i->workgroup_max_dim[0])), indent+1);
printLabelStr("y", int_to_string(
static_cast<uint32_t>(isa_i->workgroup_max_dim[1])), indent+1);
printLabelStr("z", int_to_string(
static_cast<uint32_t>(isa_i->workgroup_max_dim[2])), indent+1);
printLabelStr("Grid Max Size:", int_to_string(isa_i->grid_max_size), indent);
printLabel("Grid Max Size per Dimension:", true, indent);
printLabelStr("x", int_to_string(isa_i->grid_max_dim.x), indent+1);
printLabelStr("y", int_to_string(isa_i->grid_max_dim.y), indent+1);
printLabelStr("z", int_to_string(isa_i->grid_max_dim.z), indent+1);
printLabelInt("FBarrier Max Size:", isa_i->fbarrier_max_size, indent);
}
static hsa_status_t
AcquireAndDisplayISAInfo(const hsa_isa_t isa, uint32_t indent) {
hsa_status_t err;
isa_info_t isa_i;
isa_i.name_str = nullptr;
err = AcquireISAInfo(isa, &isa_i);
RET_IF_HSA_ERR(err);
DisplayISAInfo(&isa_i, 3);
if (isa_i.name_str != nullptr) {
delete []isa_i.name_str;
}
return err;
}
static hsa_status_t get_isa_info(hsa_isa_t isa, void* data) {
hsa_status_t err;
int* isa_int = reinterpret_cast<int*>(data);
(*isa_int)++;
std::string isa_str("ISA ");
isa_str += std::to_string(*isa_int);
printLabel(isa_str.c_str(), true, 2);
err = AcquireAndDisplayISAInfo(isa, 3);
RET_IF_HSA_ERR(err);
return err;
}
// Cache info dump is ifdef'd out as it generates a lot of output that is
// not that interesting. Define ENABLE_CACHE_DUMP if this is of interest.
#ifdef ENABLE_CACHE_DUMP
static void DisplayCacheInfo(cache_info_t *cache_i, uint32_t indent) {
printLabelStr("Name:", cache_i->name_str, indent);
printLabelInt("Level:", cache_i->level, indent);
printLabelInt("Size:", cache_i->size, indent);
}
static hsa_status_t AcquireCacheInfo(hsa_cache_t cache, cache_info_t *cache_i) {
hsa_status_t err;
uint32_t name_len;
err = hsa_cache_get_info(cache, HSA_CACHE_INFO_NAME_LENGTH, &name_len);
RET_IF_HSA_ERR(err);
cache_i->name_str = new char[name_len];
if (cache_i->name_str == nullptr) {
return HSA_STATUS_ERROR_OUT_OF_RESOURCES;
}
err = hsa_cache_get_info(cache, HSA_CACHE_INFO_NAME, cache_i->name_str);
RET_IF_HSA_ERR(err);
err = hsa_cache_get_info(cache, HSA_CACHE_INFO_LEVEL, &cache_i->level);
RET_IF_HSA_ERR(err);
err = hsa_cache_get_info(cache, HSA_CACHE_INFO_SIZE, &cache_i->size);
RET_IF_HSA_ERR(err);
return err;
}
static hsa_status_t
AcquireAndDisplayCacheInfo(const hsa_cache_t cache, uint32_t indent) {
hsa_status_t err;
cache_info_t cache_i;
err = AcquireCacheInfo(cache, &cache_i);
RET_IF_HSA_ERR(err);
DisplayCacheInfo(&cache_i, 3);
if (cache_i.name_str != nullptr) {
delete []cache_i.name_str;
}
return err;
}
static hsa_status_t get_cache_info(hsa_cache_t cache, void* data) {
hsa_status_t err;
int* cache_int = reinterpret_cast<int*>(data);
(*cache_int)++;
std::string cache_str("Cache L");
cache_str += std::to_string(*cache_int);
printLabel(cache_str.c_str(), true, 2);
err = AcquireAndDisplayCacheInfo(cache, 3);
RET_IF_HSA_ERR(err);
return err;
}
#endif // ENABLE_CACHE_DUMP
static hsa_status_t
AcquireAndDisplayAgentInfo(hsa_agent_t agent, void* data) {
int pool_number = 0;
int isa_number = 0;
hsa_status_t err;
agent_info_t agent_i;
int *agent_number = reinterpret_cast<int*>(data);
(*agent_number)++;
err = AcquireAgentInfo(agent, &agent_i);
RET_IF_HSA_ERR(err);
std::string ind(kIndentSize, ' ');
printLabel("*******", true);
std::string agent_ind("Agent ");
agent_ind += std::to_string(*agent_number).c_str();
printLabel(agent_ind.c_str(), true);
printLabel("*******", true);
DisplayAgentInfo(&agent_i);
printLabel("Pool Info:", true, 1);
err = hsa_amd_agent_iterate_memory_pools(agent, get_pool_info, &pool_number);
RET_IF_HSA_ERR(err);
printLabel("ISA Info:", true, 1);
err = hsa_agent_iterate_isas(agent, get_isa_info, &isa_number);
if (err == HSA_STATUS_ERROR_INVALID_AGENT) {
printLabel("N/A", true, 2);
return HSA_STATUS_SUCCESS;
}
RET_IF_HSA_ERR(err);
#if ENABLE_CACHE_DUMP
int cache_number = 0;
printLabel("Cache Info:", true, 1);
err = hsa_agent_iterate_caches(agent, get_cache_info, &cache_number);
if (err == HSA_STATUS_ERROR_INVALID_AGENT) {
printLabel("N/A", true, 2);
return HSA_STATUS_SUCCESS;
}
#endif
RET_IF_HSA_ERR(err);
return HSA_STATUS_SUCCESS;
}
void CheckInitialState(void) {
// Check kernel module for ROCk is loaded
FILE *fd = popen("lsmod | grep amdgpu", "r");
char buf[16];
if (fread (buf, 1, sizeof (buf), fd) <= 0) {
printf("%sROCk module is NOT loaded, possibly no GPU devices%s\n",
COL_RED, COL_RESET);
} else {
printf("%sROCk module is loaded%s\n", COL_WHT, COL_RESET);
}
// Check if user belongs to the group for /dev/kfd (e.g. "video" or
// "render")
// @note: User who are not members of "video"
// group cannot access DRM services
char u_name[32];
bool member = false;
struct passwd *pw;
int num_groups = 0;
gid_t *groups;
// Check if we can open /dev/kfd as read-write. If not, try to
// diagnose common reasons why you can't.
int open_kfd = open("/dev/kfd", O_RDWR);
if (open_kfd >= 0) {
close(open_kfd);
printf("%sAble to open /dev/kfd read-write%s\n",
COL_WHT, COL_RESET);
return;
}
printf("%sUnable to open /dev/kfd read-write: %s%s\n",
COL_RED, strerror(errno), COL_RESET);
const char *kfd_gr_name = NULL;
struct stat sb;
if (stat("/dev/kfd", &sb) == 0) {
// The owner of kfd was renamed, so avoid hard-coding the
// name. Check whatever group owns it.
if (struct group *kfd_gr = getgrgid(sb.st_gid))
kfd_gr_name = kfd_gr->gr_name;
}
if (!kfd_gr_name)
kfd_gr_name = "video";
struct group *gr_s = getgrnam(kfd_gr_name); // NOLINT
if (gr_s == nullptr) {
printf("%sFailed to get group info to check"
" for %s group membership%s\n", COL_RED, kfd_gr_name,
COL_RESET);
return;
}
if (getlogin_r(u_name, 32)) {
printf("%sFailed to get user name to check for"
" %s group membership%s\n", COL_RED, kfd_gr_name,
COL_RESET);
return;
}
pw = getpwnam(u_name); // NOLINT
if (pw == NULL) {
printf("%sFailed to find pwd entry for user %s%s\n",
COL_RED, u_name, COL_RESET);
return;
}
(void)getgrouplist(u_name, pw->pw_gid, NULL, &num_groups);
groups = new gid_t[num_groups];
if (getgrouplist(u_name, pw->pw_gid, groups, &num_groups) == -1) {
printf("%sFailed to get user group list%s\n", COL_RED, COL_RESET);
delete []groups;
return;
}
for (int i = 0; i < num_groups; ++i) {
if (gr_s->gr_gid == groups[i]) {
printf("%s%s is member of %s group%s\n", COL_WHT, u_name, kfd_gr_name, COL_RESET);
member = true;
break;
}
}
if (member == false) {
printf("%s%s is not member of \"%s\" group, the default DRM access "
"group. Users must be a member of the \"%s\" group or another"
" DRM access group in order for ROCm applications to run "
"successfully%s.\n", COL_RED, u_name, kfd_gr_name, kfd_gr_name, COL_RESET);
}
delete []groups;
return;
}
// Print out all static information known to HSA about the target system.
// Throughout this program, the Acquire-type functions make HSA calls to
// interate through HSA objects and then perform HSA get_info calls to
// acccumulate information about those objects. Corresponding to each
// Acquire-type function is a Display* function which display the
// accumulated data in a formatted way.
int main(int argc, char* argv[]) {
hsa_status_t err;
CheckInitialState();
err = hsa_init();
RET_IF_HSA_ERR(err)
// Acquire and display system information
system_info_t sys_info;
// This function will call HSA get_info functions to gather information
// about the system.
err = AcquireSystemInfo(&sys_info);
RET_IF_HSA_ERR(err);
printLabel("=====================", true);
printLabel("HSA System Attributes", true);
printLabel("=====================", true);
DisplaySystemInfo(&sys_info);
// Iterate through every agent and get and display their info
printLabel("==========", true);
printLabel("HSA Agents", true);
printLabel("==========", true);
uint32_t agent_ind = 0;
err = hsa_iterate_agents(AcquireAndDisplayAgentInfo, &agent_ind);
RET_IF_HSA_ERR(err);
printLabel("*** Done ***", true);
err = hsa_shut_down();
RET_IF_HSA_ERR(err);
}
#undef RET_IF_HSA_ERR
Reference in New Issue View Git Blame Copy Permalink