1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
|
# converts output from PoCL's text tracer to JSON for Chromium tracing
#
# idea from:
# https://www.gamasutra.com/view/news/176420/Indepth_Using_Chrometracing_to_view_your_inline_profiling_data.php
require 'tempfile'
require 'json'
require 'getoptlong'
opts = GetoptLong.new(
[ '--help', '-h', GetoptLong::NO_ARGUMENT ],
[ '--output', '-o', GetoptLong::REQUIRED_ARGUMENT ],
[ '--start', '-s', GetoptLong::OPTIONAL_ARGUMENT ],
[ '--end', '-e', GetoptLong::OPTIONAL_ARGUMENT ]
)
slice_start = nil
slice_end = nil
in_f = nil
out_f = nil
opts.each do |opt, arg|
case opt
when '--help'
puts "USAGE: $0 [-s START] [-e END] -o output_file input_file"
puts "START and END are optional timestamps (seconds.fraction) for slicing out a portion of trace. Trace begins at 0.0"
when '--output'
out_f = File.open(arg, "wb")
when '--start'
slice_start = (arg.to_f * 1000000).to_i
when '--end'
slice_end = (arg.to_f * 1000000).to_i
end
end
puts "Input files:"
puts ARGV.inspect
if (ARGV.size == 0)
puts "No input files specified"
exit 1
end
if out_f.nil?
puts "Output not specified"
exit 1
end
if slice_start and slice_end
puts "START: #{slice_start} usec END: #{slice_end} usec"
end
EVENTS = Hash.new
DEPS_FROM = Array.new
DEPS_TO = Array.new
NANOS_IN_SEC = 1_000_000_000
# 109287478675743 | EV ID 1 | DEV 1 | CQ 4 | ndrange_kernel | queued | KERNEL ID 6 | name=ML_BSSN_CL_RHS1
COMMAND_REGEXP = /^(?<timestamp>\d+) \| EV ID (?<event_id>\d+) \| DEV (?<dev_id>\d+) \| CQ (?<queue_id>\d+) \| (?<command_type>\w+) \| (?<evt_status>\w+) \|(?<remainder>.*)/
# DEP | EV ID 1 -> EV ID 2
DEP_REGEXP = /DEP \| EV ID (?<src_id>\d+) -> EV ID (?<dst_id>\d+)/
# KERNEL ID 6 | name=ML_BSSN_CL_RHS1
NDR_REGEXP = /KERNEL ID (?<kernel_id>\d+) \| name=(?<kernel_name>[\.\w]+)/
# MEM ID 34 | size=2744000 | host_ptr=0x5574f7c0b990
# unmap_mem_object | complete | MEM ID 21
MEM_ID_REGEXP = /MEM ID (?<mem_id>\d+)/
# MEM ID FROM %" PRIu64 " | MEM ID TO %" PRIu64 "
COPY_REGEXP = /MEM ID FROM (?<src_id>\d+) \| MEM ID TO (?<dst_id>\d+)/
# MEM ID %" PRIu64 " | FROM DEV %" PRIu64 " | TO DEV %" PRIu64 "
MIGR_REGEXP = /MEM 0 ID (?<mem_id>\d+) \| FROM DEV (?<src_id>\w+) \| TO DEV (?<dst_id>\w+)/
NOP_REGEXP = /MEM 0 ID (?<mem_id>\d+) \| NOP MIGRATION DEV/
def parsecommand(timestamp, match)
event_id = match[:event_id].to_i
dev_id = match[:dev_id].to_i
queue_id = match[:queue_id].to_i
if event_id.nil? or dev_id.nil? or queue_id.nil?
raise StandardError.new "unknown event 1"
end
cmd_type = match[:command_type]
event_status = match[:evt_status]
remainder = match[:remainder]
event_numeric_status = nil
case event_status
when "queued"
event_numeric_status = 0
when "submitted"
event_numeric_status = 1
when "running"
event_numeric_status = 2
when "complete"
event_numeric_status = 3
else
# map errors (<0) to complete
event_numeric_status = 3
end
meta = Hash.new
case cmd_type
when "ndrange_kernel"
match = NDR_REGEXP.match remainder
raise StandardError.new "parse error NDR" unless match
meta[:kernel_id] = match[:kernel_id]
meta[:kernel_name] = match[:kernel_name]
when "read_buffer", "write_buffer", "fill_buffer", "map_buffer", "unmap_mem_object", "read_buffer_rect", "write_buffer_rect"
match = MEM_ID_REGEXP.match remainder
raise StandardError.new "parse error MEM" unless match
meta[:mem_id] = match[:mem_id]
when "copy_buffer", "copy_buffer_rect"
match = COPY_REGEXP.match remainder
raise StandardError.new "parse error COPY" unless match
meta[:src_id] = match[:src_id]
meta[:dst_id] = match[:dst_id]
when "migrate_mem_objects"
match = NOP_REGEXP.match remainder
if match
meta[:mem_id] = match[:mem_id]
meta[:type] = "nop migration"
else
match = MIGR_REGEXP.match remainder
raise StandardError.new "parse error MIGR" unless match
meta[:mem_id] = match[:mem_id]
meta[:src_id] = match[:src_id]
meta[:dst_id] = match[:dst_id]
end
when "barrier", "svm_memcpy"
meta[:type] = "empty"
else
puts "unknown command type: #{cmd_type}"
end
h = EVENTS[event_id]
if h.nil?
h = Hash.new
EVENTS[event_id] = h
end
if h[:dev_id].nil?
h[:event_id] = event_id
h[:dev_id] = dev_id
h[:queue_id] = queue_id
# for NDRange commands, put the kernel name in the "type"
if cmd_type == 'ndrange_kernel'
h[:type] = meta[:kernel_name].to_sym
else
h[:type] = cmd_type.to_sym
end
h[:stat] = Array.new(4)
h[:meta] = meta
end
h[:stat][event_numeric_status] = timestamp
end
def parsedeps(dep_match)
src_id = dep_match[:src_id].to_i
dst_id = dep_match[:dst_id].to_i
h = EVENTS[dst_id]
if h.nil?
h = Hash.new
EVENTS[dst_id] = h
end
if h[:deps].nil?
h[:deps] = Array.new
end
h[:deps] << src_id
DEPS_FROM << src_id
DEPS_TO << dst_id
end
##############################################################################
lines = 0
starting_timestamp = nil
ARGV.each do |file|
puts "Parsing #{file}"
in_f = File.open(file, "rb")
in_f.each_line do |line|
line.strip!
next if line.empty?
cmd_match = COMMAND_REGEXP.match line
dep_match = DEP_REGEXP.match line
if cmd_match
ts = cmd_match[:timestamp].to_i
secs = ts / NANOS_IN_SEC
nsecs = ts % NANOS_IN_SEC
if starting_timestamp.nil?
starting_timestamp = ts
end
cmd_timestamp = ts - starting_timestamp
parsecommand(cmd_timestamp, cmd_match)
elsif dep_match
parsedeps(dep_match)
else
puts "Line did not match anything: \n"
puts line
end
#break if lines > 2000
lines += 1
end
puts "Parsed lines: #{lines}"
in_f.close
end
#############################################################
=begin
'cat' – the category for this event. Useful when doing larger grouping (eg "UnitUpdates")
'name' – the name of this event (eg 'PathingUpdate')
'pid' – the processor ID that spawned this event
'tid' – the thread ID that spawned this event
'ts' – the processor time stamp at the time this event was created
'ph' – the phase or type of this event
'args' – any programmatic metadata that's attached to this event
cat = pocl / pold TODO something more useful ?
name = cmdtype
pid = device
tid = queue
ts = timestamp
ph = status
args = all metadata
=end
event_json = Hash.new
event_json['cat'] = 'pocl'
out_f.puts '['
first_done = false
###############################################
###############################################
event_ids = EVENTS.keys
event_ids.sort!
event_ids.each do |event_id|
# puts EVENTS[event_id].inspect
e = EVENTS[event_id]
stat = e[:stat]
# skip broken events
next if e[:dev_id].nil? or e[:queue_id].nil? or e[:event_id].nil?
# skip events with no timing or incomplete timing
next if stat.nil? or stat[3].nil?
if slice_start and slice_end
# puts "ST: #{(stat[0] / 1000)} END: #{(stat[3] / 1000)}"
next if (stat[0] / 1000) < slice_start
next if (stat[3] / 1000) > slice_end
end
event_json['id'] = e[:event_id]
event_json['cat'] = "pocl"
event_json['name'] = e[:type]
event_json['pid'] = e[:dev_id]
event_json['tid'] = e[:queue_id]
event_json['args'] = e[:meta]
ts = stat[2]
event_json['ph'] = 'B'
event_json['ts'] = ts / 1000
out_f.puts ',' if first_done
first_done = true
out_f.puts event_json.to_json
#ts = stat[1]
#event_json['ph'] = 'I'
#event_json['ts'] = ts / 1000
#out_f.puts ',' if first_done
#out_f.puts event_json.to_json
#ts = stat[2]
#event_json['ph'] = 'I'
#event_json['ts'] = ts / 1000
#out_f.puts ',' if first_done
#out_f.puts event_json.to_json
ts = stat[3]
event_json['ph'] = 'E'
event_json['ts'] = ts / 1000
out_f.puts ',' if first_done
out_f.puts event_json.to_json
end
=begin
dep_count = DEPS_FROM.size
(0...dep_count).each do |i|
src_id = DEPS_FROM[i]
dst_id = DEPS_TO[i]
# puts "SRC DST ID: "
# puts src_id.inspect
# puts dst_id.inspect
# event_json['pid'] = e[:dev_id]
# event_json['tid'] = e[:queue_id]
event_json = Hash.new
event_json['id'] = src_id
event_json['cat'] = "pocl"
event_json['ph'] = 's'
event_json['pid'] = EVENTS[src_id][:dev_id]
event_json['tid'] = EVENTS[src_id][:queue_id]
# ts = EVENTS[src_id][:stat][3]
# event_json['ts'] = ts / 1000
out_f.puts ','
out_f.puts event_json.to_json
event_json = Hash.new
event_json['id'] = dst_id
event_json['cat'] = "pocl"
event_json['ph'] = 'f'
event_json['pid'] = EVENTS[dst_id][:dev_id]
event_json['tid'] = EVENTS[dst_id][:queue_id]
# ts = EVENTS[dst_id][:stat][0]
# event_json['ts'] = ts / 1000
out_f.puts ','
out_f.puts event_json.to_json
end
=end
###############################################
out_f.puts ']'
out_f.close
############################################
=begin
host_que = stat[1] - stat[0]
dev_que = stat[2] - stat[1]
on_dev = stat[3] - stat[2]
stat = Hash.new
stat[:start] = stat[0]
stat[:host_que] = host_que
stat[:dev_que] = dev_que
stat[:on_dev] = on_dev
e[:stat] = stat
# puts e.inspect
# puts "EVENT #{event_id}: HQ #{host_que} DQ #{dev_que} D #{on_dev}"
=end
|