def profile_adp(args):
board = runtime_util.get_device(args.model_number)
calib_obj = llenums.CalibrateObjective(args.method)
runtime = GrendelRunner()
runtime.initialize()
if args.missing:
for exp_delta_model in delta_model_lib.get_all(board):
profile_kernel(runtime,board, \
exp_delta_model.block, \
exp_delta_model.config, \
calib_obj, \
min_points=args.min_points, \
grid_size=args.grid_size)
else:
adp = runtime_util.get_adp(board, args.adp, widen=args.widen)
for cfg in adp.configs:
blk = board.get_block(cfg.inst.block)
cfg_modes = cfg.modes
for mode in cfg_modes:
cfg.modes = [mode]
profile_kernel(runtime,board,blk,cfg,calib_obj, \
args.min_points, args.max_points, \
args.grid_size, \
force=args.force, adp=adp)
def make_histogram(args):
board = runtime_util.get_device(args.model_number,layout=True)
ph = paths.DeviceStatePathHandler(board.name, \
board.model_number,make_dirs=True)
models = {}
for exp_model in delta_model_lib.get_all(board):
key = (exp_model.block, exp_model.output,\
runtime_util.get_static_cfg(exp_model.block,exp_model.config), \
exp_model.calib_obj)
if not key in models:
models[key] = []
models[key].append(exp_model)
for key,models in models.items():
blk,out,cfg,lbl = key
print("%s:%s = %d models" % (blk.name,cfg,len(models)))
png_file = ph.get_histogram_vis('merr', \
blk.name, \
out.name, \
cfg, \
lbl)
vizlib.model_error_histogram(models, \
png_file, \
num_bins=10, \
relative=True)
png_file = ph.get_histogram_vis('objf', \
blk.name, \
out.name, \
cfg, \
lbl)
vizlib.objective_fun_histogram(models, \
png_file, \
num_bins=10)
def preprocess_board_data(board):
def insert(d,ks):
for k in ks:
if not k in d:
d[k] = {}
d = d[k]
datasets = {}
models = {}
for model in exp_delta_model_lib.get_all(board):
keypath = [(model.block.name, str(model.config.mode),model.output.name), \
model.loc]
insert(models,keypath)
insert(datasets,keypath)
models[keypath[0]][keypath[1]][model.hidden_cfg] = model
datasets[keypath[0]][keypath[1]][model.hidden_cfg] = None
for data in exp_profile_dataset_lib.get_all(board):
keypath = [(data.block.name, str(data.config.mode),data.output.name), \
data.loc]
insert(datasets,keypath)
datasets[keypath[0]][keypath[1]][data.hidden_cfg] = data
return models,datasets
def update_predictor(predictor,char_board):
for model in exp_delta_model_lib.get_all(char_board):
for var,val in model.variables().items():
predictor.data.add_datapoint(model.output, var, \
dict(model.hidden_codes()), val)
predictor.fit()
def group_by_configured_block(board):
instances = {}
for model in exp_delta_model_lib.get_all(board):
key = (model.block.name, \
str(model.loc), \
model.static_cfg)
if not key in instances:
instances[key] = []
instances[key].append(model)
return instances.values()
def load_code_pool_from_database(char_board,predictor,objectives):
hidden_codes = list(map(lambda st: st.name, \
runtime_util.get_hidden_codes(predictor.block)))
code_pool= poollib.HiddenCodePool(hidden_codes,predictor,objectives)
# setting ranges of values in code pool
for hc in hidden_codes:
vals = predictor.block.state[hc].values
code_pool.set_range(hc,vals)
by_output = {}
print("===== Database Models ====")
for mdl in exp_delta_model_lib.get_all(char_board):
print(mdl)
insert_into_dict(by_output, [mdl.hidden_cfg], mdl)
print("===== Initial Code Pool ====")
for _,mdls in by_output.items():
if any(map(lambda mdl: not mdl.complete, mdls)):
continue
codes = dict(mdls[0].hidden_codes())
pred_deltavars,_ = code_pool.predictor.predict(codes)
pred_obj = code_pool.objectives.compute(pred_deltavars)
vs = {}
for mdl in mdls:
vs[mdl.output.name] = mdl.variables()
actual_obj = code_pool.objectives.compute(vs)
print("hist %s" % (codes))
for (_,expr),pred,act in zip(code_pool.objectives,pred_obj,actual_obj):
print(" obj=%s pred=%f meas=%f" % (expr,pred,act))
if not code_pool.has_code(codes):
code_pool .add_labeled_code(codes,vs)
print("==== pool (best to worst) ====")
print(predictor.config)
print("")
for fxn,score in code_pool.meas_view.order_by_dominance():
print("%s result=%s" % (fxn,score))
return code_pool
def visualize(args):
board = runtime_util.get_device(args.model_number,layout=True)
calib_obj = llenums.CalibrateObjective(args.calib_obj) if args.calib_obj else None
ph = paths.DeviceStatePathHandler(board.name, \
board.model_number,make_dirs=True)
if args.histogram:
make_histogram(args)
for delta_model in delta_model_lib.get_all(board):
if not calib_obj is None and delta_model.calib_obj != calib_obj:
continue
if delta_model.complete and \
delta_model.calib_obj != llenums.CalibrateObjective.NONE:
print(delta_model.config)
print(delta_model)
if delta_model.is_integration_op:
dataset = dataset_lib.load(board, delta_model.block, \
delta_model.loc, \
delta_model.output, \
delta_model.config, \
llenums.ProfileOpType.INTEG_INITIAL_COND)
else:
dataset = dataset_lib.load(board, delta_model.block, \
delta_model.loc, \
delta_model.output, \
delta_model.config, \
llenums.ProfileOpType.INPUT_OUTPUT)
if dataset is None:
print("-> no dataset")
continue
assert(isinstance(dataset, dataset_lib.ExpProfileDataset))
png_file = ph.get_delta_vis(delta_model.block.name, \
delta_model.loc.file_string(),\
str(delta_model.output.name), \
str(delta_model.static_cfg), \
str(delta_model.hidden_cfg), \
delta_model.calib_obj)
print(png_file)
vizlib.deviation(delta_model, \
dataset, \
png_file, \
baseline=vizlib.ReferenceType.MODEL_PREDICTION, \
num_bins=10, \
amplitude=args.max_error, \
relative=not args.absolute)
png_file = ph.get_correctable_delta_vis(delta_model.block.name, \
delta_model.loc.file_string(), \
str(delta_model.output.name), \
str(delta_model.static_cfg), \
str(delta_model.hidden_cfg), \
delta_model.calib_obj)
print(png_file)
vizlib.deviation(delta_model, \
dataset, \
png_file, \
baseline=vizlib.ReferenceType.CORRECTABLE_MODEL_PREDICTION, \
num_bins=10, \
amplitude=args.max_error, \
relative=not args.absolute)
model_file = ph.get_model_file(delta_model.block.name, \
delta_model.loc.file_string(), \
str(delta_model.output.name), \
str(delta_model.static_cfg), \
str(delta_model.hidden_cfg), \
delta_model.calib_obj)
with open(model_file,'w') as fh:
fh.write(str(delta_model.config))
fh.write(str(delta_model))
fh.write("\n\n")
fh.write(str(delta_model.spec))
def homogenous_database_get_block_info(board, use_output=True):
models = list(exp_delta_model_lib.get_all(board))
return models_get_block_info(models, use_output)
def homogenous_database_get_calibration_objective_scores(board):
all_models = list(exp_delta_model_lib.get_all(board))
for models, score in get_calibration_objective_scores(all_models):
yield models, score
def database_is_homogenous(board, enable_outputs=False):
models = list(exp_delta_model_lib.get_all(board))
return models_are_homogenous(models, \
enable_outputs=enable_outputs)
def database_is_empty(board):
for model in exp_delta_model_lib.get_all(board):
return False
return True
