gaspersic/triton
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Simple linear model

Browse Source
This commit is contained in:
Philippe Tillet
2014-09-28 19:37:56 -04:00
parent 693b8b67b0
commit 0eb56a10f0
4 changed files with 87 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
autotune/python/autotune.py
View File

@@ -12,6 +12,8 @@ from pyviennacl import backend
from pyviennacl import opencl
from pyviennacl import atidlas
from dataset import generate_dataset
from model import train_model
import tools
import utils
import vclio
@@ -66,12 +68,11 @@ def do_tuning(config_fname, spec_fname, viennacl_root):
sys.stderr.write('Warning : The device ' + device.name + ' does not support double precision! Skipping ...')
continue
#Helper
def execute(node, other_params, sizes, fname = os.devnull):
def execute(statement, other_params, sizes, fname = os.devnull):
print('-----')
print(' '.join(map(str, ("Now tuning:", datatype.__name__, '-', operation, '-'.join(other_params), '[' + device.name, '(' + device.platform.name + ')] for sizes', sizes))))
with open(fname, "w+") as archive:
with vcl.Statement(node) as statement:
return optimize.genetic(statement, ctx, TYPES[operation]['template'], lambda p: TYPES[operation]['template'](p, *other_params),
return optimize.genetic(statement, ctx, TYPES[operation]['template'], lambda p: TYPES[operation]['template'](p, *other_params),
TYPES[operation]['parameter-names'], lambda t: TYPES[operation]['perf-index']([datatype().itemsize, sizes, t]), TYPES[operation]['perf-measure'], archive)
s = map_to_list((int, p['size']))
#Vector AXPY
@@ -100,7 +101,7 @@ def do_tuning(config_fname, spec_fname, viennacl_root):
if 'all' in layouts:
layouts = ['NN', 'NT', 'TN', 'TT']
for layout in layouts:
def execution_handler(sizes, fname):
def execution_handler(sizes, fname, parameters=None):
A_trans = layout[0]
B_trans = layout[1]
A = vcl.Matrix((sizes[0], sizes[1]) if A_trans=='N' else (sizes[1],sizes[0]), context=ctx, dtype=datatype, layout=vcl.COL_MAJOR);
@@ -110,8 +111,15 @@ def do_tuning(config_fname, spec_fname, viennacl_root):
alpha = vcl.HostScalar(1.0, context=ctx, dtype = datatype)
beta = vcl.HostScalar(1.0, context=ctx, dtype = datatype)
C = vcl.Matrix((sizes[0], sizes[2]), context=ctx, dtype = datatype, layout=vcl.COL_MAJOR)
execute(vcl.Assign(C,LHS*RHS*alpha + C*beta),(A_trans, B_trans), sizes, fname)
generate_dataset(operation, execution_handler)
statement = vcl.Statement(vcl.Assign(C,LHS*RHS*alpha + C*beta))
if parameters:
TemplateType = TYPES[operation]['template']
return tools.benchmark(TemplateType(TemplateType.Parameters(*parameters),A_trans,B_trans), statement, device)
else:
execute(statement,(A_trans, B_trans), sizes, fname)
X, Y, profiles = generate_dataset(TYPES[operation]['template'], execution_handler)
train_model(X, Y, profiles)
if __name__ == "__main__":