From e076f854c9f7232044c2346dd32a04bdf6d7bacd Mon Sep 17 00:00:00 2001
From: Andreas Kloeckner <inform@tiker.net>
Date: Thu, 30 May 2013 17:34:24 -0400
Subject: [PATCH] PEP8-ify loopy.compiled.
---
loopy/compiled.py | 139 +++++++++++++++++++++++++---------------------
1 file changed, 76 insertions(+), 63 deletions(-)
diff --git a/loopy/compiled.py b/loopy/compiled.py
index c69eea427..8e5f82185 100644
--- a/loopy/compiled.py
+++ b/loopy/compiled.py
@@ -23,8 +23,6 @@ THE SOFTWARE.
"""
-
-
import pyopencl as cl
import pyopencl.array as cl_array
@@ -35,8 +33,6 @@ from pytools import Record, memoize_method
AUTO_TEST_SKIP_RUN = False
-
-
# {{{ argument checking
def _arg_matches_spec(arg, val, other_args):
@@ -69,6 +65,7 @@ def _arg_matches_spec(arg, val, other_args):
# }}}
+
# {{{ compiled kernel object
def _get_kernel_from_iterable(iterable):
@@ -89,16 +86,18 @@ def _get_kernel_from_iterable(iterable):
return result
+
class _KernelInfo(Record):
pass
+
class CompiledKernel:
def __init__(self, context, kernel, options=[], codegen_kwargs={}):
"""
:arg kernel: may be a loopy.LoopKernel, a generator returning kernels
- (a warning will be issued if more than one is returned). If the kernel
- has not yet been loop-scheduled, that is done, too, with no specific
- arguments.
+ (a warning will be issued if more than one is returned). If the
+ kernel has not yet been loop-scheduled, that is done, too, with no
+ specific arguments.
"""
import loopy as lp
@@ -125,9 +124,7 @@ class CompiledKernel:
kernel = self.kernel
import loopy as lp
- from loopy.kernel.tools import (
- add_argument_dtypes,
- get_arguments_with_incomplete_dtype)
+ from loopy.kernel.tools import add_argument_dtypes
if arg_to_dtype_set:
kernel = add_argument_dtypes(kernel, dict(arg_to_dtype_set))
@@ -146,7 +143,9 @@ class CompiledKernel:
if arg_to_has_offset[arg.name]:
offset_arg_name = vng(arg.name+"_offset")
new_args.append(arg.copy(offset=offset_arg_name))
- new_args.append(lp.ValueArg(offset_arg_name, kernel.index_dtype))
+ new_args.append(
+ lp.ValueArg(
+ offset_arg_name, kernel.index_dtype))
else:
new_args.append(arg.copy(offset=0))
else:
@@ -163,21 +162,25 @@ class CompiledKernel:
gsize_expr, lsize_expr = kernel.get_grid_sizes_as_exprs()
- if not gsize_expr: gsize_expr = (1,)
- if not lsize_expr: lsize_expr = (1,)
+ if not gsize_expr:
+ gsize_expr = (1,)
+ if not lsize_expr:
+ lsize_expr = (1,)
# }}}
from pymbolic import compile
return _KernelInfo(
kernel=kernel,
- global_size_func = compile(gsize_expr, kernel.scalar_loop_args),
+ global_size_func=compile(gsize_expr, kernel.scalar_loop_args),
local_size_func=compile(lsize_expr, kernel.scalar_loop_args),
)
@memoize_method
- def get_cl_kernel(self, arg_to_dtype_set, arg_to_has_offset_set, code_op=False):
- kernel_info = self.get_kernel_info(arg_to_dtype_set, arg_to_has_offset_set)
+ def get_cl_kernel(self,
+ arg_to_dtype_set, arg_to_has_offset_set, code_op=False):
+ kernel_info = self.get_kernel_info(
+ arg_to_dtype_set, arg_to_has_offset_set)
kernel = kernel_info.kernel
from loopy.codegen import generate_code
@@ -199,13 +202,13 @@ class CompiledKernel:
except KeyboardInterrupt:
raise
except:
- print "[Loopy] ----------------------------------------------------"
+ print "[Loopy] "+70*"-"
print "[Loopy] build failed, here's the source code:"
- print "[Loopy] ----------------------------------------------------"
+ print "[Loopy] "+70*"-"
print code
- print "[Loopy] ----------------------------------------------------"
+ print "[Loopy] "+70*"-"
print "[Loopy] end source code"
- print "[Loopy] ----------------------------------------------------"
+ print "[Loopy] "+70*"-"
raise
from loopy.kernel.data import ValueArg
@@ -235,7 +238,8 @@ class CompiledKernel:
return generate_code(kernel_info.kernel, **self.codegen_kwargs)
def get_highlighted_code(self, arg_to_dtype=None, arg_to_has_offset=None):
- return get_highlighted_code(self.get_code(arg_to_dtype, arg_to_has_offset))
+ return get_highlighted_code(
+ self.get_code(arg_to_dtype, arg_to_has_offset))
@property
def code(self):
@@ -248,8 +252,8 @@ class CompiledKernel:
# }}}
def __call__(self, queue, **kwargs):
- """If all array arguments are :mod:`numpy` arrays, defaults to returning
- numpy arrays as well.
+ """If all array arguments are :mod:`numpy` arrays, defaults to
+ returning numpy arrays as well.
If you want offset arguments (see
:attr:`loopy.kernel.data.GlobalArg.offset`) to be set automatically, it
@@ -307,6 +311,7 @@ class CompiledKernel:
args = []
outputs = []
encountered_numpy = False
+ encountered_cl = False
kwargs_copy = kwargs.copy()
@@ -323,7 +328,7 @@ class CompiledKernel:
# be set automatically, it must occur *after* the
# corresponding array argument.
- ofs, remdr = divmod(val.offset, val.dtype.itemsize)
+ ofs, remdr = divmod(val.offset, val.dtype.itemsize)
assert remdr == 0
kwargs_copy.setdefault(arg.offset, ofs)
del ofs
@@ -340,12 +345,15 @@ class CompiledKernel:
warn("argument '%s' was passed as a numpy array, "
"performing implicit transfer" % arg.name,
stacklevel=2)
+ else:
+ encountered_cl = True
# }}}
if val is None:
if not is_written:
- raise TypeError("must supply input argument '%s'" % arg.name)
+ raise TypeError(
+ "must supply input argument '%s'" % arg.name)
if isinstance(arg, lp.ImageArg):
raise RuntimeError("write-mode image '%s' must "
@@ -362,12 +370,14 @@ class CompiledKernel:
+ arg.dtype.itemsize)
if allocator is None:
- storage = cl.Buffer(queue.context, cl.mem_flags.READ_WRITE, alloc_size)
+ storage = cl.Buffer(
+ queue.context, cl.mem_flags.READ_WRITE, alloc_size)
else:
storage = allocator(alloc_size)
val = cl_array.Array(queue, shape, arg.dtype,
- strides=numpy_strides, data=storage, allocator=allocator)
+ strides=numpy_strides, data=storage,
+ allocator=allocator)
else:
assert _arg_matches_spec(arg, val, kwargs)
@@ -391,7 +401,7 @@ class CompiledKernel:
*args,
g_times_l=True, wait_for=wait_for)
- if out_host is None and encountered_numpy:
+ if out_host is None and (encountered_numpy and not encountered_cl):
out_host = True
if out_host:
outputs = [o.get(queue=queue) for o in outputs]
@@ -401,8 +411,6 @@ class CompiledKernel:
# }}}
-
-
def get_highlighted_code(text):
try:
from pygments import highlight
@@ -415,8 +423,6 @@ def get_highlighted_code(text):
return highlight(text, CLexer(), TerminalFormatter())
-
-
# {{{ automatic testing
def fill_rand(ary):
@@ -430,12 +436,10 @@ def fill_rand(ary):
fill_rand(ary, luxury=0)
-
-
-
class TestArgInfo(Record):
pass
+
def make_ref_args(kernel, queue, parameters, fill_value):
import loopy as lp
from loopy.kernel.data import ValueArg, GlobalArg, ImageArg
@@ -472,7 +476,8 @@ def make_ref_args(kernel, queue, parameters, fill_value):
is_image = isinstance(arg, ImageArg)
if is_image:
- storage_array = ary = cl_array.empty(queue, shape, arg.dtype, order="C")
+ storage_array = ary = cl_array.empty(
+ queue, shape, arg.dtype, order="C")
numpy_strides = None
alloc_size = None
strides = None
@@ -489,7 +494,8 @@ def make_ref_args(kernel, queue, parameters, fill_value):
dtype = arg.dtype
if dtype is None:
raise RuntimeError("dtype for argument '%s' is not yet "
- "known. Perhaps you want to use loopy.add_argument_dtypes "
+ "known. Perhaps you want to use "
+ "loopy.add_argument_dtypes "
"or loopy.infer_argument_dtypes?"
% arg.name)
@@ -515,7 +521,8 @@ def make_ref_args(kernel, queue, parameters, fill_value):
fill_rand(storage_array)
if isinstance(arg, ImageArg):
# must be contiguous
- ref_args[arg.name] = cl.image_from_array(queue.context, ary.get())
+ ref_args[arg.name] = cl.image_from_array(
+ queue.context, ary.get())
else:
ref_args[arg.name] = ary
@@ -535,8 +542,6 @@ def make_ref_args(kernel, queue, parameters, fill_value):
return ref_args, arg_descriptors
-
-
def make_args(queue, kernel, arg_descriptors, parameters,
fill_value):
import loopy as lp
@@ -610,12 +615,16 @@ def make_args(queue, kernel, arg_descriptors, parameters,
# create host array with test shape (but not strides)
host_contig_array = np.empty(shape, dtype=arg.dtype)
- common_len = min(len(host_ref_flat_array), len(host_contig_array.ravel()))
- host_contig_array.ravel()[:common_len] = host_ref_flat_array[:common_len]
+ common_len = min(
+ len(host_ref_flat_array),
+ len(host_contig_array.ravel()))
+ host_contig_array.ravel()[:common_len] = \
+ host_ref_flat_array[:common_len]
# create host array with test shape and storage layout
host_storage_array = np.empty(alloc_size, arg.dtype)
- host_array = as_strided(host_storage_array, shape, numpy_strides)
+ host_array = as_strided(
+ host_storage_array, shape, numpy_strides)
host_array[:] = host_contig_array
host_contig_array = arg_desc.ref_storage_array.get()
@@ -637,12 +646,16 @@ def make_args(queue, kernel, arg_descriptors, parameters,
return args
-
-
def _default_check_result(result, ref_result):
if not np.allclose(ref_result, result, rtol=1e-3, atol=1e-3):
- l2_err = np.sum(np.abs(ref_result-result)**2)/np.sum(np.abs(ref_result)**2)
- linf_err = np.max(np.abs(ref_result-result))/np.max(np.abs(ref_result-result))
+ l2_err = (
+ np.sum(np.abs(ref_result-result)**2)
+ /
+ np.sum(np.abs(ref_result)**2))
+ linf_err = (
+ np.max(np.abs(ref_result-result))
+ /
+ np.max(np.abs(ref_result-result)))
return (False,
"results do not match(rel) l_2 err: %g, l_inf err: %g"
% (l2_err, linf_err))
@@ -650,8 +663,6 @@ def _default_check_result(result, ref_result):
return True, None
-
-
def _enumerate_cl_devices_for_ref_test():
noncpu_devs = []
cpu_devs = []
@@ -676,9 +687,9 @@ def _enumerate_cl_devices_for_ref_test():
raise RuntimeError("no CL device found for test")
if not cpu_devs:
- warn("No CPU device found for reference test. The reference computation "
- "will either fail because of a timeout or take a *very* long "
- "time.")
+ warn("No CPU device found for reference test. The reference "
+ "computation will either fail because of a timeout "
+ "or take a *very* long time.")
for dev in cpu_devs:
yield dev
@@ -687,9 +698,8 @@ def _enumerate_cl_devices_for_ref_test():
yield dev
-
-
-def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], parameters={},
+def auto_test_vs_ref(
+ ref_knl, ctx, kernel_gen, op_count=[], op_label=[], parameters={},
print_ref_code=False, print_code=True, warmup_rounds=2,
code_op=None, dump_binary=False, codegen_kwargs={},
options=[],
@@ -699,8 +709,8 @@ def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], paramet
`kernel_gen`.
:arg check_result: a callable with :class:`numpy.ndarray` arguments
- *(result, reference_result)* returning a a tuple (class:`bool`, message)
- indicating correctness/acceptability of the result
+ *(result, reference_result)* returning a a tuple (class:`bool`,
+ message) indicating correctness/acceptability of the result
"""
if isinstance(op_count, (int, float)):
@@ -786,7 +796,6 @@ def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], paramet
print get_highlighted_code(ref_compiled.code)
print 75*"-"
-
ref_queue.finish()
ref_start = time()
@@ -807,7 +816,8 @@ def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], paramet
break
if not found_ref_device:
- raise RuntimeError("could not find a suitable device for the reference computation.\n"
+ raise RuntimeError("could not find a suitable device for the "
+ "reference computation.\n"
"These errors were encountered:\n"+"\n".join(ref_errors))
# }}}
@@ -896,11 +906,13 @@ def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], paramet
evt_start.wait()
evt_end.wait()
- elapsed = (1e-9*events[-1].profile.END-1e-9*events[0].profile.SUBMIT) \
+ elapsed = (1e-9*events[-1].profile.END
+ - 1e-9*events[0].profile.SUBMIT) \
/ timing_rounds
try:
elapsed_evt_2 = "%g" % \
- ((1e-9*evt_end.profile.START-1e-9*evt_start.profile.START) \
+ ((1e-9*evt_end.profile.START
+ - 1e-9*evt_start.profile.START)
/ timing_rounds)
except cl.RuntimeError:
elapsed_evt_2 = "<unavailable>"
@@ -916,8 +928,9 @@ def auto_test_vs_ref(ref_knl, ctx, kernel_gen, op_count=[], op_label=[], paramet
for cnt, lbl in zip(op_count, op_label):
rates += " %g %s/s" % (cnt/elapsed_wall, lbl)
- print "elapsed: %g s event, %s s marker-event %g s wall (%d rounds)%s" % (
- elapsed, elapsed_evt_2, elapsed_wall, timing_rounds, rates)
+ print("elapsed: %g s event, %s s marker-event %g s wall "
+ "(%d rounds)%s" % (
+ elapsed, elapsed_evt_2, elapsed_wall, timing_rounds, rates))
if do_check:
ref_rates = ""
--
GitLab