From c671e9388f5cc6ecb6428021911ea9f4936fdbb8 Mon Sep 17 00:00:00 2001
From: Andreas Kloeckner <inform@tiker.net>
Date: Sun, 17 Apr 2011 21:34:18 -0400
Subject: [PATCH] Pull in pycuda scan changes.
---
pyopencl/scan.py | 41 +++++++++--------------------------------
1 file changed, 9 insertions(+), 32 deletions(-)
diff --git a/pyopencl/scan.py b/pyopencl/scan.py
index 02679a31..ee6f74fe 100644
--- a/pyopencl/scan.py
+++ b/pyopencl/scan.py
@@ -364,31 +364,6 @@ void ${name_prefix}_final_update(
-def _div_ceil(nr, dr):
- return (nr + dr -1) // dr
-
-
-def _uniform_interval_splitting(n, granularity, max_intervals):
- grains = _div_ceil(n, granularity)
-
- # one grain per interval
- if grains <= max_intervals:
- return granularity, grains
-
- # ensures that:
- # num_intervals * interval_size is >= n
- # and
- # (num_intervals - 1) * interval_size is < n
-
- grains_per_interval = _div_ceil(grains, max_intervals)
- interval_size = grains_per_interval * granularity
- num_intervals = _div_ceil(n, interval_size)
-
- return interval_size, num_intervals
-
-
-
-
if _CL_MODE:
class _ScanKernelBase(object):
def __init__(self, ctx, dtype,
@@ -475,7 +450,8 @@ if _CL_MODE:
unit_size = self.scan_wg_size * self.scan_wg_seq_batches
max_groups = 3*max(dev.max_compute_units for dev in self.devices)
- interval_size, num_groups = _uniform_interval_splitting(
+ from pytools import uniform_interval_splitting
+ interval_size, num_groups = uniform_interval_splitting(
n, unit_size, max_groups);
block_results = allocator(self.dtype.itemsize*num_groups)
@@ -541,7 +517,7 @@ else:
scan_intervals_src, options=options, no_extern_c=True)
self.scan_intervals_knl = scan_intervals_prg.get_function(
name_prefix+"_scan_intervals")
- self.scan_intervals_knl.prepare("PIIPP", (self.scan_wg_size, 1, 1))
+ self.scan_intervals_knl.prepare("PIIPP")
final_update_src = str(self.final_update_tp.render(
wg_size=self.update_wg_size,
@@ -551,7 +527,7 @@ else:
final_update_src, options=options, no_extern_c=True)
self.final_update_knl = final_update_prg.get_function(
name_prefix+"_final_update")
- self.final_update_knl.prepare("PIIP", (self.update_wg_size, 1, 1))
+ self.final_update_knl.prepare("PIIP")
def __call__(self, input_ary, output_ary=None, allocator=None,
stream=None):
@@ -576,7 +552,8 @@ else:
max_groups = 3*dev.get_attribute(
driver.device_attribute.MULTIPROCESSOR_COUNT)
- interval_size, num_groups = _uniform_interval_splitting(
+ from pytools import uniform_interval_splitting
+ interval_size, num_groups = uniform_interval_splitting(
n, unit_size, max_groups);
block_results = allocator(self.dtype.itemsize*num_groups)
@@ -584,7 +561,7 @@ else:
# first level scan of interval (one interval per block)
self.scan_intervals_knl.prepared_async_call(
- (num_groups, 1), stream,
+ (num_groups, 1), (self.scan_wg_size, 1, 1), stream,
input_ary.gpudata,
n, interval_size,
output_ary.gpudata,
@@ -592,7 +569,7 @@ else:
# second level inclusive scan of per-block results
self.scan_intervals_knl.prepared_async_call(
- (1, 1), stream,
+ (1,1), (self.scan_wg_size, 1, 1), stream,
block_results,
num_groups, interval_size,
block_results,
@@ -600,7 +577,7 @@ else:
# update intervals with result of second level scan
self.final_update_knl.prepared_async_call(
- (num_groups, 1,), stream,
+ (num_groups, 1,), (self.update_wg_size, 1, 1), stream,
output_ary.gpudata,
n, interval_size,
block_results)
--
GitLab