mirror-github-bspeice/cuburn
1
0
Fork 0
mirror of https://github.com/stevenrobertson/cuburn.git synced 2025-02-05 11:40:04 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fix occupancy issues (1 block/SM when shuffle was on).

Browse Source 
There are 16 bar.sync() registers available per *chip*, not per block, and I
was using number 8 in the shuffle code. Evidently the driver rewrites them per
SM, but does not compact their range. Good to know.
This commit is contained in:
Steven Robertson 2010-09-12 11:09:47 -04:00
parent ce0571deaf
commit a439bf671d
2 changed files with 8 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
cuburn/device_code.py
View File

@ -36,20 +36,18 @@ class IterThread(PTXEntryPoint):
@ptx_func @ptx_func
def entry(self): def entry(self):
# For now, we indulge in the luxury of shared memory.
# Index number of current CP, shared across CTA # Index number of current CP, shared across CTA
mem.shared.u32('s_cp_idx') mem.shared.u32('s_cp_idx')
# Number of samples that have been generated so far in this CTA # Number of samples that have been generated so far in this CTA
# If this number is negative, we're still fusing points, so this # If this number is negative, we're still fusing points, so this
# behaves slightly differently (see ``fuse_loop_start``) # behaves slightly differently (see ``fuse_loop_start``)
# TODO: replace (or at least simplify) this logic
mem.shared.s32('s_num_samples') mem.shared.s32('s_num_samples')
op.st.shared.s32(addr(s_num_samples), -(features.num_fuse_samples+1)) op.st.shared.s32(addr(s_num_samples), -(features.num_fuse_samples+1))
mem.shared.f32('s_xf_sel', ctx.warps_per_cta) mem.shared.f32('s_xf_sel', ctx.warps_per_cta)
#std.store_per_thread(g_whatever, 1234)
# TODO: temporary, for testing # TODO: temporary, for testing
mem.local.u32('l_num_rounds') mem.local.u32('l_num_rounds')
mem.local.u32('l_num_writes') mem.local.u32('l_num_writes')
@ -264,8 +262,10 @@ class IterThread(PTXEntryPoint):
print '%s:' % s print '%s:' % s
for i, r in enumerate(a): for i, r in enumerate(a):
for j in range(0,len(r),w): for j in range(0,len(r),w):
print '%2d\t%s' % (i, print '%2d' % i,
'\t'.join(['%g '%np.mean(r[k]) for k in range(j,j+w)])) for k in range(j,j+w,8):
print '\t' + ' '.join(
['%8g'%np.mean(r[l]) for l in range(k,k+8)])
num_rounds_dp, num_rounds_l = ctx.mod.get_global('g_num_rounds') num_rounds_dp, num_rounds_l = ctx.mod.get_global('g_num_rounds')
num_writes_dp, num_writes_l = ctx.mod.get_global('g_num_writes') num_writes_dp, num_writes_l = ctx.mod.get_global('g_num_writes')
@ -536,9 +536,9 @@ class ShufflePoints(PTXFragment):
def shuffle(self, *args, **kwargs): def shuffle(self, *args, **kwargs):
""" """
Shuffle the data from each register in args across threads. Keyword Shuffle the data from each register in args across threads. Keyword
argument ``bar`` specifies which barrier to use. argument ``bar`` specifies which barrier to use (default is 2).
""" """
bar = kwargs.pop('bar', 8) bar = kwargs.pop('bar', 2)
with block("Shuffle across threads"): with block("Shuffle across threads"):
reg.u32('shuf_read shuf_write') reg.u32('shuf_read shuf_write')
with block("Calculate read and write offsets"): with block("Calculate read and write offsets"):

3
cuburn/ptx.py
View File

@ -458,7 +458,6 @@ class Mem(object):
class _MemFactory(_CallChain): class _MemFactory(_CallChain):
"""Actual `mem` object""" """Actual `mem` object"""
def _call(self, type, name, array=False, init=None): def _call(self, type, name, array=False, init=None):
assert len(type) == 2
memobj = Mem(type, name, array, init) memobj = Mem(type, name, array, init)
if array is True: if array is True:
array = ['[]'] array = ['[]']
@ -468,7 +467,7 @@ class _MemFactory(_CallChain):
array = [] array = []
if init: if init:
array += [' = ', init] array += [' = ', init]
self.block.code(op='.%s.%s ' % (type[0], type[1]), vars=[name, array]) self.block.code(op='.%s ' % ' .'.join(type), vars=[name, array])
self.block.inject(name, memobj) self.block.inject(name, memobj)
# TODO: move vec, addr here, or make this public # TODO: move vec, addr here, or make this public