mirror-github-bspeice/cuburn
1
0
Fork 0
mirror of https://github.com/stevenrobertson/cuburn.git synced 2025-07-12 03:05:14 -04:00
Code Issues Actions Packages Projects Releases Wiki Activity

Add faster no-L1 accum

Browse Source
This commit is contained in:
Steven Robertson
2011-10-15 00:32:30 -04:00
parent dd645bcbf6
commit c7728d3507
3 changed files with 65 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
cuburn/render.py
View File

@ -320,7 +320,7 @@ class _AnimRenderer(object):
d_seeds = self.d_seeds if on_main else self.d_alt_seeds
if not d_seeds:
seeds = mwc.MWC.make_seeds(iter.IterCode.NTHREADS *
seeds = mwc.MWC.make_seeds(a._iter.NTHREADS *
self.cps_per_block)
if self.sync:
d_seeds = cuda.to_device(seeds)
@ -363,9 +363,9 @@ class _AnimRenderer(object):
h_infos[:] = infos
cuda.memcpy_htod_async(d_info_off, h_infos, stream)
# TODO: get block config from IterCode
iter_fun(d_seeds, np.uintp(d_info_off), self.d_accum,
block=(32, 16, 1), grid=(len(block_times), 1),
iter_fun(d_seeds, np.uintp(d_info_off), np.uint64(self.d_accum),
block=(32, a._iter.NTHREADS/32, 1),
grid=(len(block_times), 1),
texrefs=[tref], stream=stream)
if self.sync and self.sleep: