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
556 Commits 4 Branches 0 Tags 2.3 MiB
31d00c8503
Commit Graph

556 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Aryan Blaauw
31d00c8503
Update Dockerfile 2019-02-18 19:06:56 +01:00
Aryan Blaauw
ac006f57c1
+ant 2019-02-18 16:45:36 +01:00
Aryan Blaauw
83297a2a03
18.04 build 2019-02-18 15:29:01 +01:00
Aryan Blaauw
5a8522686b
branched & cleanup 2019-01-29 09:32:52 +01:00
Aryan Blaauw
c19f9fa481
upgrade to 18.04 with abstraction upgrade to deploy-nvidia-docker as a base image 2019-01-29 09:25:15 +01:00
Aryan Blaauw
08f887cb1b
Update Dockerfile 2017-12-15 13:58:41 +01:00
Aryan Blaauw
7a6ee2cff3
Update Dockerfile 2017-12-11 20:36:29 +01:00
Aryan Blaauw
ff350b4111
Update Dockerfile 2017-12-05 14:27:30 +01:00
Aryan Blaauw
a5de46ae44
Update Dockerfile 2017-12-05 14:24:53 +01:00
Aryan Blaauw
26d469e156
Update Dockerfile 2017-12-05 13:41:18 +01:00
Aryan Blaauw
fb1f56fc2f
Update Dockerfile 2017-12-05 13:27:32 +01:00
Aryan Blaauw
4037218e3b
Update Dockerfile 2017-12-05 13:08:04 +01:00
Aryan Blaauw
ba3593b96a
Update Dockerfile 2017-12-05 12:34:26 +01:00
Aryan Blaauw
d4f542d800
Update Dockerfile 2017-12-05 12:17:33 +01:00
Aryan Blaauw
418e747dfb
Update Dockerfile 2017-12-05 11:03:41 +01:00
Aryan Blaauw
6b00a46ca6
Update Dockerfile 2017-12-05 10:05:49 +01:00
Aryan Blaauw
9b2a21342d
Update Dockerfile 2017-12-05 09:44:58 +01:00
Aryan Blaauw
61f106083f Update Dockerfile 2017-09-04 20:32:13 +02:00
Aryan Blaauw
c859b3ee28 Update Dockerfile 2017-09-04 20:31:34 +02:00
Aryan Blaauw
37082bb8d5 Update Dockerfile 2017-08-28 11:21:13 +02:00
Aryan Blaauw
5c076e7720 Update Dockerfile 2017-08-28 10:35:55 +02:00
Aryan Blaauw
9131a7793e Update Dockerfile 2017-08-28 10:29:07 +02:00
Aryan Blaauw
b624572641 Rename Dockerfile.light to Dockerfile 2017-08-28 10:27:22 +02:00
Aryan Blaauw
72cf04fc2a Rename Dockerfile to Dockerfile.heavy 2017-08-28 10:26:59 +02:00
Aryan Blaauw
c39a2b5c23 Create Dockerfile.light 2017-08-28 10:26:33 +02:00
Aryan Blaauw
1797f1ce3f Update Dockerfile 2017-08-28 10:21:12 +02:00
Aryan Blaauw
8a6b41f9be Update Dockerfile 2017-08-27 22:00:29 +02:00
Aryan Blaauw
e230dee5b0 Update Dockerfile 2017-08-25 10:22:42 +02:00
Aryan Blaauw
e5e1cf426f Update Dockerfile 2017-08-25 09:50:38 +02:00
Aryan Blaauw
8e18d4dd99 Update Dockerfile 2017-08-24 20:09:46 +02:00
Aryan Blaauw
cecd905ca1 Update README.md 2017-08-24 20:08:19 +02:00
Aryan Blaauw
e8142d21d5 Update README.md 2017-08-24 20:07:18 +02:00
Aryan Blaauw
5ff25fe745 Update README.md 2017-08-24 20:06:53 +02:00
Aryan Blaauw
e961ebe109 Create Dockerfile 2017-08-24 20:05:58 +02:00
Steven Robertson
93d1432f02 Cap retry delay at 10 minutes 2017-05-15 18:40:47 -07:00
Steven Robertson
a29d17b0bd Actually write the output :( 2017-05-15 18:22:13 -07:00
Steven Robertson
112a674520 Redesign distribution: now based on ssh, not zmq 2017-05-15 12:04:16 -07:00
Steven Robertson
c7654357a6 Move naming code into a common place 2017-05-15 12:01:59 -07:00
Steven Robertson
04702d7903 Add --list-devices option 2017-05-15 12:01:25 -07:00
Steven Robertson
29c595ddc5 Move most warning/info statements to stderr 2017-05-15 12:00:11 -07:00
Steven Robertson
9bcfc36b7a Retrieve out suffix without creating a renderer 2017-05-15 11:56:37 -07:00
Steven Robertson
636efcd059 Drop GL mode in main.py; sleep to reduce load 2017-05-15 00:44:15 -07:00
Steven Robertson
7dc58a0e1c Grow launch sizes and synchronize if they pile up 2017-05-15 00:43:10 -07:00
Steven Robertson
5402838a74 Disable ill-thought-through form of antialiasing 2017-05-15 00:41:30 -07:00
Steven Robertson
3528cd1da4 Force use of clang for compilation for Debian 2017-05-15 00:38:52 -07:00
Steven Robertson
f58289af53 Hotspot writeback. 10x performance increase. 
Create a map assigning two bits to every output bin. During the atomic
flush, compute a threshold for discarding writes altogether that would
keep us under 2% error - discard 1 of every 2 writes if we've already
accumulated 64 writes (hotspot value 1), 7 of 8 if we're above 256
(hotspot value 2), or 31 of 32 at 2048 (hotspot value 3). Pack this
value into a read-only buffer that can often be cached at L2, and for
particularly concentrated flames (which historically choke cuburn), L1.
During writeback, discard writes at the apporpriate rate. During the
flush of the integer accumulator to the float, scale the integer
accumulators by the discard rate.

This works because for most flames, there's not a lot of interesting
stuff in the middle regimes; either stuff is very well defined, in which
case we pretty much know exactly what the color is going to be
(remember, the max 2% relative error gets log-scaled as well), or it's
loosely defined so we should keep it at full accuracy.

Of course, a 10x boost is best-case-ish - a long, high-res render. I
realized though that I really didn't care about low quality stuff and
should go for broke optimizing this for my use case, which is
ridiculously high res HDR stuff. (On pathological flames, on the other
hand, 10x is conservative; this easily gives us 100x.)
 2017-05-09 21:16:43 -07:00
Steven Robertson
0bcde947b5 Go to 1024 contexts on Pascal 2017-05-09 21:15:03 -07:00
Steven Robertson
d1502e3b79 rings2 is not identity at high precision 2017-05-09 21:09:58 -07:00
Steven Robertson
d759d675be Always flush status lines 2017-05-09 21:09:40 -07:00
Steven Robertson
5af90b01a2 Fix a silly 'except e' (too much yavascrip in my life) 2017-05-09 21:09:00 -07:00