19 Jan
2015
19 Jan
'15
11:42 a.m.
Am 19.01.2015 um 02:52 schrieb Manuel AC:
It occasionally went a bit over 200%, pretty low.
But did you test the runtime/fps performance with Big Buck Bunny?
- Carsten
19 Jan
19 Jan
12:56 p.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Yes, the 200% is Big Buck Bunny. You got the log attached in the mail
for the benchmark.
The prores files are going over 1100%, perfect.
Manuel AC
On Mon, Jan 19, 2015 at 6:42 AM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
Am 19.01.2015 um 02:52 schrieb Manuel AC:
It occasionally went a bit over 200%, pretty low.
But did you test the runtime/fps performance with Big Buck Bunny?
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
1:17 p.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Am 19.01.2015 um 13:56 schrieb Manuel AC:
Yes, the 200% is Big Buck Bunny. You got the log
attached in the mail
for the benchmark.
Sorry, missed that log.
Very strange. Did you try the same with 1.69? Although I don't think it is something
in the code, as it actually went faster on my Mac during the recent updates.
Maybe the source disc has issues?
- Carsten
1:35 p.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
I will try with 1.69 and all logging off when the computer is available.
I used the same disk for a couple prores files using 1.69 and 1.74,
full speed, no problems.
On Mon, Jan 19, 2015 at 8:17 AM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
Am 19.01.2015 um 13:56 schrieb Manuel AC:
Yes, the 200% is Big Buck Bunny. You got the log
attached in the mail
for the benchmark.
Sorry, missed that log.
Very strange. Did you try the same with 1.69? Although I don't think it is something
in the code, as it actually went faster on my Mac during the recent updates.
Maybe the source disc has issues?
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic 
4:11 p.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
So up to now no source files are going at full speed with 1.76, but
they're ok with 1.74?
And you've tried 1.76 without logging?
Thanks,
Carl
On Mon, 19 Jan 2015, Manuel AC wrote:
I will try with 1.69 and all logging off when the
computer is available.
I used the same disk for a couple prores files using 1.69 and 1.74,
full speed, no problems.
On Mon, Jan 19, 2015 at 8:17 AM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
Am 19.01.2015 um 13:56 schrieb Manuel AC:
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
Yes, the 200% is Big Buck Bunny. You got the log
attached in the mail
for the benchmark.
Sorry, missed that log.
Very strange. Did you try the same with 1.69? Although I don't think it is something
in the code, as it actually went faster on my Mac during the recent updates.
Maybe the source disc has issues?
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
4:13 p.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Sorry, I meant 1.76.4x
1.76.42 goes full speed with prores files. Not the Bunny trailer or
the original problematic file that I thought it was prores, but looks
like not.
On Mon, Jan 19, 2015 at 11:11 AM, Carl Hetherington <cth(a)carlh.net> wrote:
So up to now no source files are going at full speed with 1.76, but
they're ok with 1.74?
And you've tried 1.76 without logging?
Thanks,
Carl
On Mon, 19 Jan 2015, Manuel AC wrote:
I will try with 1.69 and all logging off when the
computer is available.
I used the same disk for a couple prores files using 1.69 and 1.74,
full speed, no problems.
On Mon, Jan 19, 2015 at 8:17 AM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
Am 19.01.2015 um 13:56 schrieb Manuel AC:
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
Yes, the 200% is Big Buck Bunny. You got the log
attached in the mail
for the benchmark.
Sorry, missed that log.
Very strange. Did you try the same with 1.69? Although I don't think it is something
in the code, as it actually went faster on my Mac during the recent updates.
Maybe the source disc has issues?
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
20 Jan
20 Jan
1:47 a.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Logging is indeed the cause.
0.6 fps with all logging activated
1.3 fps with all logging except debugging
4.4 fps with no login
ver 1.76.40
On Mon, Jan 19, 2015 at 11:13 AM, Manuel AC <manuel.acevedo(a)civantos.com> wrote:
Sorry, I meant 1.76.4x
1.76.42 goes full speed with prores files. Not the Bunny trailer or
the original problematic file that I thought it was prores, but looks
like not.
On Mon, Jan 19, 2015 at 11:11 AM, Carl Hetherington <cth(a)carlh.net> wrote:
>
> So up to now no source files are going at full speed with 1.76, but
> they're ok with 1.74?
>
> And you've tried 1.76 without logging?
>
> Thanks,
> Carl
>
> On Mon, 19 Jan 2015, Manuel AC wrote:
>
>> I will try with 1.69 and all logging off when the computer is available.
>> I used the same disk for a couple prores files using 1.69 and 1.74,
>> full speed, no problems.
>>
>>
>>
>> On Mon, Jan 19, 2015 at 8:17 AM, Carsten Kurz <audiovisual(a)t-online.de>
wrote:
>> >
>> > Am 19.01.2015 um 13:56 schrieb Manuel AC:
>> >
>> >> Yes, the 200% is Big Buck Bunny. You got the log attached in the mail
>> >> for the benchmark.
>> >
>> >
>> > Sorry, missed that log.
>> >
>> > Very strange. Did you try the same with 1.69? Although I don't think it
is something in the code, as it actually went faster on my Mac during the recent updates.
>> >
>> > Maybe the source disc has issues?
>> >
>> > - Carsten
>> >
>> >
>> > _______________________________________________
>> > DCPomatic mailing list
>> > DCPomatic(a)carlh.net
>> > http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
>>
>>
>> _______________________________________________
>> DCPomatic mailing list
>> DCPomatic(a)carlh.net
>> http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
>>
>>
>
> _______________________________________________
> DCPomatic mailing list
> DCPomatic(a)carlh.net
> http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
3:41 a.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Am 20.01.2015 um 02:47 schrieb Manuel AC:
Logging is indeed the cause.
0.6 fps with all logging activated
1.3 fps with all logging except debugging
4.4 fps with no login
Strange. Just checked 1.76.42 on my 'standard' slow test machine (DualCoreHT
notebook) with all logging disabled vs. all logging enabled - no difference. Will need to
check with a faster machine with more cores. It's clear that more threads create more
logging, and the faster the machine, the more the logging slows is down, but I
wouldn't have thought that the impact is so huge. Maybe also should compare Mac vs.
Windows.
- Carsten
3:56 a.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Carl proposed this option at the beginning of this troubleshooting.
Just tried during the encoding with activity monitor open, and
disabling logging had an immediate and undeniable effect.
Logging is slowing the single decoder thread?
I'm sure there is many factors involved, and some of them may be
machine specific.
Thanks!
Manuel AC
On Mon, Jan 19, 2015 at 10:41 PM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
Am 20.01.2015 um 02:47 schrieb Manuel AC:
Logging is indeed the cause.
0.6 fps with all logging activated
1.3 fps with all logging except debugging
4.4 fps with no login
Strange. Just checked 1.76.42 on my 'standard' slow test machine (DualCoreHT
notebook) with all logging disabled vs. all logging enabled - no difference. Will need to
check with a faster machine with more cores. It's clear that more threads create more
logging, and the faster the machine, the more the logging slows is down, but I
wouldn't have thought that the impact is so huge. Maybe also should compare Mac vs.
Windows.
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
10:32 a.m.
New subject: [DCP-o-matic] slow OSX 10.6.8
Hi,
OK, thanks for testing. Maybe we can get to the bottom of that at some
point ...
Best,
Carl
On Mon, 19 Jan 2015 22:56:49 -0500
Manuel AC <manuel.acevedo(a)civantos.com> wrote:
Carl proposed this option at the beginning of this
troubleshooting.
Just tried during the encoding with activity monitor open, and
disabling logging had an immediate and undeniable effect.
Logging is slowing the single decoder thread?
I'm sure there is many factors involved, and some of them may be
machine specific.
Thanks!
Manuel AC
On Mon, Jan 19, 2015 at 10:41 PM, Carsten Kurz <audiovisual(a)t-online.de> wrote:
Am 20.01.2015 um 02:47 schrieb Manuel AC:
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
Logging is indeed the cause.
0.6 fps with all logging activated
1.3 fps with all logging except debugging
4.4 fps with no login
Strange. Just checked 1.76.42 on my 'standard' slow test machine (DualCoreHT
notebook) with all logging disabled vs. all logging enabled - no difference. Will need to
check with a faster machine with more cores. It's clear that more threads create more
logging, and the faster the machine, the more the logging slows is down, but I
wouldn't have thought that the impact is so huge. Maybe also should compare Mac vs.
Windows.
- Carsten
_______________________________________________
DCPomatic mailing list
DCPomatic(a)carlh.net
http://main.carlh.net/cgi-bin/mailman/listinfo/dcpomatic
3751
days inactive
3752
days old
9 comments
3 participants
participants (3)
-
Carl Hetherington -
Carsten Kurz -
Manuel AC