Re: New Benchmark Tool for LyX & Results.

2011-07-18 Thread Jean-Marc Lasgouttes
Le 14/07/11 14:45, John McCabe-Dansted a écrit : Nice, but the benchmark still reports 800M of memory used in r39299. I've attached some output from valgrind/massif. This can be pretty printed using ms_print This massif output indicates that you have compiled with stdlib-debug enabled (see th

Re: New Benchmark Tool for LyX & Results.

2011-07-14 Thread Abdelrazak Younes
On 14/07/2011 12:26, Abdelrazak Younes wrote: On 14/07/2011 12:03, Abdelrazak Younes wrote: On 14/07/2011 10:14, Abdelrazak Younes wrote: On 14/07/2011 08:37, John McCabe-Dansted wrote: Anyway: I have reproduced the increase in memory use with =rel as well. It appears to be caused by the addit

Re: New Benchmark Tool for LyX & Results.

2011-07-14 Thread Abdelrazak Younes
On 14/07/2011 12:03, Abdelrazak Younes wrote: On 14/07/2011 10:14, Abdelrazak Younes wrote: On 14/07/2011 08:37, John McCabe-Dansted wrote: Anyway: I have reproduced the increase in memory use with =rel as well. It appears to be caused by the additions to lib/layouttranslations. Does LyX make a

Re: New Benchmark Tool for LyX & Results.

2011-07-14 Thread John McCabe-Dansted
On Thu, Jul 14, 2011 at 2:37 PM, John McCabe-Dansted wrote: > Anyway: I have reproduced the increase in memory use with =rel as > well. It appears to be caused by the additions to > lib/layouttranslations. Does LyX make an in memory copy of the > translations for every inset? LyX appears to be usi

Re: New Benchmark Tool for LyX & Results.

2011-07-14 Thread Abdelrazak Younes
On 14/07/2011 10:14, Abdelrazak Younes wrote: On 14/07/2011 08:37, John McCabe-Dansted wrote: Anyway: I have reproduced the increase in memory use with =rel as well. It appears to be caused by the additions to lib/layouttranslations. Does LyX make an in memory copy of the translations for every

Re: New Benchmark Tool for LyX & Results.

2011-07-14 Thread Abdelrazak Younes
On 14/07/2011 08:37, John McCabe-Dansted wrote: Anyway: I have reproduced the increase in memory use with =rel as well. It appears to be caused by the additions to lib/layouttranslations. Does LyX make an in memory copy of the translations for every inset? I had a look at the code and it doesn'

Re: New Benchmark Tool for LyX & Results.

2011-07-13 Thread John McCabe-Dansted
On Thu, Jul 14, 2011 at 6:27 AM, Tommaso Cucinotta wrote: > Il 13/07/2011 20:32, John McCabe-Dansted ha scritto: > one trivial questions (sorry to ask): > 1) are you sure power management was disabled on your machine while running > these benchmarks ? Powermanagement was probably ondemand. The re

Re: New Benchmark Tool for LyX & Results.

2011-07-13 Thread Richard Heck
On 07/13/2011 06:27 PM, Tommaso Cucinotta wrote: > Il 13/07/2011 20:32, John McCabe-Dansted ha scritto: >> I have begun developing benchmark utilities for LyX. These can be >> found by doing >>git clone g...@github.com:gmatht/Jankey.git >>cd Jankey/Benchmarks/ >> >> Running these benchmarks

Re: New Benchmark Tool for LyX & Results.

2011-07-13 Thread Tommaso Cucinotta
Il 13/07/2011 20:32, John McCabe-Dansted ha scritto: I have begun developing benchmark utilities for LyX. These can be found by doing git clone g...@github.com:gmatht/Jankey.git cd Jankey/Benchmarks/ Running these benchmarks reports that that the resource usage of LyX has been growing in a

New Benchmark Tool for LyX & Results.

2011-07-13 Thread John McCabe-Dansted
I have begun developing benchmark utilities for LyX. These can be found by doing git clone g...@github.com:gmatht/Jankey.git cd Jankey/Benchmarks/ The Benchmark basically involves spamming the following keycodes at LyX \Ac \D1 \Ac \D9 \D9 \Ao \D1\Av\D10 \Ai n n nnn \[Right] asdf \r \r iadsf