Program size getting huge

Case number:699969-995659
Topic:General
Opened by:auntdeen
Status:Open
Type:Suggestion
Opened on:Monday, August 5, 2013 - 02:29
Last modified:Friday, August 9, 2013 - 18:01

The size of the program is increasing at a rapid rate (Mac OS 10.7.5 desktop)

I just had 4 clients running for two and a half days, and the log.txt was an astounding 58.6 MB.

In early July the size of my program was 262.8 MB - two different clients, one DP and one Main (I keep old clients so I just checked). Currently my client is a whopping 923.7 MB, and that's up from 868 MB yesterday. I'm obviously going to need to do a fresh install, with the hassle of uploading every save for myself first.

If this is result of the precise wiggle, I suspect that I will not be using it. As it is, it makes my fans loud. The only other thing that gets my fans cranking is mutate, so I do very little of those puzzles.

It's getting to the point where computing power is becoming supreme here. I am lucky enough to have a 1 year old, top of the line Mac. I tested the precise versus the less precise on 755 by running an overall rebuild script that has 30 loops. The less precise took 12 hours to complete. I finally killed the precise run after 48 hours that only made it to loop 27. Yes, the precise gained more points.

But how many folders have the computing power to test the different server models, with run times like that for scripts, and with the current bloat in the program?

(Mon, 08/05/2013 - 02:29  |  11 comments)


MurloW's picture
User offline. Last seen 22 weeks 5 days ago. Offline
Joined: 11/21/2012

58.6MB is astounding? Right now, after 3 days and 2 nights of scripting 3 clients my log.txt is 96.85MB, which is not out of the ordinary for what I've seen. I've reported before on logs becoming over 100MB in size, but nobody went "Wow, that's big!". Like that one time, I posted in feedback that FoldIt was crashy (it still is, but that post had specifics) and my log.txt was ~115MB, (definitely over 100MB, I couldn't upload the thing to my feedback post.) with about 3/4 of it being empty lines. Just blank space, no dev thought that was odd, or at least they didn't react as such.

My program size is near 1GB if the puzzle running is a large, mutating symmetry; also: subsequent clients being opened require more memory, regardless of puzzle opened.

For example: first client, puzzle 755 solo, 650~800MB running size. Second client: puzzle 753 solo, 800~950MB run size. Third client: puzzle 755 evo, 750~900MB run size. Fourth client, puzzle 753 evo, 900MB~1GB+ run size. Fifth client, and any more: any puzzle, 1GB+ run size guaranteed.

Obviously the puzzle being worked on affects running size, but the systematic increase in memory needed is there even if I opened 5 clients of the same puzzle and left them idle.

My machine is a few years old, but it was hardcore when I put it together;
and I am certainly not testing every server model or trying multiple poses on the symmetry because I simply do not have the computing power to pull it off in a week. A few months ago, I did. And it was fun.

jflat06's picture
User offline. Last seen 7 hours 31 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

Can you give me a the build-id of the client that is only using 262.8 MB? Thanks!

Joined: 04/19/2009

Here they are - hopes this helps:

20130724-39923ab87c-osx_x86-devprev

20130702-6bc01bba0e-osx_x86

Joined: 05/19/2009
Groups: Contenders

Auntdeen, how much real memory (RAM) do you have installed ? The more the merrier your system will run.

Joined: 04/19/2009

I have 8 MG.

Probably getting "dated" now - but a year ago the configuration was in the $2500 neighborhood (which Apple gave me free because I had had 4 displays go bad on the previous Mac).

Joined: 05/19/2009
Groups: Contenders

4 clients in 8 Gigabyte is on the edge, I would recommend more RAM. I unfortunately do not have enough knowledge of Macs to tell you how to switch off the disk swapping (virtual memory system) because that is what makes the system very very slow, it swaps programs and data out from memory to disk and back. 16 or more Gb is recommendable. You may also not have more CPU cores to handle work so that could be a limiting factor.

marie_s's picture
User offline. Last seen 1 year 37 weeks ago. Offline
Joined: 05/18/2008
Groups: None

On the new wiggle, I appreciate the result it gives in hand folding and I have no problem on scripts.
I have a new good laptop on windows.

We need more example of players playing on older computer on some recipes to see if there is a problem.
If some pleyer cannot play :
- 20 generations of gab bis,
- a loop a loop rebuild,
- 30 steps of drw.
with precise wiggle then there is a problem

I think that 30 loops of overall rebuild is not mandatory to play Foldit with pleasure and results.

Joined: 09/21/2011
Groups: Void Crushers

Marie, I have done some comparisons earlier. I have a pretty fast laptop, but running DRW with precise wiggle is not an option for me. I run everything on less precise.
Notice the measurements where on a small puzzle. Running times are at least kwadratic on the puzzle size.
So on the latest big one 1 precize mini DRW cycle will probably take about 2 hours or more.
Even the less precision wiggle is now too slow for comfort!!!
------
Here are the measurements I did:
Comparison when the new wiggle came into devprev.
My DRW 2.3.0 1 1 mini cycle (so only 1 part is rebuild)
Main: 12 minutes
Devprev: 13 minutes (but 1 qstab more which is 1 wiggle all(2) extra)
Devprev in background: 7 minutes

Another script where wiggle all(2) is called for every segment:
Main: 13 minutes
Devprev: 11:45 minutes

Conclusion: the playing field is now level. (Add sarcasm)
Scripts are becoming too slow for confort.

And this was on puzzle 747, only 71 segments. Bigger puzzles will be a lot worse.

After the latest update (with less precision):
Ran the same scripts as with the previous update on the same position:
DRW took 4 minutes
The other script also took 4 minutes
So the speed is 3 times as fast.
The other script however (an endgame script) did not find 10 pts like before but 8 pts, as can be expected.

frood66's picture
User offline. Last seen 20 min 21 sec ago. Offline
Joined: 09/20/2011
Groups: Marvin's bunch

This is probably one of the best feedbacks for a long time - will be interesting to see if it is ignored

frood66's picture
User offline. Last seen 20 min 21 sec ago. Offline
Joined: 09/20/2011
Groups: Marvin's bunch

just scrolled up - 20 gens GAB bis. Marie - I dream of this. I can do perhaps 5 gens on a smaller puzzle if left o'nite (if I'm lucky)

jflat06's picture
User offline. Last seen 7 hours 31 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

The increase in size seems to be due to us merging in the most recent changes to Rosetta. There's the potential that we don't need all of this stuff, so we'll look into shrinking it.

Sitemap

Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons