Design puzzles manual elimination

Case number:699969-1998821
Topic:General
Opened by:spmm
Status:Closed
Type:Suggestion
Opened on:Tuesday, December 2, 2014 - 12:05
Last modified:Monday, February 16, 2015 - 21:04

Instead of having us spend days wrestling with filters only to find that top scoring solutions are still helixes why not try a different approach?
Scientists or anyone delegated to do so eyballs the results and immediately disqualifies all helix solutions unless they are expressly permitted. No correspondence will be entered into.
I would prefer to wait 48 hours to get puzzle results, certainly better than wasting time and wrecking my computer in the process of seriously attempting design puzzles.
You could give disqualified solutions a small score maybe.

The only indicator we have is the score, if it doesn't work then just add the human factor on top of the mechanical score. Rather than adding more and more filters which imho wreck the game play and more importantly player hardware.
Also makes it very difficult for players with low end machines to even compete.

Just a thought

(Tue, 12/02/2014 - 12:05  |  5 comments)


Angus's picture
User offline. Last seen 1 day 11 hours ago. Offline
Joined: 06/04/2008
Groups: Beta Folders

I see a lot of problems with this. Where does the line get drawn on how big a helix you can have, or how many small helices? No helices ever allowed in design puzzles? It introduces subjective interference into the scoring which will always cause dissent and disagreement.

I also don't get the "wrecking the computer" thing. Running a PC flat out is not harmfull unless the cooling is sub-standard and allows the CPU to overheat. Many of us have run farms of PCs flat out for years running BOINC projects, fully maxing the CPUs the whole time. Buy a workstation quality machine, or build one. Don't expect a home-use PC rated for a low duty cycle to run flat out, they aren't designed for that. Foldit has always stressed the CPU, with or without filters, don't blame them.

spmm's picture
User offline. Last seen 11 weeks 5 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers

The scientists say what they want in the design puzzle notes - they are the arbiters and imo are perfectly well qualified to make the decision and are analysing the design puzzles anyway. There is no dissent allowed, in the same way that you don't get to choose if your design will be built in the lab now. Quote: 'We like x puzzle solution' not 'this was the best scoring solution'.

I run a workstation quality machine and your advice is not very helpful for the people who can't afford them but who may have good design ideas. I wasn't assigning blame just pointing out that the current heavily filtered design puzzles are slow and overtax computers particularly if people want to work on other puzzles as well.

You have also contradicted yourself by saying 'running a PC flat out is not harmful' followed by 'home-use PC aren't designed for that' thereby implying that they will be damaged by running flat out, you need to add the workstation qualifier to the first statement; but I don't want to have a pedantic argument just made a suggestion.

Angus's picture
User offline. Last seen 1 day 11 hours ago. Offline
Joined: 06/04/2008
Groups: Beta Folders

You left off the rest of the sentence. "... unless the cooling is sub-standard and allows the CPU to overheat." That is one of the more obvious differences between a home-quality machine set up for light duty cycles, and a workstation machine that is designed with improved cooling to allow the CPU to run flat out for long periods. They may use the exact same CPU, but running the home/consumer grade flat out will push the CPU core temps to an unacceptably high level, potentially shortening the life of the CPU.

It is the USERs choice to run their PC beyond the specified duty cycle, and the USER runs the risk of damaging their machine. IF they are concerned, they should not try to run a client for every core 24 hours a day.

All of this goes double for most laptops !

Joined: 09/24/2012
Groups: Go Science

In order to save time on filtered puzzles, I use recipes with filter turning off just before wiggle and on just after, as first proposed by Brow42 here:
DRW for FIlters TvdL 2.1.1

I know that many authors, like TvdL - after experimenting - do not recommend this.

But personally, during hand folding, I prefer to see the counter moving fast on short term, and I believe this strategy is safe for any wiggle side chains and shake, or on end game, when the protein is stable and the filters already high.

Other recipes that use this strategy:
Acid Tweeker V2.4.5
Rav3n_pl GAB v2.0.5 filter

spmm's picture
User offline. Last seen 11 weeks 5 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers
Status: Open » Closed

Closed

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, Boehringer Ingelheim, RosettaCommons