Fix of Ram Issue #764 -- Pre-proposal discussion, Dev attention needed if possible

[Cryptosi]

This is the First effort to have an external developer fix an issue and have his/her work funded by the PIVX budget.

This pre-proposal is to discuss developer VThomas fixing Issue #764 in PIVX Github.

please read this google doc for more information.

IDP issue #764 - PIVX

Fix of Ram Issue #764 Independent development proposal for issue #764 https://github.com/PIVX-Project/PIVX/issues/764 Prepared by Cryptosi & VThomas OVERVIEW & PURPOSE This proposal is for VThomas to carry out independent Development work, towards solving the Github issue #764 which relates to h...

docs.google.com

All feedback is considered useful.

Thanks - Cryptosi

 

[warrows]

Disclaimer: Following are my personal views and I don't represent the whole dev team or other individual devs in any way.

It sounds like a mostly good idea. However I ran valgrind a few days ago and found only 8MB of leaks. Some of it are in libraries and for the rest I'll have fixes up soon. Anyway, it's not much.

For the rest, I'm pretty sure there is progress to be made and memory to be saved in the codebase, but I'm also pretty sure it won't be a miracle. PIVX is taking up memory mostly because its features like zPIV require it (even if it's disabled, it's still in the code, which might change soon or not but won't depend on an audit)

Anyway that's one of the things on my todo list. I don't mind seeing it done by someone else.

However I'm not sure paying months of work upfront without a detailed plan is the best way to go. And I don't see a detailed plan nor any track record here: Github is linked but empty. No way to know if the guy is good or not.
I'd like to at the very least see some code or recommandations. If not, the best way to go is to submit one or two small PR to show one's worth and then request money for more work. It's open source after all.

 

[presstab]

Replied to this on github. Would be happy to help implement a fix to this issue, for a fraction of the proposed cost. Have already chopped off 0.8gb without too much time spent on it.

 

[vthomas]

Well I am pretty new to Freelancing and Github. I have contributed to open source when it was on sourceforge and a C++ based Machine Learning Library. Apart from that I have worked in the Industry for companies like Symantec and Siemens. I guess that should prove my credentials.

If presstab has really done what he said...What was stopping him all this while. why did he not do this earlier??

What i propose is a total redo of the code with the memory footprint in focus. Any 32 bit process that takes like 1GB of RAM is asking for trouble. Trust me I know this from experience.
Even 64 bit processes can have problems with this if you do not hit the Cache often. So performance tuning is another aspect that I will be doing.

Also this code has to run on varied computers of nodes where it will have to perform well for its reputation.

So
1) I guess you have my credentials to prove by looking at some sample C++ code on Github. Most of my C++ code is proprietary.
2) I only asked for 4000 PIVX, which is like 2800 USD at todays rate. Thats not much in my opinion.

@warrows...if you need any more information about my credentials or something...please contact me through cryptosi.

 

[vthomas]

OK, Since 12K PIV for 3 months is something not agreeable to everyone. I am willing to reduce this to 3000 PIV for 1 month. Hopefully I can reduce or do as good a job as @presstab