chatted with chyan, we are easier to remove current repository than editing history, since we just have only one commit.
removing binaries from master does not work since they remain as objects in repository, so options are to rebuild repository without these binaries, or to modify repository to use lfs for binaries. But linker objects or built binaries are useless to be in repository, so we are about to select an option to rebuild this repository without these binaries.
I've got archive of the master, so I may be possible to push it to our server as archive/backup.

if anyone have any objection, comment to this ticket in a few days (like by 12 Dec 2017).

I think you can just go ahead and do this without asking for permission via JIRA.

If you want to modify a repo, the easiest way is git bfg

I encountered many times that entire my work had no mean after quite a while (incl. a year span) after I did some work or documentation, by comments or strong direction by management, so taking a time for formal review should be safer.

Thank you for pointing bfg cleaner, but VS has a feature to remove all resultant binaries, so it should be safer to use such official tool, I suppose.

What does, "VS" mean? We're talking git here, recovering from a mistake. Once you've git rm the binaries (and added them to .gitignore) the problem won't reoccur, so the only problem is cleaning up the mess.

I agree that deleting the entire repo and starting over is one way, and git bfg is the semi-standard other way; it's easier than using various git filtering commands.

So I must be missing something.

VS is Visual Studio used for building codes/binaries in this repository.
We may be possible to get a list of extensions for binaries, and supply them to bfg, but building new is easier since there is no history in this repository we need to keep.

Why are we using visual studio to build this code (I assume that we are building on a Linux box)? And what's the relationship between VS removing binaries and this problem?

To be clear I'm perfectly happy with your with deleting and rebuilding this git repository – go ahead – but please explain what the build tools have to do with stopping this problem from reoccurring.

As we heard during a breakout session for PFI control of science WG gathering at MPA 2017, msim is MPS simulator which runs GUI window on Windows OS and simulates MPS to command from GUI and list of commands directly to Cobra FPGA. This ics_msim_proto is the repository for msim code, so this repository is built with VS to make binaries running on Windows OS.

I close this as it is obsolete.