I've just noticed that the 2.17 i'm running on 64bit system with 64bit boinc, named minirosetta_2.17_windows_x86_64.exe is actually a 32bit application. Is it a mistake or it is the way it was meant to be? Was it always this way? If it is meant to be this way, why creating "64bit" version at all?

Same on linux:

mercury boinc.bakerlab.org_rosetta # file minirosetta_2.17_x86_64-pc-linux-gnu
minirosetta_2.17_x86_64-pc-linux-gnu: ELF 32-bit LSB executable, Intel 80386, version 1 (GNU/Linux), statically linked, stripped

---

It's a "wrapper" to make it run on 64-bit systems. But the application itself is 32-bit.

It's always been that way.

---

True, and as to why, if no such application were made available, BOINC server would not send any work to 64bit machines.

---

Rosetta Moderator: Mod.Sense

True, and as to why, if no such application were made available, BOINC server would not send any work to 64bit machines.

Other projects send 32bit application to 64bit systems just like that, without confusing names. Some projects with new server-side software can send 32bit application to 64bit systems even if 64bit application is available, just to figure out which is faster and use faster one later on. But it usually makes people complain. SIMAP had to roll-back to earlier server version because of this not so long ago.
Anyway, is Rosetta's server software so old that it can't even send 32bit application to 64bit system on itself?

---

I'm not positive. Perhaps it was a necessary work around at one point in time. All I can say for sure is that it is correct that there is no true 64bit application.

---

Rosetta Moderator: Mod.Sense

Alright, thanks.

---