It's in the code release directory again.

I've made two changes:

1.At the end of calculate integrals there is:

        #ifdef GMLE_BOINC
                int retval = write_checkpoint(es);
                if (retval) {
                        fprintf(stderr,"APP: astronomy checkpoint failed %d\n",retval);
                        return retval;
                }
        #endif

So in the rare case that this is the last checkpoint calculated (and there hasn't been a new one from the next integral calculation or likelihood calculation), the app will recalculate an integral which would give a bad result.

The other change I did was move the code to checkpoint from the innermost integral loop, to the middle loop. This should speed things up a little (and maybe more than a little on older architectures).

With the updates to 0.14 it seems like the bad results were quite a bit less frequent, so I'm hoping this was the last bit of the issue.

---

http://milkyway.cs.rpi.edu/milkyway/download/code_release/ <-- this is better

---

me@rescam.org

http://milkyway.cs.rpi.edu/milkyway/download/code_release/ <-- this is better

I fail at the interwebs tonight :(

---

http://milkyway.cs.rpi.edu/milkyway/download/code_release/ <-- this is better

I fail at the interwebs tonight :(

Then I await v0.16 in the morning. Nice RAC.

---

me@rescam.org