Celso L. Mendes

@MichaelBrim , I am *not* using the MPI_File_sync any more, I was just counting on the sync that Adam enabled via the setting of UNIFYFS_CLIENT_WRITE_SYNC=1. But you raised a good...

@kathrynmohror , @adammoody : I reran today this reproducer example above, under UnifyFS, using my Unify build based on PR#581. This time, with Unify, I obtained the following output: ```...

@adammoody , for this reproducer code posted here it doesn't make any difference to have WRITE_SYNC or not: both versions produce the same output file. However, for the original ROMIO...

I reran this today, on LLNL's Lassen (IBM system), under Unify v.0.9.3, i.e. using a Unify build based on PR#619. The original MPI test code misc.c works correctly, produces the...

@adammoody , many thanks for looking into this. In regard to your comments: a) I'm unclear about fcntl as you have found: didn't Rob Latham tell us that by using...

@adammoody , I did some more tests today with the example code above. I started with the original code, and tried to set in my job script UNIFYFS_CLIENT_LOCAL_EXTENTS=0, but that...

@adammoody , yes, I can certainly try the MPI_File_sync as you suggest, I'll let you know. Just to answer quickly regarding spillover: I had set it to *zero*. That might...

@adammoody , for the records, to address one of the points that had been suggested here: the MPI_File_sync did *not* help, in the original example (i.e. with the default data...

@adammoody , and going back to the chunk-size issue, when I disable data sieving (to produce correct values), and move the SIZE value back to to 5000 as originally, but...

@adammoody , thanks a lot for finding the source of the "-1" for this issue! Also, as you had suggested, I've run this example *without* Unify, and employed the Recorder...