Checkpoint?

Message boards : Number crunching : Checkpoint?

Author	Message
dannyridel Send message Joined: 21 Jul 19 Posts: 63 Credit: 15,355,619 RAC: 74,346	Message 6181 - Posted: 14 Apr 2020, 6:16:19 UTC
	I don't seem to see checkpointing in v2 mfakto windows. Is there any problem? On my side? On the app side? Should I be seeing a checkpoint?
	ID: 6181 · Rating: 0 · rate: / Reply Quote

Gigacruncher [TSBTs Pirate] Send message Joined: 28 Mar 20 Posts: 51 Credit: 8,419,360 RAC: 0	Message 6182 - Posted: 14 Apr 2020, 6:25:41 UTC
	http://srbase.my-firewall.org/sr5/forum_thread.php?id=1344&postid=6176#6176
	ID: 6182 · Rating: 0 · rate: / Reply Quote

dannyridel Send message Joined: 21 Jul 19 Posts: 63 Credit: 15,355,619 RAC: 74,346	Message 6183 - Posted: 14 Apr 2020, 6:30:05 UTC
	I'm on Windows with v2, not linux.
	ID: 6183 · Rating: 0 · rate: / Reply Quote

dannyridel Send message Joined: 21 Jul 19 Posts: 63 Credit: 15,355,619 RAC: 74,346	Message 6184 - Posted: 14 Apr 2020, 13:10:34 UTC
	It seems, that when the task is at about 60% a .ckpt file and a .ckpt.bu file appears. I know the .bu is backup, but why do they appear so late and why does the .bu file have a later checkpoint than the original one? see below: .ckpt: 104864329 73 74 4620 mfakto 0.14-Win: 4595 0 026B80C3 .ckpt.bu: 104864329 73 74 4620 mfakto 0.14-Win: 4515 0 FB8118C9
	ID: 6184 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,685,639 RAC: 42,633	Message 6192 - Posted: 14 Apr 2020, 15:03:40 UTC - in response to Message 6184.
	It seems, that when the task is at about 60% a .ckpt file and a .ckpt.bu file appears. I know the .bu is backup, but why do they appear so late and why does the .bu file have a later checkpoint than the original one? see below: .ckpt: 104864329 73 74 4620 mfakto 0.14-Win: 4595 0 026B80C3 .ckpt.bu: 104864329 73 74 4620 mfakto 0.14-Win: 4515 0 FB8118C9 .chpt.bu should only be a backup.
	ID: 6192 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,685,639 RAC: 42,633	Message 6193 - Posted: 14 Apr 2020, 15:04:57 UTC - in response to Message 6181.
	I don't seem to see checkpointing in v2 mfakto windows. Is there any problem? On my side? On the app side? Should I be seeing a checkpoint? Depends on your GPU, on a 5500xt it does every 5min.
	ID: 6193 · Rating: 0 · rate: / Reply Quote

dannyridel Send message Joined: 21 Jul 19 Posts: 63 Credit: 15,355,619 RAC: 74,346	Message 6201 - Posted: 15 Apr 2020, 0:29:08 UTC - in response to Message 6193.
	Okay, on vega8 it happens about that long AFTER 5 hours passed idk why :(
	ID: 6201 · Rating: 0 · rate: / Reply Quote

pls Send message Joined: 18 Jul 20 Posts: 6 Credit: 27,307,980 RAC: 0	Message 6637 - Posted: 18 Jul 2020, 21:24:21 UTC - in response to Message 6201.
	Has there been a real answer to this? I'm seeing that whenever BOINC syspends an SRBASE work unit to run another project, the SRBASE unit is resumed from the beginning. This is wasted work unless I'm only running this one project. Are there any plans to fix checkpointing so that it works correctly? Thanks.
	ID: 6637 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7479 Credit: 43,685,639 RAC: 42,633	Message 6638 - Posted: 18 Jul 2020, 22:00:44 UTC - in response to Message 6637. Last modified: 18 Jul 2020, 22:02:37 UTC
	Has there been a real answer to this? I'm seeing that whenever BOINC syspends an SRBASE work unit to run another project, the SRBASE unit is resumed from the beginning. This is wasted work unless I'm only running this one project. Are there any plans to fix checkpointing so that it works correctly? Thanks. Checkpoints are every 10min (depends on the CPU speed), the runtime is not saved by the wrapper. On windows you can check the stderr.txt file. You can also check the FAQ
	ID: 6638 · Rating: 0 · rate: / Reply Quote

dannyridel Send message Joined: 21 Jul 19 Posts: 63 Credit: 15,355,619 RAC: 74,346	Message 6641 - Posted: 19 Jul 2020, 14:07:57 UTC - in response to Message 6637.
	Has there been a real answer to this? I'm seeing that whenever BOINC syspends an SRBASE work unit to run another project, the SRBASE unit is resumed from the beginning. This is wasted work unless I'm only running this one project. Are there any plans to fix checkpointing so that it works correctly? Thanks. If these tasks are LLR (CPU tasks), don't fret. BOINC doesn't show progress correctly and LLR checkpoints itself. For TF, I'm not sure.
	ID: 6641 · Rating: 0 · rate: / Reply Quote

Post to thread

Message boards : Number crunching : Checkpoint?