Ryzen on Linux

Message boards : Number crunching : Ryzen on Linux
Message board moderation

To post messages, you must log in.

Previous · 1 · 2

AuthorMessage
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1877073 - Posted: 6 Jul 2017, 6:00:09 UTC - in response to Message 1877003.  

My NV code will not do any checkpoints in the next version. A restart of a task will start from 0% but that is not so a big loss, is it?

If you're running a TitanXp, no.
But if you're running less capable hardware that takes more than a couple of minutes to process a WU, then yes, it is a big loss.
Grant
Darwin NT
ID: 1877073 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2786
Credit: 685,657,289
RAC: 835
Canada
Message 1877076 - Posted: 6 Jul 2017, 8:04:57 UTC - in response to Message 1877073.  

I have been scratching my head on that comment from Petri for awhile. Sure it would be fine to not checkpoint for a dedicated cruncher, but definitely not for a user that has 'only when idle' selected. Most of the time mine are set to 2 minute checkpoints, since they seldom turn off.

My runtimes (max) are around...
750Ti - 13.5 min
980 - 5 min
1070 - 4 min
1080 - 3.5 min

I think a command line switch to turn it off, or reading user prefs for idle flag (to turn it on) would be required. I'm guessing the reason you would like to do this is to save a little processing time recording those values.

But I do think for any non-dedicated machine it would render any small improvements gained very little compared to what would be lost by most people.

I would lose an average of ~5 minutes on my 750Ti for each restart, but for me it isn't bad since My AMD also is sitting on day 19 of uptime, so very little. But compare that to having an idle time set, and it could add up to hours of lost work per day.

Adding a command line option would be the way to go, but that is probably a heck of a pile of coding for little change.
ID: 1877076 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13736
Credit: 208,696,464
RAC: 304
Australia
Message 1877080 - Posted: 6 Jul 2017, 8:55:06 UTC - in response to Message 1877076.  

My runtimes (max) are around...
750Ti - 13.5 min

13.4min done & something happens. That's a lot of crunching to redo for the sake of a checkpoint.
Grant
Darwin NT
ID: 1877080 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34258
Credit: 79,922,639
RAC: 80
Germany
Message 1877097 - Posted: 6 Jul 2017, 14:39:21 UTC

Win 7 is up and running now.

Looks much better now.
There is still a difference but not that much.


With each crime and every kindness we birth our future.
ID: 1877097 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1877099 - Posted: 6 Jul 2017, 15:18:50 UTC - in response to Message 1877097.  

Win 7 is up and running now.

Looks much better now.
There is still a difference but not that much.

That looks exactly the same for Arecibo VLARs on the AVX app for me. I find the BLC tasks run about 30 minutes faster than the Arecibo.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1877099 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34258
Credit: 79,922,639
RAC: 80
Germany
Message 1879909 - Posted: 23 Jul 2017, 14:13:25 UTC

Here is a pic of my mobo.




With each crime and every kindness we birth our future.
ID: 1879909 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1879918 - Posted: 23 Jul 2017, 15:39:19 UTC - in response to Message 1879909.  

Nice! Curious ..... what case is that? Is it by chance a Coolermaster HAF922? Looks very familiar since I have that case still in the closet as a cast-off.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1879918 · Report as offensive
Profile Mike Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 17 Feb 01
Posts: 34258
Credit: 79,922,639
RAC: 80
Germany
Message 1879983 - Posted: 23 Jul 2017, 22:34:40 UTC

Its a Corsair Graphite 600T.


With each crime and every kindness we birth our future.
ID: 1879983 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1879986 - Posted: 23 Jul 2017, 22:52:38 UTC - in response to Message 1879983.  

Thanks. Wondered that's all. I was trying to reconcile the case floor grommets but the hard drive cages looked familiar. From memory as I hadn't bothered to unbury the case from the closet.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1879986 · Report as offensive
Profile Darrell
Volunteer tester
Avatar

Send message
Joined: 14 Mar 03
Posts: 267
Credit: 1,418,681
RAC: 0
United States
Message 1879992 - Posted: 23 Jul 2017, 23:33:01 UTC - in response to Message 1879909.  

When was the last time you dusted that unit out?:)
ID: 1879992 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13164
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1880013 - Posted: 24 Jul 2017, 2:26:17 UTC - in response to Message 1879992.  

When was the last time you dusted that unit out?:)

No, the HAF922 is a skeleton now. I am a terrible packrat. I never throw anything out and castoff hardware takes up an entire room. What I usually do is survey the bits and pieces on the shelf gathering dust and see what I might have to procure to make another system. That is the evolution of the new Linux special app machine. I have kept relatives in computers since the late 80's.

The working computers get blown out at least every couple months or earlier if I am installing hardware or something. The cat hair from 3 "children" gets into everything.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1880013 · Report as offensive
Profile -= Vyper =-
Volunteer tester
Avatar

Send message
Joined: 5 Sep 99
Posts: 1652
Credit: 1,065,191,981
RAC: 2,537
Sweden
Message 1880059 - Posted: 24 Jul 2017, 14:06:02 UTC - in response to Message 1877003.  

My Ryzen isn`t running very hot.
I usually have 51° to 55°C running 6 instances on CPU plus GPU.
I have a Noctua D15 installed and its not even running full load atm.
Of course you are overclocked with your 1700 which is different.

I have Linux running on a 64 GB USB 3 stick. That was just 60 Bucks.
It runs smoothly and very fast.
You just need some time to read thruough everything.


A warning is at a place here (Eine kleine warnung).

Boinc and the seti app both write to disk (USB). BOINC writes just as it is told to do in the user interface (once in a minute or so).
The (GPU) App writes checkpoints hundreds of times a second unless it is modified in the source code to not to do so.
Please make a backup for your own sake.

My NV code will not do any checkpoints in the next version. A restart of a task will start from 0% but that is not so a big loss, is it?

Viel Grüss,
Petri


This is solved by Reading this https://lonesysadmin.net/2013/12/22/better-linux-disk-caching-performance-vm-dirty_ratio/ and check the section for Approach 2!

_________________________________________________________________________
Addicted to SETI crunching!
Founder of GPU Users Group
ID: 1880059 · Report as offensive
Previous · 1 · 2

Message boards : Number crunching : Ryzen on Linux


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.