Problems and Technical Issues with Rosetta@home

Message boards : Number crunching : Problems and Technical Issues with Rosetta@home

To post messages, you must log in.

Previous · 1 . . . 171 · 172 · 173 · 174 · 175 · 176 · 177 . . . 300 · Next

AuthorMessage
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5691
Credit: 5,859,226
RAC: 0
Message 104656 - Posted: 4 Feb 2022, 18:18:10 UTC - in response to Message 104655.  

Give up on a M.2 for BOINC. BOINC seems to hate them

I don't know. I have 3 computers running BOINC on M.2 drives. 2 on PCIE M.2, and one on a SATA M.2. All work fine . One is on Windows 11, two on Linux Mint 20.2 or 3. I had issues trying to run BOINC with Virtual Box on a Windows 10 computer, with 2.5 inch SATA SSD drives.


I am not saying that issues don't exist. But it is probably not M.2 per se. Possibly the type or brand of disk, the impolementation of the M.2 spec on the motherboard, or the driver. Similarly, I had issues with a cheap Chinese SSD (Foresee brand) that came bundled with my firewall hardware. It would simply not run FreeBSD (PFSense or OPNSense) correctly. Replacing the SSD with a cheap yet better one (Trancend) fixed the problem. The format was mSATA. Not exactly the same as an M.2 SATA, but the general idea applies.

Anyway, the issues could be with M.2 but more likely with something else revolving around the M.2 connexion such as the drive itself or the driver.

My 2 cents. hope it helps.



It could be a board/drive thing sending mixed signals.
I'll repeat the scenario again, the M.2 will hold data and operate as a drive for windows purposes.
BUT as a BOINC data holder it wipes out the ability to see the GPU's (no windows errors, no errors from FAH either and no errors saying PCIe lanes were full) only the BOINC error saying no GPU's detected.
I asked in BOINC forum and they had no answers other than to check everything that we already checked here and what I have tried. I spent a week with a few hours a night trying to figure things out with no luck.

I'm not going to waste time or money to go through a bunch of M.2 drives to find one that will work.
I'm picking up my Samsung EVO 500GB (BOINC Data was stored on my EVO 250GB before Python made me determine I need a separate drive for Python and other projects as multiple python projects swallowed up the leftover drive space in addition to my other projects) this weekend. I'm running off a 1TB HDD now that is supposed to be my deep storage drive.

The drive was not defective, even though the Firmware could not be flashed (also odd since it flashed ok at the stores tech bench) and it could store information and read and write ok.

Just something about a M.2 980, a MSI x 470 gaming plus and BOINC mix that causes this.
Even dug down into how many PCIe lanes were in use and yes, I still have 4 left for a M.2 drive.

I'm just pushing my machine to use EVERY aspect of its capability, Vbox,normal boinc, gpu cards and the such along with a 8 core 16 thread processor (15 dedicated to BOINC) plus FAH and internet (spotify you tube, DVD playback)
ID: 104656 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104659 - Posted: 4 Feb 2022, 22:04:03 UTC - in response to Message 104653.  

One of the points I made was one I got from you - that if it was possible to allocate 150% of RAM to Boinc, people would, because tasks never actually used that amount or anything near it while running - often less than 10% of it.
Well bugger me from behind, someone took me seriously. I'm actually a very intelligent person with 1000% common sense, which is probably why nobody else has it anymore.

This is a very good example of your posts.
One sentence that makes sense, while another that no-one in their right mind would believe.
Just because most of what you write is nonsense doesn't detract from the germ of an idea that has something to it.
I'm a glass half-full person.
(Well, tbh I'm not, but for the purposes of this weird role I've given myself, I am)

My point at the time being, people get angry or upset when tasks don't or can't run because they're so eager and keen to be able to contribute their CPU time for the project, so anything that can be done to make it possible for them to contribute as much as possible makes it win-win for everyone. And they met us halfway.
You're going to have to tell me which of my sentences makes sense and which you don't believe.
ID: 104659 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
.clair.

Send message
Joined: 2 Jan 07
Posts: 274
Credit: 26,399,595
RAC: 0
Message 104660 - Posted: 4 Feb 2022, 22:05:41 UTC

I recently put in a Samsung 870 evo ssd sata3 2.5" drive on an asus Z9PE-D8 WS board , and it went in and works without problem , sort of . .
now my drive for backups `E` keeps unmounting itself and disappearing of the system and I have to "rescan disks" in `disk management` to use it,
and that is with 14 sata ports and 7 pcie slots to get lost in , and I have tried moving to other plugs cables etc
boinc works , even though I today found it using 325GB of disk space [and all the work units to go with it]
and more zombie tasks to kill
ID: 104660 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104661 - Posted: 4 Feb 2022, 22:12:22 UTC - in response to Message 104660.  

I just spent 4 hours shuffling 15 RAM sticks around between two Xeon servers to find what combination would work best. Sometimes one would be disabled due to a "membist failure" which I assume is a self test, even though I have that turned off in the BIOS. But in a different slot, that one is ok. All I was trying to do originally is stop them saying "press F1 to continue" at startup due to the RAM not being in the most efficient layout. Turns out I have 4 different "geometries" of RAM (never realised that would matter) and there's no way to make them both happy. But then I found a "disable F1 at startup" in the BIOS. Why didn't I see that before? Maybe it came in with a BIOS upgrade. Stupid Dell design!
ID: 104661 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Sid Celery

Send message
Joined: 11 Feb 08
Posts: 2118
Credit: 41,165,934
RAC: 13,955
Message 104662 - Posted: 5 Feb 2022, 1:43:45 UTC - in response to Message 104659.  

You're going to have to tell me which of my sentences makes sense and which you don't believe.

You're probably the only person who doesn't know, especially as I already gave the answer as to which I paid attention to...
ID: 104662 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
.clair.

Send message
Joined: 2 Jan 07
Posts: 274
Credit: 26,399,595
RAC: 0
Message 104663 - Posted: 5 Feb 2022, 2:36:04 UTC

different "geometries" of RAM

I take it you found the layout of ram `banks` like 2Rx4 and 4R4 on ram sticks shure do funky stuff if the board can`t take it
ID: 104663 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104664 - Posted: 5 Feb 2022, 3:46:03 UTC - in response to Message 104662.  

You're going to have to tell me which of my sentences makes sense and which you don't believe.

You're probably the only person who doesn't know, especially as I already gave the answer as to which I paid attention to...
I'm not going to waste time on your childish riddles.
ID: 104664 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104665 - Posted: 5 Feb 2022, 3:48:35 UTC - in response to Message 104663.  
Last modified: 5 Feb 2022, 3:51:33 UTC

different "geometries" of RAM
I take it you found the layout of ram `banks` like 2Rx4 and 4R4 on ram sticks shure do funky stuff if the board can`t take it
The board will take any of them, but if they're not all the same, it can't dual channel them etc.

Do you know what those things mean? For example, I have 16GB 2Rx4, also 16GB 4Rx8. Surely 4Rx8 is 4 times bigger than 2Rx4? It's twice as big in both dimensions. I would have expected a 2Rx8 and a 4Rx4 to be equal in size. Perhaps there are more chips on each. I can't be bothered looking.
ID: 104665 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
.clair.

Send message
Joined: 2 Jan 07
Posts: 274
Credit: 26,399,595
RAC: 0
Message 104671 - Posted: 5 Feb 2022, 21:45:09 UTC
Last modified: 5 Feb 2022, 22:00:32 UTC

I did look it up a while ago and each stick is internaly divided up into banks of chips for some purpose I don't remember
Had another look.
type this int search - "what does 4rx8 ram meaning"
this brings up a lot of stuff to do with how the memory controler addresses the memory chips and how they are phisicaly connected on each stick
A16GB stick is a 16GB stick however it is internaly wired,
older motherboards and domestic boards cant take the multi bank type`s for servers
Now I have a headache, time for some parotcetamol . . .
ID: 104671 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104674 - Posted: 5 Feb 2022, 23:15:54 UTC - in response to Message 104671.  
Last modified: 5 Feb 2022, 23:16:22 UTC

I did look it up a while ago and each stick is internaly divided up into banks of chips for some purpose I don't remember
Had another look.
type this int search - "what does 4rx8 ram meaning"
this brings up a lot of stuff to do with how the memory controler addresses the memory chips and how they are phisicaly connected on each stick
A16GB stick is a 16GB stick however it is internaly wired,
older motherboards and domestic boards cant take the multi bank type`s for servers
Now I have a headache, time for some parotcetamol . . .
I assume this is the same thing as "high density memory" which was being sold very cheaply on Ebay about 10 years ago. I bought a metric tonne of it for 500 computers at work. Only half of them acknowledged its existence. The sellers often said things like "only suitable for high end gaming machines" as though it was some special kind of memory. In fact it was just much cheaper to make, or was found incompatible with so many MBs that it got sold off cheap. We had Dell machines at work, and Dell is notorious for being incompatible with everything. I used as much as I could then resold the rest.
ID: 104674 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104684 - Posted: 6 Feb 2022, 22:14:24 UTC

Damnit Greg! You deleted your post while I was replying to it, and now mine has vanished!

The main points:

Doesn't make sense, you would think a drive access problem would return a file not found error, not a GPU error. I've like you only used C or C/E, so I don't know about D, but I'm sure someone must have done it and would have reported this bug.
Try making your new drive be F.
Or don't split install, just put it on the new drive only.
ID: 104684 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5691
Credit: 5,859,226
RAC: 0
Message 104685 - Posted: 6 Feb 2022, 22:24:38 UTC

Problem solved and so freaking simple!
apparently data has to be stored in a folder on the drive rather than out on the drive itself

So maybe this also solves the M.2 problem?
I am on SATA now...so no idea
ID: 104685 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104686 - Posted: 6 Feb 2022, 22:40:41 UTC - in response to Message 104685.  
Last modified: 6 Feb 2022, 22:44:19 UTC

Problem solved and so freaking simple!
apparently data has to be stored in a folder on the drive rather than out on the drive itself

So maybe this also solves the M.2 problem?
I am on SATA now...so no idea
Ah! Good to know (although not so good for you, you've lost out on an M2). I'll report this as a bug to the programmers.
ID: 104686 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5691
Credit: 5,859,226
RAC: 0
Message 104688 - Posted: 6 Feb 2022, 22:59:36 UTC - in response to Message 104686.  

Problem solved and so freaking simple!
apparently data has to be stored in a folder on the drive rather than out on the drive itself

So maybe this also solves the M.2 problem?
I am on SATA now...so no idea
Ah! Good to know (although not so good for you, you've lost out on an M2). I'll report this as a bug to the programmers.



So if out in the open on the drive you lose the GPU, if in a folder on the drive then everything is fine.
Took me 2 weeks and a lot of hassle to figure that out, it was only after the error that I got a light bulb idea.
God, how annoying.

So if you get a M.2 then try this first if you get GPU errors.
I reported it in my thread on BOINC Forum.
Maybe someone there will notice as well.
The only guy helping me is a Dutch guy who has some RL issues taking priority.
Several other piped in at the start with all the usual stuff.
Anyway....bed time here..monday comes fast.
ID: 104688 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104689 - Posted: 6 Feb 2022, 23:39:21 UTC - in response to Message 104688.  

Problem solved and so freaking simple!
apparently data has to be stored in a folder on the drive rather than out on the drive itself

So maybe this also solves the M.2 problem?
I am on SATA now...so no idea
Ah! Good to know (although not so good for you, you've lost out on an M2). I'll report this as a bug to the programmers.



So if out in the open on the drive you lose the GPU, if in a folder on the drive then everything is fine.
Took me 2 weeks and a lot of hassle to figure that out, it was only after the error that I got a light bulb idea.
God, how annoying.

So if you get a M.2 then try this first if you get GPU errors.
I reported it in my thread on BOINC Forum.
Maybe someone there will notice as well.
The only guy helping me is a Dutch guy who has some RL issues taking priority.
Several other piped in at the start with all the usual stuff.
Anyway....bed time here..monday comes fast.
Programmer is asking if you have Boinc installed as a service?
ID: 104689 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
.clair.

Send message
Joined: 2 Jan 07
Posts: 274
Credit: 26,399,595
RAC: 0
Message 104691 - Posted: 7 Feb 2022, 2:46:14 UTC

On mine it`s
computer/F/ProgramData/BOINC/notices+projects+slots folders and all the other files, and F/Program Files/BOINC/locals+skins folders and the other files
is that what you ment by in a folder
I first created the `Program data` and `Program files` folders on the disk
Then installed boinc to the drive by changing the `advanced` settings , it did not like that because it created another folder named `ProgramData` different spelling.
I did not chose `F` that was the next available drive letter.
It does now seem an appropriate letter of the alphabet for it to use considering all the funky stuff it has done.
ID: 104691 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5691
Credit: 5,859,226
RAC: 0
Message 104692 - Posted: 7 Feb 2022, 11:32:44 UTC - in response to Message 104689.  

Tell programmer guy, BOINC was installed on its default settings, only change was where the data was stored.

So it was C: program files/BOINC

Then data was attempted to be installed as just D: (no folder) and after the errors D: Data (folder name) because bit was C: BOINC Data folder and E: BOINC Data folder.

But if it wants a folder, then why not say so during installation or in the startup log into?

It's weird that CPU will run with no folder but GPU requires a folder.

Have the person explain that!
ID: 104692 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mr P Hucker
Avatar

Send message
Joined: 12 Aug 06
Posts: 1600
Credit: 11,724,373
RAC: 9,452
Message 104694 - Posted: 7 Feb 2022, 14:05:03 UTC - in response to Message 104691.  
Last modified: 7 Feb 2022, 14:06:02 UTC

On mine it`s
computer/F/ProgramData/BOINC/notices+projects+slots folders and all the other files, and F/Program Files/BOINC/locals+skins folders and the other files
is that what you ment by in a folder
I first created the `Program data` and `Program files` folders on the disk
Then installed boinc to the drive by changing the `advanced` settings , it did not like that because it created another folder named `ProgramData` different spelling.
I did not chose `F` that was the next available drive letter.
It does now seem an appropriate letter of the alphabet for it to use considering all the funky stuff it has done.
I have mine in E:Program DataBoinc, which I created. I only noticed it was different when trying to copy configs between computers and kept getting path not found. The ones that had put it on C: themselves as the default had no space between program and data. A Microsoft bug. FFS, all the backslashes go away in this forum. I even tried the code tag but it still removed them!
ID: 104694 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
.clair.

Send message
Joined: 2 Jan 07
Posts: 274
Credit: 26,399,595
RAC: 0
Message 104698 - Posted: 7 Feb 2022, 17:10:18 UTC

Oops , I forgot about that , I was thinking in Linux // Vs
And so it duz I put two backslashes after the Vs to try it
I suppose it is seen as a part of the BB forum tags and so is hidden
Arrr :-) , but if you put fore in you get one \
so lets try ten \\ Hmm got two
Ok twenty backslashes \\\\\ I know I am being silly . . . .
Ok so I got five
Game over :-}
ID: 104698 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Greg_BE
Avatar

Send message
Joined: 30 May 06
Posts: 5691
Credit: 5,859,226
RAC: 0
Message 104700 - Posted: 7 Feb 2022, 18:19:10 UTC

I have 49 gigs of memory and Python x 13 used up close to 65% of that memory and nothing else could run.
Then BOINC got a brain and sidelined a bunch and put SiDock x 5 into play. and I am down to 56% usage.

Such pigs of memory these Pythons.

BOINC complained with this:

LHC@home: Notice from server
CMS Simulation needs 7629.39MB more disk space. You currently have 0.00 MB available and it needs 7629.39 MB.
2/7/2022 7:55:21 AM DESKTOP-LFM92VN
--------------------------------------------------------------------------------
LHC@home: Notice from server
Theory Simulation needs 7629.39MB more disk space. You currently have 0.00 MB available and it needs 7629.39 MB.
2/7/2022 7:55:21 AM DESKTOP-LFM92VN
--------------------------------------------------------------------------------
LHC@home: Notice from server
ATLAS Simulation needs 9536.74MB more disk space. You currently have 0.00 MB available and it needs 9536.74 MB.
2/7/2022 7:55:21 AM DESKTOP-LFM92VN
--------------------------------------------------------------------------------
QuChemPedIA@home: Notice from server
NWChem needs 4096.03MB more disk space. You currently have 0.00 MB available and it needs 4096.03 MB.
2/7/2022 7:56:29 AM DESKTOP-LFM92VN
--------------------------------------------------------------------------------
Rosetta@home: Notice from server
rosetta python projects needs 19073.49MB more disk space. You currently have 0.00 MB available and it needs 19073.49 MB.
2/7/2022 7:55:34 AM DESKTOP-LFM92VN
--------------------------------------------------------------------------------
SiDock@home: Notice from server
CurieMarieDock on BOINC + zipped input, checkpoints and progress bar needs 128.00MB more disk space. You currently have 0.00 MB available and it needs 128.00 MB.
2/7/2022 10:14:32 AM DESKTOP-LFM92VN

Good grief..do I need 4 massive sticks of memory now? I just dropped a 150 on two big sticks. Plus 65 on a new 550GB drive (78% free) I think I might put a project limit of 12 tasks into play unless someone has a way of isolating Python with a code? I've put to much money into this computer now.
ID: 104700 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 . . . 171 · 172 · 173 · 174 · 175 · 176 · 177 . . . 300 · Next

Message boards : Number crunching : Problems and Technical Issues with Rosetta@home



©2024 University of Washington
https://www.bakerlab.org