Posts by Greg_BE

1) Message boards : Questions and problems : M.2 drives and GPU (Message 107006)
Posted 6 Feb 2022 by Greg_BE
Post:
ok...so...after going to SATA...I found the same problem when trying to store data out in the open.
So after deleting everything and starting over, I created a folder on the SATA drive and put the data in that and everything works.

That could have been the issue on the M.2 as well. But being I returned it already, hard to know. But if it works on SATA that way it must work on the M.2 that way.

As for the other thing about firmware..well no idea.
2) Message boards : Questions and problems : M.2 drives and GPU (Message 106992)
Posted 3 Feb 2022 by Greg_BE
Post:
Well I gave up on the M.2 and sent it back.
So this is a dead thread now for me.

Someone else over in RAH was considering a M.2 and he has the same board.
So after my issues, he's not to keen on getting a M.2

I have been in contact with Samsung EU in Paris via whatsapp and they would pass this problem and all the information I gave them to their back office testers.

It's strange how a single board could cause a specific brand of M.2 to not work.
But that being said.....it does work with Windows. That is not the problem.
The problem is that it does not work with BOINC at least not where the GPU's usage is concerned.
CPU is no problem.
So this is a very very specific problem.

But that is for someone else to complain and try to fix.
I'm done with it.

So don't worry about this anymore. It's finished and I'm outta here on this issue.
Thanks anyway neighbor.

Enjoy the sunny weekend.
3) Message boards : Questions and problems : M.2 drives and GPU (Message 106962)
Posted 2 Feb 2022 by Greg_BE
Post:
Magician not being able to flash the drive seems quite a big clue. But you're not the only one...

I'll check your log tomorrow as it's midnight for me as well, Nederland weet je?


Hi neighbor! I'm in Beveren.

As far as flashing goes, the store was able to flash on their setup.
I can't get it to do that on mine.

See your comments tonight after work.
4) Message boards : Questions and problems : M.2 drives and GPU (Message 106959)
Posted 1 Feb 2022 by Greg_BE
Post:
Any other ideas?
So far you have suggested everything I have already tried for trouble shooting and investigating.
Only the HWINFO stuff I had not looked at, but unfortunately I don't have time now to look deeply at it as it is midnight and I am up in 7.5 hours.
There is a link in one of the messages to a HTML output of what HWINFO found.
5) Message boards : Questions and problems : M.2 drives and GPU (Message 106958)
Posted 1 Feb 2022 by Greg_BE
Post:
Download the HWiNFO64 installer, install and run it, on the first screen uncheck "Sensors only", then click Run.
Next check in the Bus section what's all using PCIe lanes and how many.



https://drive.google.com/file/d/1G-8dVAMW-jPLmGqU66O_yX3W1RjBltkn/view?usp=sharing

This folder contains the log in HTML of all the HWINFO64 items.
Have a look and see if you can find what I am asking or anything else of interest.
It is midnight in Europe, so I have to get to bed.
6) Message boards : Questions and problems : M.2 drives and GPU (Message 106957)
Posted 1 Feb 2022 by Greg_BE
Post:
4) Why does BOINC work fine on SSD (SATA) and HDD (SATA) but not on M.2?
BOINC worked fine on both my Samsung SSDs, one NVMe M.2, one SATA.
How do you have everything connected? Your HDD and optical drive are connected to which SATA ports?

Furthermore, your motherboard details warn:
SATA1 port will be unavailable when installing SATA M.2 SSD in M2_2 slot.
PCI_E6 slot will be unavailable when installing PCIe M.2 SSD in M2_2 slot.


Did you install Samsung Magician and check if the drive needs a firmware update? Also check the drives for errors.


Not using SATA1 already been through that
Not using PCI 6 either. That's the last slot and its x4
I am using 1 and 3 which are x8 slots.

If memory serves correctly SATA 2,4,6 are in use (top row)

Unable to flash firmware. Samsung magician just stops the process.
The store was able to flash the first returned drive in their setup.
So not sure why they can and I can't.
Magician is the latest version.
7) Message boards : Questions and problems : M.2 drives and GPU (Message 106956)
Posted 1 Feb 2022 by Greg_BE
Post:
You would think there would be an error about the data if there was no more data to be seen on D:
Well, there is, BOINC would prompt you that there are no projects added, and make a new data directory. But before it can get there it has to go through its startup routine, among which is the detection of the GPU(s). And if it fails there, it'll just exit.



What I didn't tell you is that all my CPU projects ran just fine while this GPU problem was going on.
So BOINC had no issue with data. Just data for GPU's from a M.2
8) Message boards : Questions and problems : M.2 drives and GPU (Message 106953)
Posted 1 Feb 2022 by Greg_BE
Post:
I've been through the idea of used up PCIe lanes. But that does not make sense. A Ryzen 7 3700x has 24 lanes. The GPU's (1080 and 1050) use 8 lanes each. The CPU uses 4. That's 20. There should be 4 left.
SATA and USB (as well as Ethernet, Thunderbolt, U.2, ExpressCard and OcuLink) use PCIe lanes as well.

But, having said that, the GPU error you get may indicate you have a memory error on one of the GPUs or a problem with your video card driver.
I'd check the latter first, as that's easiest solved: just download the newest Nvidia driver and clean install it for both GPUs.

Else, check your GPU's memory: this site has 6 free programs to check your GPU memory (and it's written by HAL9000, I see).

Edit: after you moved your data directory to the M.2 SSD, did you tell BOINC (via the installer, or registry) that this is the new data directory?


1) GPU tests ok -that was never in doubt since FAH uses GPU's
2) Drivers are up to date
3) Question: If there are not enough PCI lanes left for the M.2 then why is it I can store and read/write data to it, but BOINC does not like it? And wouldn't there be some sort of windows error or other error by BOINC to indicate this?
4) Why does BOINC work fine on SSD (SATA) and HDD (SATA) but not on M.2?

GPU's according to the board mfg are 1 = 16 and 2 = 2 x 8, so again only 16.
SATA = 1 HDD 1SSD 1DVD ROM.
Ethernet is handled by onboard realtek PCIe ethernet controller.

Is there some program that can show how many PCIe lanes are in use to eliminate all this guess work about are they full or are there 4 free?
9) Message boards : Questions and problems : M.2 drives and GPU (Message 106951)
Posted 1 Feb 2022 by Greg_BE
Post:
Edit: after you moved your data directory to the M.2 SSD, did you tell BOINC (via the installer, or registry) that this is the new data directory?[/quote]

Clean install each time the drive was changed.
You would think there would be an error about the data if there was no more data to be seen on D:
10) Message boards : Questions and problems : M.2 drives and GPU (Message 106948)
Posted 1 Feb 2022 by Greg_BE
Post:
I tried to install a Samsung 980 M.2 drive and store the BOINC data there because RAH Python tasks in multiples are eating up all the remaining space on my Samsung SSD 250GB C: drive.

The M.2 is seen by windows ok, partions ok, stores data ok.
But when I boot up BOINC then I get a series of errors.

It sees data directory D: (the m.2)
goes into running under account Greg
Then the error GPU detection failed, process exited with status 0x1: Incorrect function (0x 1)
No usable GPU's found
app version refers to missing GPU type NVIDIA
Application uses missing NVIDIA GPU

I tried both m.2 slots, always this error.
If I move the data to HDD or SSD, then it is happy.
But if the data is on the M.2 , forget it.

I've been through the idea of used up PCIe lanes. But that does not make sense. A Ryzen 7 3700x has 24 lanes. The GPU's (1080 and 1050) use 8 lanes each. The CPU uses 4. That's 20. There should be 4 left.

Now the wild thing is, the GPU's are ok as long as the data is on SSD or HDD. FAH being a windows program files project runs off the SSD so again no problem. This eliminates the full PCIe lanes theory.

So does BOINC have an issue with M.2 drives being used as data drives?
Nothing else seems to. Windows OS sees it. I can partion it, I can store data on it. Just BOINC does not like it. But why?

I swapped out my first M.2 for another M.2 and the same problem came up. Again HDD and SSD work fine for data.

I can not find anything that says BOINC hates M.2's but it appears to be that way. I would like to know what you all know. Can you help?

Configuration:

MSI X470 gaming plus
AMD Ryzen 7 3700x
GTX 1050 and GTX 1080
1x 250GB SATA SSD
1x 1TB HDD

Returned M.2 was a Samsung 980 500GB drive.
I don't have that anymore since my refund time runs out tomorrow.
11) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106059)
Posted 10 Nov 2021 by Greg_BE
Post:
No more answers..ok...then it is what it is.
Not going to monitor this thread anymore.
Thanks for the ideas though, to bad they didn't work.
And Jim1348, your post was the least useful.
Thanks Joseph Stateson for you ideas.
Richard Haselgrove your answer was interesting, thanks.
12) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106035)
Posted 8 Nov 2021 by Greg_BE
Post:

1. Create a venue over at RAH: "School" and select only Python <-- not possible. RAH does not have that option. If it did I would have tried that already. You can not isolate Python from 4.2 tasks. I already tried to exclude 4.2. BOINC just coughed up errors. I can do the rest however. New profile and suspend everything else.


Bummer : I was not aware of that. Just went over there and read the following quote

Rosetta@home does not currently divide its tasks into sub-projects that users can select

Obviously there is injustice in having some projects suffer because only a few users wants to contribute to them Clearly this project and its administrators are fully Woke to this injustice!

That means my way or the highway and I took the highway long ago.

The following might not work in R@H project directory (app_config.xml) but you could try it.

<app_config>
<app>
<name>Woke</name>
<max_concurrent>1</max_concurrent>
</app>
</app_config>


There will be an error message "No Woke projects" and POSSIBLY there will be a list of sub-project names that can be used to exclude.

At least this works on other projects. They print out the names that can be used. I they spell out something like "RH4_2" then put that in place of "Woke" and only one of those 4.2 will be run. Hopefully you will get a viperload of Pythons.

HTH



No dice, same as before: Your app_config.xml file refers to an unknown application 'Woke'. Known applications: 'rosetta'

It just knows rosetta and I guess python falls under "rosetta" or something.
Stopping with Sidock.
Maybe the only solution is to set all CPU projects to "no new" tasks and introduce each project back one at a time. But somehow I don't think that will change anything either.
Setting work load to 1 full day, no reserve.
13) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106034)
Posted 8 Nov 2021 by Greg_BE
Post:
When Python first came out, I got about 10 tasks.
Then all of sudden they no longer came to my system,
Nothing has changed in the setup. I have current Vbox since I run LHC ATLAS and never had any problems with Vbox in the past.

Re-posting the same issue on every forum you can find probably won't fix it.
More likely, freeing up memory and trying the other suggestions will. The BOINC scheduler is also rather particular, and won't download some work if you tie it up on other projects.
I would think you should have found that out already though.



FYI only 2 forums. Here and RAH. Since they guys at RAH are not expert enough I came here to look for help and not to be insulted with one liners.
I will set SiDock to no new tasks and let it run out and see what changes.

What I am trying to get at is what if any error messages I should be looking for, what physical things could be preventing it, what software if any is acting up. I don't need wild speculation.
14) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106031)
Posted 8 Nov 2021 by Greg_BE
Post:
RAH could easily stop sending 4.2 tasks and put Python in and there would be enough memory for a mix of Python and 4.2. But I get 0 Python these days


RAH is native to Windows. Python runs in Vbox only. The project cannot just swap a Python for an RAH like you suggest. Vbox has a say in what it wants and allows. It is not allowing any Python tasks and only it knows why?

Taking a guess: Can you log onto the guest OS and check resources? I assume it it linux or whatever or (got help you) Max OS. I can not help with any of those OS'es


What you should do if you want to sell car, house, spouse, and everything you own for the love of PYTHON

1. Create a venue over at RAH: "School" and select only Python
2. Set your system to use the "School" venue
3. Suspend all other BOINC and FAH tasks
4. Reboot
5 Wait (don't hold your breath) for some PYTHON to start up

Once you got a bunch of PYTHON running, then resume all other tasks and start looking for error messages and / or tasks that are refuse to run. If FAH no longer wanting to run go ask over at FAH what is going on.



1. Create a venue over at RAH: "School" and select only Python <-- not possible. RAH does not have that option. If it did I would have tried that already. You can not isolate Python from 4.2 tasks. I already tried to exclude 4.2. BOINC just coughed up errors. I can do the rest however. New profile and suspend everything else.


Made a school profile, changed to school profile, update request sent to RAH, no change.
Still wants to download 4.2 tasks even though everything but GPU tasks (no FAH) are running.
No LHC even, so I don't know anymore.
I even reset the project, no change.

Vbox is windows flavor...as to how to log into a guest OS, now your talking French to me.
I have heard stories that newer Vbox versions can cause issues, but then if I revert to something older then it is possible something in Sidock certificates would cough up an error. Crazy

I am really puzzled by all this maybe it will maybe it wont, maybe if I kick it things would work.
15) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106027)
Posted 8 Nov 2021 by Greg_BE
Post:
RAH could easily stop sending 4.2 tasks and put Python in and there would be enough memory for a mix of Python and 4.2. But I get 0 Python these days


RAH is native to Windows. Python runs in Vbox only. The project cannot just swap a Python for an RAH like you suggest. Vbox has a say in what it wants and allows. It is not allowing any Python tasks and only it knows why?

Taking a guess: Can you log onto the guest OS and check resources? I assume it it linux or whatever or (got help you) Max OS. I can not help with any of those OS'es


What you should do if you want to sell car, house, spouse, and everything you own for the love of PYTHON

1. Create a venue over at RAH: "School" and select only Python
2. Set your system to use the "School" venue
3. Suspend all other BOINC and FAH tasks
4. Reboot
5 Wait (don't hold your breath) for some PYTHON to start up

Once you got a bunch of PYTHON running, then resume all other tasks and start looking for error messages and / or tasks that are refuse to run. If FAH no longer wanting to run go ask over at FAH what is going on.



1. Create a venue over at RAH: "School" and select only Python <-- not possible. RAH does not have that option. If it did I would have tried that already. You can not isolate Python from 4.2 tasks. I already tried to exclude 4.2. BOINC just coughed up errors. I can do the rest however. New profile and suspend everything else.
16) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106026)
Posted 8 Nov 2021 by Greg_BE
Post:
I don't think fah is a boinc app so it does not know about its resource needs correct?
Correct. FAH and BOINC studiously ignore each other's existence. If both are running on the same machine, each will assume that it has exclusive use of the resources found. Everything will be double-counted, and, roughly speaking, everything will drop back to 50% efficiency or worse.

Each platform would need to be manually restrained to work within whatever proportion of the machine resources you are willing to assign it.


Again...FAH is GPU only. GPU BOINC is Einstein.
1 or 2 cores go to support the 2 GPU's I have. I have 16 cores.
That should not affect CPU resources or RAM resources.
RAM as my figures have indicated has a peak load of 93% and a average of 73.7%
That is stats from HWINFO
So if according to that program, my system is not overloaded and both programs work fine, then why is RAH holding out on Vbox while LHC ATLAS pushes them out no problem?
And again, I HAD in the past when Python was first coming out, some tasks from that group of tasks, then all of a sudden they don't come anymore.
That is the question.
17) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106024)
Posted 8 Nov 2021 by Greg_BE
Post:
Asked at Rosetta, to complex for the few guys there that can help, so referred here.


Memory is set at 100% all the time
Non GPU is selected
page swap is 100%
Vbox is using 25% CPU and 8.9-9.1 MB
FAH is using 12.4% CPU combined (tasks) and 1.9 GB RAM Controller uses 0-0.6% and 19.7MB
Firefox uses .3% CPU and 556.5 MB
Edge (FB) uses 3.3% max and 348.7 MB
RAH at the moment is running 9 tasks with 3 on hold using +/- 3.38 GB (the running tasks) the three on hold use 588 MB

My total physical memory commitment from everything is 75% at the moment.
Vbox with ATLAS is a combined 6609 MB (ATLAS task is 6600 and Vbox 9 and just now found Vbox wrapper 1,6 MB.

HWINFO says virtual is this:
current use 29,345, Min 29,243, Max 36,3684 average 35.163
available 7465, min. 5636, max. 12,646, avg 6485
Load 79.7,69.9,86.5,84.3%

physical used is: 10,956, 10,827,22,770,18,078
Available is: 13,545,1,731,13,673,6,422
Load is: 44.7,44.1,92.9,73.7%

page file 17.3,8.1,17.3,11.7%

To me it does not seem to be a memory issue.
RAH could easily stop sending 4.2 tasks and put Python in and there would be enough memory for a mix of Python and 4.2. But I get 0 Python these days.

I have unchecked leave in memory before I came here, that did not make a difference either. Still nothing.

Vbox runs in the background and never gives me any problems unless for some reason the environment gets gummed up and I need to go manually clean it. Then I restart BOINC and everything is ok again.
18) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106018)
Posted 8 Nov 2021 by Greg_BE
Post:
Which memory size do you need to know?
Virtual or Working?
On some tasks there is a huge difference between the two.

I am at this moment only using 63% of the memory on average (Windows task manager)
According to BOINC Tasks program, the memory values as of this moment are as follows:

Einstein 161.92 and 161.04 MB (GPU tasks)
WCG Cancer 81.41 and 75.55 (and climbing very slightly each 2 second update .04 %)
SiDock 140.08, 151.63, 150.24
*RAH degrader - 545.75,483.52,344.19,354.33* - Now 30 minutes later they have grown to 1.10 to 1.26 GB
LHC ATLAS - (4 cores) fixed at 6600 MB, never changes according to the experts on that forum.
Task manager says BOINC is using 1.8xx GB of memory
Firefox with 9 tabs is 857 MB
FAH is using 1.95 GB and 364.2 MB for tasks and the client is using 31.5 MB
System processes use the rest of that percentage.

So it is 13 cores in use for CPU and I would guess 2 cores monitoring the GPU and that leaves me 1 with system duties.

For logging:
Which flags should I have checked in diagnostics to see if there is a problem?
I have task,app_msg_send, cpu_sched_debug,rr_simulation,sched_op_debug,statefile_debug,work_fetch_debug and sched_ops enabled, but to me the messages look all routine.

As for RAH tasks that are active, no more than 4 or 5 are active at any given time. The rest are in queue.
19) Message boards : Questions and problems : Can not get Rosetta Python (Vbox) tasks (Message 106008)
Posted 7 Nov 2021 by Greg_BE
Post:
When Python first came out, I got about 10 tasks.
Then all of sudden they no longer came to my system,
Nothing has changed in the setup. I have current Vbox since I run LHC ATLAS and never had any problems with Vbox in the past.

I give 15/16 cores to BOINC to run all my projects. RAH,LHC ATLAS,WCG,Einstein,SiDock. 1 core is used to run my 2 GPU's (1080 and 1050) and handle system work (mainly web browsing and some videos).

So why can't I get any Python tasks?
I just keep getting Rosetta 4.2.

System specs:
AuthenticAMD
AMD Ryzen 7 3700X 8-Core Processor [Family 23 Model 113 Stepping 0]
(16 processors) [2] NVIDIA NVIDIA GeForce GTX 1080 (4095MB) driver: 472.12 OpenCL: 3.0 Microsoft Windows 10
Professional x64 Edition, (10.00.19043.00)
Memory is 24GB
Vbox 6.1.28 r147628 with current extension pack.

Currently only using 52% of the memory with 5 WCG, 4 SiDock, LHC (4 cores always) which is only 13/15 cores available. Also run GPU only FAH and Einstein GPU.

BOINC is set for 6 hour runs before switching projects.
RAH is set for 6 hour run times.
Cache is 1 full day of work plus .15 days of extra
I run 16 hours a day +/- an hour.

So any ideas why Python Vbox tasks are not coming through?
20) Message boards : Questions and problems : Unable to get Vbox 5.2.16 to work (Message 87588)
Posted 12 Aug 2018 by Greg_BE
Post:
I downloaded BOINC with its old 5.2.8 Vbox and then upgraded to 5.2.16, but now I am unable to get it to connect even though BOINC can "see" it.

So now I am unable to run any VM tasks.

Virtualization is enabled in BIOS. Double and tripled and quadrupled checked that. So there is something software wrong with BOINC and Vbox. But what?




Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.