Thanks. The driver on that host is 436.30 not 436.48 like on the Windows 10 PC which is not enlisted in Science United.
Tullio

You’ll have to clarify what host you’re actually talking about. That is the only host under your account that has a GPU. The host I linked to was last running the 436.48 driver according to the host details page.

But that host also hasn’t contacted the SETI servers for over 36hrs now.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

That is the only host under your account that has a GPU.

He's made it clear that he's talking about a host on a Science United account, not his personal account. You're talking at cross purposes.

That is the only host under your account that has a GPU.

He's made it clear that he's talking about a host on a Science United account, not his personal account. You're talking at cross purposes.

Not sure what Science United is. But if he’s doing work for SETI@home, there should be a host details page and associated task list for that host.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

The Window 10 host, not linked to Science United, has a GTX 1060 GPU. The Windows 8.1, enlisted in Science United, has a 1050 Ti GPU. The same host has a guest SuSE Linux Tumbleweed with kernel 5.3.5-1 via Virtual Box, also linked to Science United, but this OS is unable to see the GPU, so it cannot run GPU tasks. CPU are Ryzen 5 1400 on the Windows 10 host, A10-6700 on the Wiindows 8.1. RAMs are 24 GB on the Windows 10 PC and 20 GB on the 8.1 PC. They are connected by WiFi.
Tullio
Science United was created by Dave Anderson and I am getting a mail from him every week stating how many tasks I have completed and the hours it has taken. I am running mostly Milkyway@home and Asteroids@home tasks both CPU and GPU, but also SETI@home. My ScienceUnited home page says I have completed 426 SETI tasks and zero failures.

---

Science United will have sent you a perfectly normal BOINC client and Manager. If you:

• Open the Manager
  
• Switch to 'Advanced' view (if you aren't using that already)
  
• Click on the 'Projects' tab
  
• Click on the 'SETI@Home' project line
  
• Click on the 'Properties' button
  

You'll get a page of information including a Computer ID, valid for this project only. Please post that number here.

I am an old SETI@home user. On April 28 I enlisted in ScienceUnited stating that my interests were astronomy, physics and chemistry. I started getting a number of tasks on one of my PCs, a Windows 8.1 and went on crunching all I got. I have three PCs listed in Science United but two are Virtual Box Suse Tumbleweed, which are updated very frequently It is only a single Virtual Computer but every time I have a kernel update Science United thinks I have a new computer. I do not want to enlist my fastest computer, the one with Windows 10 and Ryzen 5 1400 since I want to decide myself what I want to do. It is mostly running LHC@home tasks, also using Virtual Box, and I got a letter of thanks from CERN and a polo shirt with the LHC tag. On Science United I do not see my credits, since I am an anonymous user and my RAC is the sum of all ScienceUnited user RACS. This might not by liked by users who want to be credit guzzlers, but this is not my case. When I started a SETI session on my fastest computer I saw a number of tasks, coming from Arecibo, using very little GPU. I read the message boards and saw this was related to the nVidia driver 436.48, which works perfectly in Einstein and GPUGRID and on SETI tasks coming from Green Bank. So I aborted them. That is all.
Tullio

---

That’s all fine and good, but doesn’t really answer the question asked.

Follow the steps that Richard posted and post the Host ID of the system in question. That way we can see the progress with the 436 drivers.

But if you’re just aborting all Arecibo work to avoid the problem, then it’s really not an indication that the problem is solved with the 436.48 drivers.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

I am now running SETI on my Windows 10 PC, not enlisted in Science United. The driver is 436.48 and they all run well. On the Windows 8.1 PC, enlisted in Science United, the driver is 436.30. I cannot get any data on my membership in ScienceUnited, since I have no password for every project, only a general password for all Science United.
Tullio
All my GPU tasks on the ScienceUnited PC were failing this morning. Since I had a major Windows 8.1 update, I suspected my nVidia driver 436.30 was overwritten and I reinstalled it. All run well now and I am not updating it to 436.48 as suggested by nVidia.
It is now running Amicable numbers on 4 CPUs and SETI@home on GPU.

---

I am now running SETI on my Windows 10 PC, not enlisted in Science United. The driver is 436.48 and they all run well. On the Windows 8.1 PC, enlisted in Science United, the driver is 436.30. I cannot get any data on my membership in ScienceUnited, since I have no password for every project, only a general password for all Science United.
Tullio

Yes they will until we get another dose of Arecibo VHAR work. ;-)

Cheers.

I am now running SETI on my Windows 10 PC, not enlisted in Science United. The driver is 436.48 and they all run well. On the Windows 8.1 PC, enlisted in Science United, the driver is 436.30. I cannot get any data on my membership in ScienceUnited, since I have no password for every project, only a general password for all Science United.
Tullio
All my GPU tasks on the ScienceUnited PC were failing this morning. Since I had a major Windows 8.1 update, I suspected my nVidia driver 436.30 was overwritten and I reinstalled it. All run well now and I am not updating it to 436.48 as suggested by nVidia.
It is now running Amicable numbers on 4 CPUs and SETI@home on GPU.

You don’t need any password to get the host id. Richard explained exactly how to do that. You just have to navigate around the existing BOINC window to open it up.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

OK. I am getting lost on Windows and its foibles.
The 8.1 PC, connected to Science United, has ID 8815395
The 10.0 PC, not connected to Science United, has ID 8609032
Tullio

---

OK. I am getting lost on Windows and its foibles.
The 8.1 PC, connected to Science United, has ID 8815395
The 10.0 PC, not connected to Science United, has ID 8609032
Tullio

When I click on your name and then Computers and Select ID 8609032, then select "Errors" the top of page reads;

Error tasks for computer 8609032

Next 20
State: All (205) · In progress (92) · Validation pending (75) · Validation inconclusive (1) · Valid (15) · Invalid (0) · Error (22)
Application: All (205) · AstroPulse v7 (0) · SETI@home v8 (205)

The first task (Workunit 3695652996) has these details;

<core_client_version>7.14.2</core_client_version>
<![CDATA[
<message>
exceeded elapsed time limit 4441.77 (1414546.15G/318.46G)</message>
<stderr_txt>

and lower

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 2.564348

That is what you get with Aricebo VHAR tasks using the latest drivers.

And for host 8815395:

State: All (189) · In progress (30) · Validation pending (32) · Validation inconclusive (3) · Valid (19) · Invalid (0) · Error (105)

The errors seem to be mostly different, like result 8138751749:

<message>
too many boinc_temporary_exit()s</message>

and

WARNING: BOINC failed to provide OpenCL device, using own enumeration abilities
ERROR: OpenCL kernel/call 'clGetDeviceIDs (second call)' call failed (-32) in file ..\..\..\src\GPU_lock.cpp near line 1310.
Waiting 30 sec before restart...

That sounds like a computer with an incomplete driver (missing the OpenCL component).

OpenCL is often missing in the versions of the driver supplied by Microsoft for Windows 10: but this is your Windows 8.1 computer, and OpenCL is recorded on the Computer information page.

In either case, the cure is to install a complete driver from the NVidia website. Choose a driver from the 431 range, and choose the custom install option to remove all previous drivers.

Just because a driver is compatible with one application (Einstein), you can't say that it is 'working perfectly' in all environments. The SETI applications may utilise code pathways that are not used in the Einstein code, and encounter bugs there.

yes, he did mention that he had an update which errored out all of his tasks. he suspected a driver update via Windows, which has been known to cause issues like that.

the important thing is that I think it's clear that the issue is not fixed in the 436 drivers yet. aborting Arecibo work is just a band aid to work around the issue.

I can't understand why someone wouldn't just revert to the known good 431 drivers.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

I can't understand why someone wouldn't just revert to the known good 431 drivers.

Maybe because the 431 is the "Studio Driver" and the 436 are "Game Ready Driver" and the gamer like to keep their driver updated.
The question could be: Do you run a game or a GPU who rely need the 436 driver?
IMHO...
if the answer is NO and wish to keep crunching (SETI at least) without any issue, stay on the 431 until Nvidia fix the issue.
if the answer is YES you have 2 options: stop crunching (SETI at least) until Nvidia fix the issue or do like he was doing, abort all Arecibo WU received. That is radical, yes, but what else the user could do?
One point is clear, keep crunching High VAR (>2.??) Arecibo WU with the available 436 series driver is a waste of time & resources.

---

One point is clear, keep crunching High VAR (>2.??) Arecibo WU with the available 436 series driver is a waste of time & resources.

And for simplicity: it's easy to recognise Arecibo VHAR tasks by their short deadlines, three weeks from issue. Today's downloads of that type have just flipped over to 6 November (UTC) deadlines.

I can't understand why someone wouldn't just revert to the known good 431 drivers.

Maybe because the 431 is the "Studio Driver" and the 436 are "Game Ready Driver" and the gamer like to keep their driver updated.
The question could be: Do you run a game or a GPU who rely need the 436 driver?
IMHO...
if the answer is NO and wish to keep crunching (SETI at least) without any issue, stay on the 431 until Nvidia fix the issue.
if the answer is YES you have 2 options: stop crunching (SETI at least) until Nvidia fix the issue or do like he was doing, abort all Arecibo WU received. That is radical, yes, but what else the user could do?
One point is clear, keep crunching High VAR (>2.??) Arecibo WU with the available 436 series driver is a waste of time & resources.

431.60 is a Game Ready Driver and includes the recent security fixes.

I doubt anyone here NEEDS the 436 drivers.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

I am not a gamer. I solved the problem on the Windows 8.1 PC by reinstalling the 436.30 driver which seems to work. But all my other projects use the 436.48 driver. I am running 6 projects, plus Science United.
Tullio

---

Windows 7 & 8.x don't have a problem with the 436.x driver. If you would just reinstall the 431.x driver in Windows 10, then you wouldn't have any problems at all.

I am not a gamer. I solved the problem on the Windows 8.1 PC by reinstalling the 436.30 driver which seems to work. But all my other projects use the 436.48 driver. I am running 6 projects, plus Science United.
Tullio

what we keep telling you is that you did not "solve" the original problem. the 436 drivers still have problems with Arecibo HVAR tasks. you only rectified the issue of windows changing the driver on you from an update, which has been known to cause issues, regardless of driver version.

you can truly solve this problem by reverting to the 431 series drivers (whichever you wish) and you will then be able to process all task types without aborting them.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours