Hey Everyone,

My primary SETI rig has an NVIDIA GeForce GTX 650 Ti as well as an AMD Radeon HD 7700 series.

My AMD card produces 99% valid results, even though it is a little slower. My GTX 650 unfortunately only yields about 85% valid results, and on a bad day can really churn out a lot of invalid results. (Right now, there are 26 invalids.)

Over time, I have posted several times looking for advice on how to reduce the invalid results from my NVIDIA card, but not much seems to help. I have two GTX 650 cards, and they both don't do so well, which leads me to believe that they are just not a perfect match for my system.

Onto my point; the GTX 650 card DOES however produce 99% valid AP work units. Inside of the SETI preferences, you can tell a BOINC whether or not to use NVIDIA on a particular machine. I would like to use my NVIDIA card, however I would like to ONLY do AP work units with this card.

If not, I am unfortunately considering doing Einstein with my NVIDIA card, since during the SETI break I was producing valid NVIDIA Einstein results as well. (What the heck is it about MB that my NVIDIA dislikes? :)

Is it possible to only run AP tasks with the NVIDIA card? Would it be as simple as not selecting MB for NVIDIA during the lunatics installer?

Thanks to anyone who may have some insight.

mherr170 i have 2 gtx 650 cards running and they do not produce the problems you are getting and are in the same machine .
With a problem like this i would need more info to figure out why it's doing it .

how many units per GPU are you running ?
how many units are you running on the CPU ?
have you upgraded to latest Bionic ?
have updated Luna to 0.043a ?
what Nvida driver are you using ?

You can limit what GPU's use what app if you use a config file but others will have to help you on that one

And lastly if you have overclock the GTX cards return them back to stock as there should already be overclocked at the factory and don't like being pushed any faster .

---

Hey Everyone,

My primary SETI rig has an NVIDIA GeForce GTX 650 Ti as well as an AMD Radeon HD 7700 series.

My AMD card produces 99% valid results, even though it is a little slower. My GTX 650 unfortunately only yields about 85% valid results, and on a bad day can really churn out a lot of invalid results. (Right now, there are 26 invalids.)

Over time, I have posted several times looking for advice on how to reduce the invalid results from my NVIDIA card, but not much seems to help. I have two GTX 650 cards, and they both don't do so well, which leads me to believe that they are just not a perfect match for my system.

Onto my point; the GTX 650 card DOES however produce 99% valid AP work units. Inside of the SETI preferences, you can tell a BOINC whether or not to use NVIDIA on a particular machine. I would like to use my NVIDIA card, however I would like to ONLY do AP work units with this card.

If not, I am unfortunately considering doing Einstein with my NVIDIA card, since during the SETI break I was producing valid NVIDIA Einstein results as well. (What the heck is it about MB that my NVIDIA dislikes? :)

Is it possible to only run AP tasks with the NVIDIA card? Would it be as simple as not selecting MB for NVIDIA during the lunatics installer?

Thanks to anyone who may have some insight.

Yes it is exactly that simple if that is what you would like to do. There is a way to do it in the cc_config.xml using <exclude_gpu>, but if you are not going to run an app. Why have it on the system(s)?

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Hey Glenn,

Thanks for responding. I have borderline given up on trying to solve my Nvidia MB issues, but I'll answer your questions. Maybe you have some hints.

I am running 2 MB per 650ti (only have 1 installed currently)
0 on the CPU
Boinc is recent, but not very latest (The issue occured before and after the recent upgrade anyway)
Installed Lunatics 43a
and updated to driver 344.75

I have to assume that the answer to all of my questions fall right in line with a working GPU, but my machine doesn't play 100% nicely with them.

Yes it is exactly that simple if that is what you would like to do. There is a way to do it in the cc_config.xml using <exclude_gpu>, but if you are not going to run an app. Why have it on the system(s)?

Thanks HAL. Would it be wise to uninstall and reinstall Lunatics if I am looking to remove the MB app for NVIDIA? Maybe an obvious question.

I'm running Seti MB GTX 650ti with Lunatics 0.43a and no problems.

Here's an example of cc_config that I use if it is of help.

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
<options>
<exclude_gpu>
<url>http://setiathome.berkeley.edu/</url>
<device_num>0</device_num>
<type>NVIDIA</type>
<app>astropulse_v7</app>
</exclude_gpu>
</options>
<options>
<exclude_gpu>
<url>http://setiathome.berkeley.edu/</url>
<device_num>1</device_num>
<type>NVIDIA</type>
<app>astropulse_v7</app>
</exclude_gpu>
</options>
</cc_config>

---

mherr170 ahh upgrade to the latest version of Bionic i had to upgrade it myself .

If you have try'd a CLEAN install of the Nvida driver it won't be that .
You are running win 8.1 ! maybe there is a prob with that but i'm shore other will let you know .

it not the igpu as your machine is a quaddy , you have 3 gig of ram thou and your running win 8.1 ?

win 7 is a biautch if you only got 2 gig of ram and run some of the projects but not shore about win 8 +

Do you have anything running in the background , or are you using anther app while crunching ?

otherwise i can only think off it's running ATI and Nvidia cards and somehow it's causing a prob !

---

...(What the heck is it about MB that my NVIDIA dislikes? :)
...

As the MB Cuda application developer, I would like to know this as well. Looking at the mentioned invalids, I see spontaneous overflow results that the wingmen disagrreed with (at a higher rate than might be expected by some faulty app being in circulation etc).

As you mention Einstein and AP both seem fine, I would first compare loadings (single instance) (GPU usage in e.g. evga precision X or similar), and temperatures. Since we know that different applications push the devices differently, there can be more pressure on VRAM, BUS, GPU itself, temperatures. You're looking for anything that stands out as particularly different in the MB application compared to the others.

Generally speaking, Cuda MB is relatively high reliability due to longer maturity. Single instance, compute efficiency on new apps like AP and Einstein is a bit lower, and therefore MB will push the envelope with multiple instances and a marginal factory voltage/overclock. With temperature cocnerns eliminated/controlled, One often effective means to correct marginal factory overclocks is a slight boost to core voltage (temperatures & reasonable safety permitting).

In the future, I'll be adding a throttle to Cuda MB, since going flat chat will never suit every system/user.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

Yes it is exactly that simple if that is what you would like to do. There is a way to do it in the cc_config.xml using <exclude_gpu>, but if you are not going to run an app. Why have it on the system(s)?

Thanks HAL. Would it be wise to uninstall and reinstall Lunatics if I am looking to remove the MB app for NVIDIA? Maybe an obvious question.

I think you can just rerun it & select the new app combination you like. Then the app_info.xml will be overwritten by a new one.
It has been a while since I played around with the installer. As I normally just extract the apps and do things manually these days.

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

...(What the heck is it about MB that my NVIDIA dislikes? :)
...

As the MB Cuda application developer, I would like to know this as well. Looking at the mentioned invalids, I see spontaneous overflow results that the wingmen disagrreed with (at a higher rate than might be expected by some faulty app being in circulation etc).

As you mention Einstein and AP both seem fine, I would first compare loadings (single instance) (GPU usage in e.g. evga precision X or similar), and temperatures. Since we know that different applications push the devices differently, there can be more pressure on VRAM, BUS, GPU itself, temperatures. You're looking for anything that stands out as particularly different in the MB application compared to the others.

Generally speaking, Cuda MB is relatively high reliability due to longer maturity. Single instance, compute efficiency on new apps like AP and Einstein is a bit lower, and therefore MB will push the envelope with multiple instances and a marginal factory voltage/overclock. With temperature cocnerns eliminated/controlled, One often effective means to correct marginal factory overclocks is a slight boost to core voltage (temperatures & reasonable safety permitting).

In the future, I'll be adding a throttle to Cuda MB, since going flat chat will never suit every system/user.

Thanks very much for the response! Out of all the folks here who run NVIDIA 650s or similar, I really seem to be the only person having issues. My computer is a frankenstein of mixed parts, old parts, and started out as an experiment to see if I could crunch SETI at all, let alone do it efficiently. So the fact that some things don't work exactly the way I'd like doesn't surprise me.

My two graphics cards sit next to each other, so heat has always been a question to me. However, running various monitoring programs, I've never seen it reach 70C. It usually hovers in the mid 60's.

I did try to mess with the overclocking initially, but I really didn't know what I was doing, so that effort fell a bit to the way side out of fear of breaking something.

As far as the single instance loadings, I'll have to read up on exactly what this entails. I recently uninstalled a program that I used months ago to attempt to adjust voltages on the card, so I wonder if that is enough to set it back to 'factory' so to speak. I may try cuda MB one more time to see if it made any difference.

I'm running Seti MB GTX 650ti with Lunatics 0.43a and no problems.

Here's an example of cc_config that I use if it is of help.

<cc_config>
<options>
<use_all_gpus>1</use_all_gpus>
</options>
<options>
<exclude_gpu>
<url>http://setiathome.berkeley.edu/</url>
<device_num>0</device_num>
<type>NVIDIA</type>
<app>astropulse_v7</app>
</exclude_gpu>
</options>
<options>
<exclude_gpu>
<url>http://setiathome.berkeley.edu/</url>
<device_num>1</device_num>
<type>NVIDIA</type>
<app>astropulse_v7</app>
</exclude_gpu>
</options>
</cc_config>

Thank you very much! I will compare to mine to see if there is something strange.

Yes it is exactly that simple if that is what you would like to do. There is a way to do it in the cc_config.xml using <exclude_gpu>, but if you are not going to run an app. Why have it on the system(s)?

Thanks HAL. Would it be wise to uninstall and reinstall Lunatics if I am looking to remove the MB app for NVIDIA? Maybe an obvious question.

I think you can just rerun it & select the new app combination you like. Then the app_info.xml will be overwritten by a new one.
It has been a while since I played around with the installer. As I normally just extract the apps and do things manually these days.

Thanks HAL! I think I'll run a few more tests with MB, but if they fail good to know that it is pretty easy to get NVIDIA to only run AP>