CUDA app version changed to 1.07 for both Windows x86 and x86_64.
I have removed the plan_classes "cuda" and "cuda23" because the app required compute capability 2.1+ cards & I haven't seen even a single task completed with success on "cuda23" plan_class.

The new version was recompiled to allow compute capability 2.0 cards:
GeForce GTX 590, GeForce GTX 580, GeForce GTX 570, GeForce GTX 480, GeForce GTX 470, GeForce GTX 465, GeForce GTX 480M (source: https://en.wikipedia.org/wiki/CUDA)
so in theory anything from Fermi up should work.

The CUDA app requires Visual Studio 2015 c++ redistributable package installed to run.

---

M4 Project homepage
M4 Project wiki

Thanks for trying but the new v1.07 is still not working with my 580's. It's the same error as before.

Exit status 195 (0x000000C3) EXIT_CHILD_FAILED

http://www.enigmaathome.net/result.php?resultid=278292779


PS: Is there an update for when a fix will be available for running the app on more than just GPU-0 ? (in multi GPU systems)

Thanks

Thanks for trying but the new v1.07 is still not working with my 580's. It's the same error as before.

Exit status 195 (0x000000C3) EXIT_CHILD_FAILED

http://www.enigmaathome.net/result.php?resultid=278292779

Are the required Visual Studio 2015 c++ redistributable packages nstalled on your system?

Yes it's installed on all my systems. The Enigma GPU app works fine on all my PC's with the only excepting being the PC with the GTX 580's in it. However the 580 based system works fine at all other boinc GPU projects including the new xansons cuda app.

I be interested in finding out if any fermi based GPU's has ever produced valids results so far?

Here is a full list of GPUs that returned valid workunits. It's updated in realtime.
http://www.enigmaathome.net/gpu_os.php

Your issue may or may not be related to the card itself.
There are others reporting that the app will not start on their setup, producing the same error and their GPUs are known to work. The reason for that is unknown at this moment.

---

M4 Project homepage
M4 Project wiki

Thanks TJM, the only v2.0 card that appears on that list is the 560. The 550 ti & 560 ti are both v2.1 cards.

Just out of interest is it possible to add to that list how many of each card is producing those numbers? Is that just one 560 producing those numbers or is there a dozens?

Also I wonder if there's a correlation in the driver versions too? Although i'm not sure if it would be possible add that to your list?

Cheers


PS: Any update on a fix for the multi GPU problem?

Thanks TJM, the only v2.0 card that appears on that list is the 560. The 550 ti & 560 ti are both v2.1 cards.

Long time ago I had 560 and AFAIR it was compute capability 2.1
Just did a quick google search and first result seems to confirm that: http://www.geeks3d.com/20100606/gpu-computing-nvidia-cuda-compute-capability-comparative-table/
So most probably the app does not work on compute capability 2.0 cards.

Just out of interest is it possible to add to that list how many of each card is producing those numbers?

In theory that's possible and the project used to do that (for CPUs), but over time the table started getting too large for realtime processing and eventually that feature was removed.

Is that just one 560 producing those numbers or is there a dozens?

There are at least 3 of them that returned results, but as mentioned above I think they're compute capability 2.1 anyway.

---

M4 Project homepage
M4 Project wiki

Not to worry thank you for trying anyway.

PS: Any update on a fix for the multi GPU problem?

http://www.enigmaathome.net/forum_thread.php?id=836&postid=5852#5852 + a couple more following posts.

---

M4 Project homepage
M4 Project wiki

Hi,
I'm having the same EXIT_CHILD_FAILED issue with a GT 630m GPU (Compute Capability 2.1) in Version 1.10 of the cuda_fermi app.
So far I've updated the driver and reinstalled Visual Studio 2015 (both 32-bit and 64-bit versions) which allowed the app to run for more than a second, but now the app is causing driver crashes which is causing the same errors further down the line. Still not had a single successful WU on this card, and looking at the chart you posted above, nor has anybody else yet.
Just wondered if there's any settings I may be able to change to stop these driver crashes, or whether anything can be changed your end?
Perhaps even a few shorter WUs/different app for the "m" variants of Nvidia's GPUs, as I've heard before that there's some minor differences in some of the drivers for these laptop GPUs as opposed to the desktop variants?

I'm crushing for Enigma from years, and this beloved time seems to be over.
I have a laptop vith a Cuda 2.0 NVIDIA GPU (GT 460M), and when executing an Enigma GPU task, my screen is flashing up and down a few times, I get a message telling the Driver has been restored, and the computer crashes.
This should be quite interresting at Christmas time if lasting for more, but is there any chance for me to crush again with my laptop ? Do I need to buy a new one ?

have errors on CUDA apps even with Visual Studio 2015 c++ redistributable package installed.
http://www.enigmaathome.net/result.php?resultid=287091326

I installed it too (both as they are 2). No change.
Please notice that SETI GPU tasks (CUDA and OpenCl) work fine on my PC.

My GPU isn't a GT 460M, but a GT 540M (2048MB) driver: 385.41 OpenCL: 1.1.

The strange thing is that as soon as started, the Enigma GPU tasks interfere with the display, and the computer stops within 10 seconds. This may not happen when you have 2 graphic sets, as the Intel chipset will drive the display when the Nvidia GPU is free for calculation. When only one graphic set, display and calculation are done by the same device, and the programmer has to take care not to mix both.

I think that both the 540M and 460M are too slow to run the GPU app, which is optimized for faster cards.

---

M4 Project homepage
M4 Project wiki

I think that both the 540M and 460M are too slow to run the GPU app, which is optimized for faster cards.

What about the 630m I have that led to this thread being reopened? It begins the tasks fine, but is hitting computation errors at about 30% complete after around 10-15 mins?

I see the 550m, 640m and 650m all returning WUs. So it's not the parallelisation to 96 CUDA cores, as that's the same as the 550m, and it's not the Fermi architecture or memory size, so I'm scratching my head a little.

Edit: Now only running for a maximum of 2 minutes per time, often only 30-40 seconds. Not sure what's changed in the past month but it's getting worse.

There is an Update 3 version available too. Haven't tried running with either version yet. I just know my WUs are all erring out without it.

http://www.microsoft.com/en-us/download/details.aspx%3Fid%3D53840

Is there a preference?

Update 3 works just fine on my i7 laptop.

That'll be because your laptop is running with a GTX 960m, a vastly more superior card than this post set out to address. (Compute Capability 5.0, versus Compute Capability 2.0 & 2.1)

I also have tried update 3, but the issue seems to be more with either WU size or something more specific to my GT 630m... which should work (CC 2.1, Fermi architecture, more up to date/powerful than a number of cards returning WUs) but for some reason isn't working.

It's literally just an updated version of the same base chip (GF108) in the GT 550m that's listed as working (14 WUs so far) http://www.enigmaathome.net/gpu_os.php

Whereas the 2GB version of the GT 630m uses the GF117 chip, and I kinda wish that was the version I had, cos I'm pretty sure that would breeze it. (More CUDA cores, higher clock speed, higher memory bandwidth)

Funny Thing: The Update 3 version from Microsoft is actually a lower version # than the one I have installed. I picked up a newer version via another source... I just can't figure out from where now...

*EDIT: Managed to track down the link for the latest version.

https://www.microsoft.com/en-us/download/details.aspx?id=48145
Microsoft Visual C++ 2015 Redistributable (x64) - 14.0.23026

https://www.microsoft.com/en-us/download/details.aspx?id=52685
Microsoft Visual C++ 2015 Redistributable (x64) - 14.0.24123
(This is the Update 3 RC)

https://www.microsoft.com/en-us/download/details.aspx?id=53840
Microsoft Visual C++ 2015 Redistributable (x64) - 14.0.24215
(This is the Official Update 3)

Aside from no checkpointing (don't know if that's a general issue in itself), I have this latest version installed on an older Dell i5 with a NVS 4200M and Engima runs. Just not sure if it'll complete properly yet...