Folding Forum

Please note that work is happening to ensure that FAHControl is moved to Python 3 (https://github.com/FoldingAtHome/fah-control/pulls) Should also switch to real systemd based service handling at least in redhat based rpm. rhel6 is coming to EOL and RHEL7 is systemd based, lika all supported fedora...

OpenMM requires OpenCL 1.2 or greater. Some WUs will run if your GPU will only support OCL 1.1 but you'll get a number of errors. I know of no advantage to having greater that 1.2. CUDA is not supported yet, but I understand it's coming soon. I'm not sure what the minimum CUDA version will be (Prob...

toTOW wrote:Never heard about this ...

Well it's official AMD opensource project for OpenCL and more. I think they do not work much with proprietary OpenCL drivers anymore. ROCm is their future so it would be good idea to get FAH also support it if it doesn't already.

No, they are not all Polaris cards, the 470 and 480 are Ellesmere. But even if they were, the details of the implementation are enough different between them that it makes a lot of difference in how some code interacts with them. And as you mention, the 590 is even based on a different chip from th...

Joe_H wrote:The GPU folding cores do not work with the open source Mesa driver. To use for folding you would need to install the AMD drivers and OpenCL support.

How about with ROCm?

https://github.com/RadeonOpenCompute/ROCm

This is normal because task manager have filters and your only looking 3D, Copy, Video encode/decode.

If you click that 3D text what have little arrow pointing down (means it have menu) you can select to show Compute 1 and then task manager should agree with radeon control panel.

My hunch is that AMD may have shot themselves in the foot by releasing so many different GPU genrations with the same binary code identifying the drivers. This problem has been escalated to the OpenMM developers. No they haven't. 470/480/570/580/590 are all polaris cards and based exactly same GPU....

No, it's not Core22 that's the problem. Your system is unstable. I suspect your system is excessively overclocked. FAH often puts a greater load on the hardware than whatever bench-marking tool you used to establish stability. I''ll bet it Will run fine at stock clock rates. No my system is at DEFA...

There is something broken in core22 and corona projects, I'm getting same issue. Sometimes I can fold short time but then WU crashes. https://pastebin.com/2DgYRRta I had to put it to pastebin because it was too large to post here. This is full log after restart. As you can see there is some WU's wha...

02:59:52:WU01:FS01:0x22:ERROR:exception: Error invoking kernel sortShortList: clEnqueueNDRangeKernel (-5) This message usually is associated with a problem with the OpenCL setup. But since your system can also proceed at other times, do you happen to be folding on a system that is a laptop that swi...

Is anyone else seeing lot of failured with WU's like right after download or after some folding? 02:59:31:WU01:FS01:Download 99.73% 02:59:33:WU01:FS01:Download complete 02:59:33:WU01:FS01:Received Unit: id:01 state:DOWNLOAD error:NO_ERROR project:11746 run:0 clone:2935 gen:5 core:0x22 unit:0x0000000...

Same here and when you get WU, download is very slow. 07:37:48:WU00:FS01:Downloading 50.83MiB 07:37:56:WU00:FS01:Download 0.25% 07:38:05:WU00:FS01:Download 0.49% 07:38:21:WU00:FS01:Download 0.98% 07:38:40:WU00:FS01:Download 1.48% 07:38:47:WU00:FS01:Download 3.93% 07:39:01:WU00:FS01:Download 4.30% 07...

Since the Nvidia driver seem locked on OpenCL 1.2, it really doesn't matter if at some point in the future OpenCL and Vulkan have the same API. The API F@H has to write to is OpenCL 1.2, or it loses cross vendor compatibility. Deciding to follow CUDA or OpenCL 2.2 means different projects for diffe...

I think direction is more focused to Vulkan, maybe FAH should start to look in Vulkan compute API also? I think Khronos is slowly fading away from OpenCL and OpenGL and focus is in Vulkan now.

https://pcper.com/2017/05/breaking-open ... to-vulkan/

The network admins at Temple.edu have been working on this problem but apparently it's an intermittent problem which makes it difficult for them to isolate (and correct). About the time they do something to get the backlog of uploads accepted by either the WS or the CS, it seems like the problem ha...