Folding Forum

In W10 with client working on cpu the gpu 0:2 SKL GT2 (HD Graphics 515) seems not to work on my Samsung tablet. Is there a adjustment that can be made? I have client working cpu and gpu on four other x86 machines.

Think the gen is high enough ... do you have the beta gpu flag set to true? ... https://foldingforum.org/viewtopic.php? ... ag#p346637

Neil-B wrote:Think the gen is high enough ... do you have the beta gpu flag set to true? ... https://foldingforum.org/viewtopic.php? ... ag#p346637

Now it's getting interesting. I tried to enter the GPU-Beta but it was not accepted anywhere... Where to enter it?

Open FAHControl ... Click Configure ... Click on Slots Tab ... Select GPU Slot and Click on Edit ... Click on Add under the extra slot options ... Enter gpu-beta in the name field and true i the value field then click on ok ... cant remember if a restart is necessary after setting this flag - if it doesn't seem to work straight away then try a restart.

Having said that have just noticed that my iGPU is currently on 11 attempts to get a new wu so their may be a shortage of appropriate wus for iGPUs at the moment - even if it fails to get an assignment you should still see i the log and in the control that the slot is actually functional (if not actually folding due to lack of wu)

Neil-B wrote:Open FAHControl ... Click Configure ... Click on Slots Tab ... Select GPU Slot an[...]

Oka, thanks for the explenation. Thats what I tried but it was not working even after several restarts. Maybe I added constantly typos...
And if it works it is not said, that the iGPU does not tell me "Disabled"

I am not exactly sure if it is even worth the effort. Ho many PPD can I expect with my massive 24 UHD620 shaders?

not much tbh - but mine does complete within timeout so yours might and for me running the igpu rather than the cpu kept temperatures in better position for iirc the same amount of points ... I guess you are putting the "-" in gpu-beta - and I typed the flag all lower case in case it makes a difference ? ... when you reedit the gpu slot is the flag still there?

.... and techpowerup has your igpu with 192 shaders same as mine

.... post your log (inc header) and we can see if there is something in the setup?

If it is ready to work and is just awaiting assignment of a wu then the statuys view sin FAHControl should show the slot as ready with the number of attempts and next attempt timer showing in the panel to the right ... the attempts to get an assignment should also show in the logs

Neil-B wrote:not much tbh - but mine does complete within timeout so yours might and for me running the igpu rather than the cpu kept temperatures in better position for iirc the same amount of points ... I guess you are putting the "-" in gpu-beta - and I typed the flag all lower case in case it makes a difference ? ... when you reedit the gpu slot is the flag still there?

.... and techpowerup has your igpu with 192 shaders same as mine

.... post your log (inc header) and we can see if there is something in the setup?

After booting yesterday evening suddely the iGPU was READY... But as you said, I could not assign a WU yet. Obviously there are not so many suitable Projects for iGPUs.

I went to Intel and was encouraged to download Intel Driver and Support Assistant. That chose a driver update for the UHD Gaphics for me, and I was in the situation described above, 'Ready', but repeatedly failing to be assigned a WU. That lasted a couple of days, but last time I looked it was working on a Moonshot rapid sprint job
(Windows 10, Core i7 in my laptop)

Yup there is an igpu project back in play

FACTS:
The Intel iGPs are extremely slow compared to other GPUs. Also if you also fold with a CPU slot, it will slow down, increasing your total throughput only slightly. In other words, don't expect much.

Speed it really important to FAH projects and slow GPUs do slow down the whole project. Therefore assignments to your iGP will often be throttled. There may or may not be specific projects that will benefit from assigning WUs to Intel iGPs.

Hope I'm ok to post this here, rather than in the GPU projects forum. I'm getting quite a few core restarts and even some aborts with project 13439, running on the Intel igpu (UHD Graphics 'Comet Lake'). Just wondering if it's just me, or are other people seeing the same?
I'm also folding on the cpu (paused at the moment: I was hoping to see how fast the igpu could work) and an Nvidia Geforce MX250. I know the work rate of the igpu is very small, but it's the only one of my three slots that's getting covid work units now.
Many units complete successfully. When it goes wrong, I get this first (seems it can happen at any stage):
and then if it happens a second time, this:

bruce wrote:FACTS:
The Intel iGPs are extremely slow compared to other GPUs. Also if you also fold with a CPU slot, it will slow down, increasing your total throughput only slightly. In other words, don't expect much.

Speed it really important to FAH projects and slow GPUs do slow down the whole project. Therefore assignments to your iGP will often be throttled. There may or may not be specific projects that will benefit from assigning WUs to Intel iGPs.

Would it -in theory- be possible to create WUs in the future where the Intel IGPUs and CPU fold together?

PeteHobbis wrote:Hope I'm ok to post this here, rather than in the GPU projects forum. I'm getting quite a few core restarts and even some aborts with project 13439, running on the Intel igpu (UHD Graphics 'Comet Lake'). Just wondering if it's just me, or are other people seeing the same?
I'm also folding on the cpu (paused at the moment: I was hoping to see how fast the igpu could work) and an Nvidia Geforce MX250. I know the work rate of the igpu is very small, but it's the only one of my three slots that's getting covid work units now.
Many units complete successfully. When it goes wrong, I get this first (seems it can happen at any stage):

Code: Select all

11:04:12:WU02:FS02:0x22:Completed 840000 out of 1000000 steps (84%)
11:07:24:WU02:FS02:0x22:Completed 850000 out of 1000000 steps (85%)
11:07:24:WU02:FS02:0x22:Checkpoint completed at step 850000
11:09:20:WU02:FS02:0x22:An exception occurred at step 855658: Particle coordinate is nan
11:09:20:WU02:FS02:0x22:ERROR:98: Attempting to restart from last good checkpoint by restarting core.
11:09:20:WU02:FS02:0x22:Folding@home Core Shutdown: CORE_RESTART
11:09:20:WARNING:WU02:FS02:FahCore returned: CORE_RESTART (98 = 0x62)
11:09:20:WU02:FS02:Starting
11:09:20:WU02:FS02:Running FahCore: "C:\Program Files (x86)\FAHClient/FAHCoreWrapper.exe" C:\ProgramData\FAHClient\cores/cores.foldingathome.org/win/64bit/22-0.0.13/Core_22.fah/FahCore_22.exe -dir 02 -suffix 01 -version 706 -lifeline 7944 -checkpoint 15 -opencl-platform 1 -opencl-device 0 -gpu-vendor intel -gpu 0 -gpu-usage 100
11:09:20:WU02:FS02:Started FahCore on PID 47548

and then if it happens a second time, this:

Code: Select all

11:19:46:WU02:FS02:0x22:Completed 880000 out of 1000000 steps (88%)
11:23:08:WU02:FS02:0x22:Completed 890000 out of 1000000 steps (89%)
11:23:09:WU02:FS02:0x22:An exception occurred at step 890045: Particle coordinate is nan
11:23:09:WU02:FS02:0x22:Max number of attempts to resume from last checkpoint (2) reached. Aborting.
11:23:09:WU02:FS02:0x22:ERROR:114: Max number of attempts to resume from last checkpoint reached.
11:23:09:WU02:FS02:0x22:Saving result file ..\logfile_01.txt
11:23:09:WU02:FS02:0x22:Saving result file globals.csv
11:23:09:WU02:FS02:0x22:Saving result file science.log
11:23:09:WU02:FS02:0x22:Saving result file state.xml.bz2
11:23:09:WU02:FS02:0x22:Folding@home Core Shutdown: BAD_WORK_UNIT
11:23:09:WARNING:WU02:FS02:FahCore returned: BAD_WORK_UNIT (114 = 0x72)
11:23:09:WU02:FS02:Sending unit results: id:02 state:SEND error:FAULTY project:13439 run:13014 clone:9 gen:0 core:0x22 unit:0x00000009000000000000347f000032d6
11:23:09:WU02:FS02:Uploading 290.84KiB to 18.188.125.154
11:23:09:WU02:FS02:Connecting to 18.188.125.154:8080
11:23:11:WU01:FS02:Connecting to assign1.foldingathome.org:80
11:23:14:WU01:FS02:Assigned to work server 18.188.125.154
11:23:14:WU01:FS02:Requesting new work unit for slot 02: gpu:0:2 CML GT2 [UHD Graphics] from 18.188.125.154
11:23:14:WU01:FS02:Connecting to 18.188.125.154:8080
11:23:16:727:127.0.0.1:New Web session
11:23:16:WU02:FS02:Upload complete
11:23:16:WU02:FS02:Server responded WORK_ACK (400)
11:23:16:WU02:FS02:Cleaning up

I've seen this sporadically with very old drivers on Windows so far, can you please report what driver version you're using, then update to the latest ones from intel.com? https://downloadcenter.intel.com/product/80939/Graphics

The UHD Graphics driver is version 27.20.100.9168, which Intel Driver & Support Assist says is up to date. I only installed that recently (maybe 12 Feb, the day of my first post above)
I see there's a 27.20.100.9313 listed as Beta - I'll try downloading that.
Thanks!