Project 8690 bad WU

Moderators: Site Moderators, PandeGroup

bruce
Posts: 22470
Joined: Thu Nov 29, 2007 10:13 pm
Location: So. Cal.

Re: Project 8690 bad WU

Post by bruce » Wed Jul 24, 2019 11:47 pm

Maybe I missed it, but I don't see an indication of what kind of GPU you have. It would be reported (only) near the bottom of the first page of the log.

As a wild guess, perhaps your GPU does not support DoublePrecision and OpenCL 1.2. This is a requirement for (almost) every project now but it's not well enforced. Except for change in FAH's initialization panel, I don't know what other symptoms might be indicated; if it could be causing the INTERRUPTED error.

Does FAHBench run in double precision on your system?

vincent89147
Posts: 10
Joined: Sun Jul 14, 2019 11:46 pm

Re: Project 8690 bad WU

Post by vincent89147 » Thu Jul 25, 2019 2:56 am

My slot is of the CPU type, not GPU so I'm not sure that that matters. The GPU is an NVidia 8300 (on board the M3N78 PRO motherboard) but not properly configured for use as anything but graphics. I have added this slot using "slot-add cpu", so I'm not sure that I should get WUs that require a GPU slot. If I do, though, and this project requires a GPU to run, then that may explain the interruptions.

FAHBench does not work:

Code: Select all

vv@nestor:~/FAHBench-2.3.2-Linux/bin$ ./FAHBench-cmd 
FAHBench Simulation
-------------------
Plugin directory: "/home/vv/FAHBench-2.3.2-Linux/lib/openmm"
Work unit: dhfr
WU Name: Dihydrofolate reductase
WU Description: A common system for benchmarking molecular dynamics
System XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/system.xml
Integrator XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/integrator.xml
State XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/state.xml
Step chunk: 40
Device ID 0; Platform OpenCL; Platform ID 0
Run length: 60s

Loading plugins from plugin directory
Number of registered plugins: 2
Deserializing input files: system
Deserializing input files: state
Deserializing input files: integrator
Creating context (may take several minutes)

Something went wrong:
Error initializing context: clGetPlatformIDs (-1001)
Thanks!

vincent89147
Posts: 10
Joined: Sun Jul 14, 2019 11:46 pm

Re: Project 8690 bad WU

Post by vincent89147 » Fri Jul 26, 2019 3:46 am

I took some time today to fix the GPU configuration and after some tinkering and installing the nvidia-legacy-340xx drivers, CUDA and openCL libraries I have gotten it to work properly:

Code: Select all

19:50:22:******************************* System ********************************
19:50:22:            CPU: AMD Phenom(tm) 9950 Quad-Core Processor
19:50:22:         CPU ID: AuthenticAMD Family 16 Model 2 Stepping 3
19:50:22:           CPUs: 4
19:50:22:         Memory: 7.54GiB
19:50:22:    Free Memory: 7.09GiB
19:50:22:        Threads: POSIX_THREADS
19:50:22:     OS Version: 4.19
19:50:22:    Has Battery: false
19:50:22:     On Battery: false
19:50:22:     UTC Offset: -7
19:50:22:            PID: 1050
19:50:22:            CWD: /var/lib/fahclient
19:50:22:             OS: Linux 4.19.0-5-amd64 x86_64
19:50:22:        OS Arch: AMD64
19:50:22:           GPUs: 1
19:50:22:          GPU 0: Bus:2 Slot:0 Func:0 NVIDIA:1 C77 [GeForce 8300]
19:50:22:  CUDA Device 0: Platform:0 Device:0 Bus:0 Slot:0 Compute:1.1 Driver:6.5
19:50:22:OpenCL Device 0: Platform:0 Device:0 Bus:0 Slot:0 Compute:1.0 Driver:340.107
19:50:22:***********************************************************************
However, FAHBench-cmd will not finish successfully:

Code: Select all

vv@nestor:~/FAHBench-2.3.2-Linux/bin$ ./FAHBench-cmd 
FAHBench Simulation
-------------------
Plugin directory: "/home/vv/FAHBench-2.3.2-Linux/lib/openmm"
Work unit: dhfr
WU Name: Dihydrofolate reductase
WU Description: A common system for benchmarking molecular dynamics
System XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/system.xml
Integrator XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/integrator.xml
State XML: /home/vv/FAHBench-2.3.2-Linux/share/fahbench/workunits/dhfr/state.xml
Step chunk: 40
Device ID 0; Platform OpenCL; Platform ID 0
Run length: 60s

Loading plugins from plugin directory
Number of registered plugins: 2
Deserializing input files: system
Deserializing input files: state
Deserializing input files: integrator
Creating context (may take several minutes)
Checking accuracy against reference code
Creating reference context (may take several minutes)
Comparing forces and energy

Something went wrong:
Error downloading array interactionCount: clEnqueueReadBuffer (-36)

Post Reply

Return to “Issues with a specific WU”