No OpenCL platforms found Error #120

charith26 · 2021-06-08T12:43:26Z

charith26
Jun 8, 2021

I'm trying to setup TornadoVM on a GPU backed EC2 instance. I'm getting the following error when I run the tornadoDeviceInfo command. What could be the issue?

$ tornadoDeviceInfo
WARNING: Using incubator modules: jdk.incubator.vector, jdk.incubator.foreign
[TornadoVM-OCL-JNI] ERROR : clGetPlatformIDs -> Returned: -1001
[TornadoVM-OCL-JNI] ERROR : clGetPlatformIDs -> Returned: -30
Exception in thread "main" java.lang.ExceptionInInitializerError
at tornado.drivers.common@0.10-dev/uk.ac.manchester.tornado.drivers.TornadoDeviceQuery.main(TornadoDeviceQuery.java:61)
Caused by: uk.ac.manchester.tornado.api.exceptions.TornadoBailoutRuntimeException: [WARNING] No OpenCL platforms found. Deoptimizing to sequential execution.
at tornado.drivers.opencl@0.10-dev/uk.ac.manchester.tornado.drivers.opencl.OCLDriver.(OCLDriver.java:59)
at tornado.drivers.opencl@0.10-dev/uk.ac.manchester.tornado.drivers.opencl.OCLTornadoDriverProvider.createDriver(OCLTornadoDriverProvider.java:50)
at tornado.runtime@0.10-dev/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.loadDrivers(TornadoCoreRuntime.java:168)
at tornado.runtime@0.10-dev/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.(TornadoCoreRuntime.java:150)
at tornado.runtime@0.10-dev/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.(TornadoCoreRuntime.java:98)
... 1 more

Thanks,

Answered by stratika

Jun 8, 2021

I think @jjfumero is right. I tried recently to run setup TornadoVM in AWS EC2 GPU instances and have created some notes about installing all the requirements and the OpenCL drivers for the instances that have the AMD and the Nvidia GPUs. I hope it can be helpful.

View full answer

jjfumero · 2021-06-08T12:59:29Z

jjfumero
Jun 8, 2021
Maintainer

You need to install the OpenCL driver for your GPU. If the Amazon instance contains an NVIDIA GPU, you can install CUDA SDK to get OpenCL as well as the PTX backends.

0 replies

stratika · 2021-06-08T13:04:27Z

stratika
Jun 8, 2021
Maintainer

I think @jjfumero is right. I tried recently to run setup TornadoVM in AWS EC2 GPU instances and have created some notes about installing all the requirements and the OpenCL drivers for the instances that have the AMD and the Nvidia GPUs. I hope it can be helpful.

2 replies

charith26 Jun 9, 2021
Author

Thank you for pointing this out. The guide really helped too. I chose a DeepLearing AMI when launching the instance and assumed it would work.

stratika Jun 9, 2021
Maintainer

Nice! Unfortunately, we have not tested the DeepLearning AMI yet.

jnorthrup · 2022-07-11T19:33:35Z

jnorthrup
Jul 11, 2022

hello

I have a similar outcome for a bare metal gentoo installation and would like to know how to best diagnose what might be link errors since the system reports relative success from the build script.

i am targetting nvidia on tornado v14.0

if i build only for ptx the nvidia error responses look similar.

if i include spirv in the script the java throws a stacktrace.

tornado --devices
WARNING: Using incubator modules: jdk.incubator.vector, jdk.incubator.foreign
[TornadoVM-OCL-JNI] ERROR : clGetDeviceIDs -> Returned: 4294967295
[TornadoVM-OCL-JNI] ERROR : clGetDeviceIDs -> Returned: 4294967266
[TornadoVM-OCL-JNI] ERROR : clCreateContext -> Returned: -30

Number of Tornado drivers: 2
Driver: OpenCL
  Total number of OpenCL devices  : 1
  Tornado device=0:0
        OPENCL --  [NVIDIA CUDA] -- NVIDIA GeForce RTX 3060 Ti
                Global Memory Size: 7.8 GB
                Local Memory Size: 48.0 KB
                Workgroup Dimensions: 3
                Total Number of Block Threads: 1024
                Max WorkGroup Configuration: [1024, 1024, 64]
                Device OpenCL C version: OpenCL C 1.2

Driver: PTX
  Total number of PTX devices  : 1
  Tornado device=1:0
        PTX -- PTX -- NVIDIA GeForce RTX 3060 Ti
                Global Memory Size: 7.8 GB
                Local Memory Size: 48.0 KB
                Workgroup Dimensions: 3
                Total Number of Block Threads: 2147483647
                Max WorkGroup Configuration: [1024, 1024, 64]
                Device OpenCL C version: N/A

[TornadoVM-OCL-JNI] ERROR : clReleaseContext -> Returned: -34

the system has the following packages installed

dev-libs/opencl-icd-loader-2022.01.04::gentoo  USE="-test" ABI_X86="32 (64) (-x32)" 0 KiB
dev-util/spirv-headers-1.3.211::gentoo  0 KiB
dev-util/spirv-tools-1.3.211-r1::gentoo  ABI_X86="(64) -32 (-x32)" 0 KiB
dev-util/nvidia-cuda-sdk-11.5.1:0/11.5.1::gentoo  USE="cuda doc* examples* mpi* opencl -debug" 0 KiB
dev-util/spirv-llvm-translator-14.0.0-r1:14::gentoo  USE="tools -test" 0 KiB

schedulling a java static function shows something similar

(1621, 1.0)
********************************************************
[TornadoVM-OCL-JNI] ERROR : clGetDeviceIDs -> Returned: 4294967295
[TornadoVM-OCL-JNI] ERROR : clGetDeviceIDs -> Returned: 4294967266
[TornadoVM-OCL-JNI] ERROR : clCreateContext -> Returned: -30
Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: Index 0 out of bounds for length 0
        at tornado.drivers.opencl@0.14/uk.ac.manchester.tornado.drivers.opencl.OCLDriver.checkAndInitBackend(OCLDriver.java:112)
        at tornado.drivers.opencl@0.14/uk.ac.manchester.tornado.drivers.opencl.OCLDriver.getDefaultBackend(OCLDriver.java:173)
        at tornado.drivers.opencl@0.14/uk.ac.manchester.tornado.drivers.opencl.OCLDriver.getProviders(OCLDriver.java:205)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.tasks.TornadoTaskSchedule.addInner(TornadoTaskSchedule.java:480)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.tasks.TornadoTaskSchedule.addInner(TornadoTaskSchedule.java:1837)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.tasks.TornadoTaskSchedule.addTask(TornadoTaskSchedule.java:1898)
        at tornado.api@0.14/uk.ac.manchester.tornado.api.TaskSchedule.task(TaskSchedule.java:215)
        at org.bereft.LauncherEnv$Companion.evaluate1(LauncherEnv.kt:246)

TIA!

1 reply

jjfumero Jul 12, 2022
Maintainer

The minimum OpenCL implementation required for TornadoVM is 2.1 from v0.14. We need to update this in the documentation.

Could you please update your NVIDIA drivers?
What's the error you get for the SPIR-V build? You need an Intel Integrated GPU with the compute-runtime driver installed. Then, the build script of TornadoVM will install all dependencies (e.g., Level Zero, etc).

jnorthrup · 2022-07-12T11:02:50Z

jnorthrup
Jul 12, 2022

thank you so much for adding clarity here. per below, it looks like i have some ldconfig hell to trace through

https://developer.nvidia.com/blog/nvidia-is-now-opencl-3-0-conformant/ is non-specific whether i get opencl 3.0 in the nvidia-drivers-215 package or if i must downgrade to 465.

if i set the bios on alder lake to use both intel+nvidia graphics concurrently x11-server and steam tend to get overrun by gremlins.

if ptx can work with recent drivers i can focus on opencl and spirv later once i get a known-good acceleration baseline

Number of platforms                               1
  Platform Name                                   Clover
  Platform Vendor                                 Mesa
  Platform Version                                OpenCL 1.1 Mesa 22.0.3
  Platform Profile                                FULL_PROFILE
  Platform Extensions                             cl_khr_icd
  Platform Extensions function suffix             MESA

  Platform Name                                   Clover
Number of devices                                 0

NULL platform behavior
  clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...)  No platform
  clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...)   No platform
  clCreateContext(NULL, ...) [default]            No platform
  clCreateContext(NULL, ...) [other]              
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM)  No devices found in platform
  clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL)  No devices found in platform

second question spirv-only build:

$ tornado --devices
WARNING: Using incubator modules: jdk.incubator.foreign, jdk.incubator.vector
Exception in thread "main" java.lang.ExceptionInInitializerError
        at tornado.drivers.common@0.14/uk.ac.manchester.tornado.drivers.TornadoDeviceQuery.main(TornadoDeviceQuery.java:73)
Caused by: uk.ac.manchester.tornado.api.exceptions.TornadoBailoutRuntimeException: [Warning] No SPIRV platforms found. Deoptimizing to sequential execution
        at tornado.drivers.spirv@0.14/uk.ac.manchester.tornado.drivers.spirv.SPIRVDriver.<init>(SPIRVDriver.java:67)
        at tornado.drivers.spirv@0.14/uk.ac.manchester.tornado.drivers.spirv.SPIRVTornadoDriverProvider.createDriver(SPIRVTornadoDriverProvider.java:47)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.loadDrivers(TornadoCoreRuntime.java:173)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.<init>(TornadoCoreRuntime.java:107)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.<clinit>(TornadoCoreRuntime.java:83)
        ... 1 more

using only --ptx i get, now, with nvidia-cuda-toolkit 11.7

 $ tornado --devices
WARNING: Using incubator modules: jdk.incubator.vector, jdk.incubator.foreign
[TornadoVM-PTX-JNI] ERROR : cuDeviceGetCount -> Returned: 3
Exception in thread "main" java.lang.ExceptionInInitializerError
        at tornado.drivers.ptx@0.14/uk.ac.manchester.tornado.drivers.ptx.PTXDriver.<init>(PTXDriver.java:50)
        at tornado.drivers.ptx@0.14/uk.ac.manchester.tornado.drivers.ptx.PTXTornadoDriverProvider.createDriver(PTXTornadoDriverProvider.java:48)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.loadDrivers(TornadoCoreRuntime.java:173)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.<init>(TornadoCoreRuntime.java:107)
        at tornado.runtime@0.14/uk.ac.manchester.tornado.runtime.TornadoCoreRuntime.<clinit>(TornadoCoreRuntime.java:83)
        at tornado.drivers.common@0.14/uk.ac.manchester.tornado.drivers.TornadoDeviceQuery.main(TornadoDeviceQuery.java:73)
Caused by: uk.ac.manchester.tornado.api.exceptions.TornadoBailoutRuntimeException: [WARNING] No CUDA devices found. Deoptimizing to sequential execution.
        at tornado.drivers.ptx@0.14/uk.ac.manchester.tornado.drivers.ptx.PTXPlatform.<init>(PTXPlatform.java:36)
        at tornado.drivers.ptx@0.14/uk.ac.manchester.tornado.drivers.ptx.PTX.<clinit>(PTX.java:49)
        ... 6 more


 $ cuda-config -s
8.5 9.4 9.5 10.3 10.4 11.1 11.2 11.3

6 replies

jnorthrup Jul 13, 2022

i have not run strace on this yet to look for the ld.so attempts but here's where I get with using an example with --ptx only

any experience with 701 ?

any checklist of necessary kernel support tools on the path?

WARNING: Using incubator modules: jdk.incubator.foreign, jdk.incubator.vector
Size = 2000
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 0 expected: 1006.04297 but found: 478.93076
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 1 expected: 1001.0076 but found: 480.91138
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 2 expected: 1006.0006 but found: 498.72565
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 3 expected: 1036.0298 but found: 494.9982
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 4 expected: 991.1084 but found: 480.5183
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 5 expected: 997.1881 but found: 491.35217
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 6 expected: 1010.89386 but found: 494.37778
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 7 expected: 1010.51605 but found: 485.75693
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 8 expected: 989.6739 but found: 487.36652
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 9 expected: 1001.0973 but found: 471.50443
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 10 expected: 1000.9882 but found: 489.22482
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 11 expected: 1008.3447 but found: 484.15054
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 12 expected: 978.73914 but found: 482.29733
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 13 expected: 990.96326 but found: 487.25134
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 14 expected: 990.9804 but found: 491.7143
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 15 expected: 1021.61707 but found: 511.43347
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 16 expected: 1004.4417 but found: 491.15067
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 17 expected: 996.26575 but found: 485.1386
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 18 expected: 1013.6961 but found: 485.07043
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 19 expected: 1002.1915 but found: 482.53036
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 20 expected: 1001.0439 but found: 496.05768
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 21 expected: 996.7283 but found: 490.23972
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 22 expected: 1013.0728 but found: 487.82623
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 23 expected: 981.6559 but found: 483.82727
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 24 expected: 1009.90656 but found: 487.6148
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 25 expected: 999.3026 but found: 473.87613
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 26 expected: 1011.14545 but found: 486.07614
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 27 expected: 1007.3181 but found: 488.49084
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 28 expected: 1007.4837 but found: 486.489
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 29 expected: 1027.1473 but found: 496.34912
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 30 expected: 1022.0302 but found: 495.81546
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 31 expected: 1003.18536 but found: 482.57388
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 32 expected: 1011.30066 but found: 494.21515
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 33 expected: 987.6176 but found: 482.77774
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 34 expected: 986.91113 but found: 470.44162
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 35 expected: 1018.9551 but found: 490.21936
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 36 expected: 1017.9433 but found: 494.35437
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 37 expected: 989.24774 but found: 488.03482
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 38 expected: 1016.0163 but found: 492.51285
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 39 expected: 1011.9941 but found: 495.54044
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 40 expected: 1000.03485 but found: 485.291
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 41 expected: 1010.7331 but found: 498.4033
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 42 expected: 1020.6166 but found: 505.5043
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 43 expected: 1019.32434 but found: 496.1492
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 44 expected: 998.2298 but found: 481.54813
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 45 expected: 998.9188 but found: 492.49448
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 46 expected: 986.9938 but found: 484.2024
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 47 expected: 996.6094 but found: 493.58575
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 48 expected: 991.18915 but found: 487.0468
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 49 expected: 996.77985 but found: 486.67905
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 50 expected: 1011.4355 but found: 493.63373
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 51 expected: 1009.1317 but found: 502.3132
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 52 expected: 980.5745 but found: 478.65298
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 53 expected: 1024.4088 but found: 498.16995
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 54 expected: 977.2212 but found: 481.36386
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 55 expected: 994.79175 but found: 486.2373
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 56 expected: 978.0242 but found: 475.95874
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 57 expected: 1009.9985 but found: 492.74326
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 58 expected: 1004.59357 but found: 486.27618
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 59 expected: 1001.07166 but found: 493.84146
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 60 expected: 1021.1369 but found: 491.91885
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 61 expected: 987.83307 but found: 471.33344
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 62 expected: 1010.7604 but found: 490.95288
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 63 expected: 995.0433 but found: 480.2782
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 64 expected: 994.46936 but found: 489.16202
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 65 expected: 988.2541 but found: 483.6861
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 66 expected: 993.86115 but found: 487.25354
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 67 expected: 1026.3273 but found: 504.1928
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 68 expected: 978.5004 but found: 483.05728
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 69 expected: 1000.665 but found: 487.04144
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 70 expected: 979.9706 but found: 489.0993
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 71 expected: 998.10046 but found: 498.07794
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 72 expected: 1000.6339 but found: 487.2054
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 73 expected: 1010.50696 but found: 505.6599
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 74 expected: 1001.0631 but found: 494.0623
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 75 expected: 1000.06793 but found: 499.11456
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 76 expected: 1011.17004 but found: 491.9201
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 77 expected: 979.24493 but found: 481.11234
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 78 expected: 1004.01324 but found: 495.51587
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 79 expected: 1021.4427 but found: 511.7313
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 80 expected: 1009.30023 but found: 499.7407
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 81 expected: 1035.3447 but found: 503.0629
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 82 expected: 1010.5714 but found: 500.18015
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 83 expected: 999.9466 but found: 495.70538
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 84 expected: 991.0792 but found: 475.34433
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 85 expected: 999.2709 but found: 491.34882
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 86 expected: 997.34973 but found: 492.60266
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 87 expected: 989.7774 but found: 478.69858
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 88 expected: 998.4244 but found: 486.64175
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 89 expected: 1003.4582 but found: 493.50598
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 90 expected: 1007.8076 but found: 485.01868
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 91 expected: 988.6098 but found: 486.7797
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 92 expected: 1018.08563 but found: 506.57208
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 93 expected: 982.3063 but found: 472.12485
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 94 expected: 984.2389 but found: 471.07062
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 95 expected: 1003.5598 but found: 484.50415
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 96 expected: 1007.6399 but found: 499.35742
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 97 expected: 976.1551 but found: 473.158
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 98 expected: 1020.7483 but found: 494.68597
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 99 expected: 1000.18695 but found: 490.4196
[TornadoVM-PTX-JNI] ERROR : cuLaunchKernel -> Returned: 701
Result is not correct - iteration: 100 expected: 1001.4723 but found: 485.52365
Median TotalTime: 433786.0

the code

package org.bereft.greatexpenses;

import uk.ac.manchester.tornado.api.TaskSchedule;
import uk.ac.manchester.tornado.api.annotations.Parallel;
import uk.ac.manchester.tornado.api.annotations.Reduce;

import java.util.ArrayList;
import java.util.Collections;
import java.util.Random;
import java.util.stream.IntStream;

class ConfigurationReduce {

    public static final int MAX_ITERATIONS = 101;
}

class Stats {

    public static double computeMedian(ArrayList<Long> input) {
        Collections.sort(input);
        double middle = input.size() >> 1;
        if (input.size() % 2 == 1) {
            middle = (input.get(input.size() / 2) + input.get(input.size() / 2 - 1)) >> 1;
        }
        return middle;
    }
}

public class /*package uk.ac.manchester.tornado.examples.reductions;*/ ReductionIrregular {

    private static void reduceFloats(float[] input, @Reduce float[] output) {
        for (@Parallel int i = 0; i < input.length; i++) {
            output[0] += input[i];
        }
    }

    private void run(final int inputSize) {

        float[] input = new float[inputSize];
        float[] result = new float[]{0.0f};
        Random r = new Random(101);

        //@formatter:off
        TaskSchedule task = new TaskSchedule("s0")
                .streamIn(input)
                .task("t0", ReductionIrregular::reduceFloats, input, result)
                .streamOut(result);
        //@formatter:on

        ArrayList<Long> timers = new ArrayList<>();
        for (int i = 0; i < ConfigurationReduce.MAX_ITERATIONS; i++) {

            IntStream.range(0, inputSize).parallel().forEach(idx -> {
                input[idx] = r.nextFloat();
            });
            float[] sequential = new float[1];
            reduceFloats(input, sequential);

            long start = System.nanoTime();
            task.execute();
            long end = System.nanoTime();
            timers.add((end - start));

            if (Math.abs(sequential[0] - result[0]) > 0.1f) {
                System.out.println("Result is not correct - iteration: " + i + " expected: " + sequential[0] + " but found: " + result[0]);
            } else {
                System.out.println("Iteration: " + i + " is correct");
            }
        }

        System.out.println("Median TotalTime: " + Stats.computeMedian(timers));

    }

    public static void main(String[] args) {
        int inputSize = 2000;
        if (args.length > 0) {
            inputSize = Integer.parseInt(args[0]);
        }
        System.out.println("Size = " + inputSize);
        new ReductionIrregular().run(inputSize);
    }
}

the run script

source ~/work/TornadoVM/source.sh

tornado --debug -Xmx9G -XX:+PrintFlagsFinal -XX:+UseFMA -XX:+UseNUMA    \
        -XX:-UseZGC -XX:-UseG1GC -XX:+UseParallelGC -XX:-UseShenandoahGC \
        -XX:+AlwaysCompileLoopMethods -ea -XX:-UseCompressedOops        \
-cp "$PWD/target/classes:$PWD/target/lib/*" org.bereft.greatexpenses.ReductionIrregular

pom

 
       <dependencies> 
        <dependency>
            <groupId>tornado</groupId>
            <artifactId>tornado-api</artifactId>
            <version>0.15-dev</version>
        </dependency>
        <dependency>
            <groupId>tornado</groupId>
            <artifactId>tornado-matrices</artifactId>
            <version>0.15-dev</version>
        </dependency>
    </dependencies>

jnorthrup Jul 13, 2022

CUDA_ERROR_LAUNCH_OUT_OF_RESOURCES - 701
This indicates that a launch did not occur because it did not have appropriate resources.

This error usually indicates that the user has attempted to pass too many arguments to the device kernel, or the kernel launch specifies too many threads for the kernel's register count. Passing arguments of the wrong size (i.e. a 64-bit pointer when a 32-bit int is expected) is equivalent to passing too many arguments and can also result in this error.

jjfumero Jul 13, 2022
Maintainer

Thanks for the report. We do not see this error in our testbed (GTX 1050, 2060 and GP100). Can you please open a bug-report with this use case?

jjfumero Jul 13, 2022
Maintainer

M.. I actually checked on my new RTX 3070 with NVIDIA 510 driver and I also get this error. Something happened here. We will take a look.

jnorthrup Jul 15, 2022

cuda/opencl shows up in flatpak builds also. im libc-challenged to un-shar the nvidia distribution of cuda-toolkit but the flatpak ubuntu-ish binary distribution is behaving same way for driver 515.x

opencl works for my isolated sample but my own project code in progress is exerting a jvm critical error. will prepare that as issue as more is known.

jnorthrup · 2022-07-13T13:28:47Z

jnorthrup
Jul 13, 2022

running with opencl produced errors due to mesa compiled with intel opencl 1.2 stubs.

rebuilding without mesa opencl produces what might be correct results from the code above and script.

Size = 2000
Task info: XXX__GENERATED_REDUCE0.reduce_seq0
        Backend           : OPENCL
        Device            : NVIDIA GeForce RTX 3060 Ti CL_DEVICE_TYPE_GPU (available)
        Dims              : 0
        Global work offset: [0]
        Global work size  : [1]
        Local  work size  : [1, 1, 1]
        Number of workgroups  : [1]

Iteration: 0 is correct
Task info: XXX__GENERATED_REDUCE0.reduce_seq0
        Backend           : OPENCL
        Device            : NVIDIA GeForce RTX 3060 Ti CL_DEVICE_TYPE_GPU (available)
        Dims              : 0
        Global work offset: [0]
        Global work size  : [1]
        Local  work size  : [1, 1, 1]
        Number of workgroups  : [1]

Iteration: 1 is correct
Task info: XXX__GENERATED_REDUCE0.reduce_seq0
        Backend           : OPENCL
        Device            : NVIDIA GeForce RTX 3060 Ti CL_DEVICE_TYPE_GPU (available)
        Dims              : 0
        Global work offset: [0]
        Global work size  : [1]
        Local  work size  : [1, 1, 1]
        Number of workgroups  : [1]

Iteration: 2 is correct
Task info: XXX__GENERATED_REDUCE0.reduce_seq0
        Backend           : OPENCL
        Device            : NVIDIA GeForce RTX 3060 Ti CL_DEVICE_TYPE_GPU (available)
        Dims              : 0
        Global work offset: [0]
        Global work size  : [1]
        Local  work size  : [1, 1, 1]
        Number of workgroups  : [1]

Iteration: 3 is correct
Task info: XXX__GENERATED_REDUCE0.reduce_seq0
        Backend           : OPENCL
        Device            : NVIDIA GeForce RTX 3060 Ti CL_DEVICE_TYPE_GPU (available)
        Dims              : 0
        Global work offset: [0]
        Global work size  : [1]
        Local  work size  : [1, 1, 1]
        Number of workgroups  : [1]

[...]

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

No OpenCL platforms found Error #120

{{title}}

Replies: 5 comments 9 replies

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Select a reply

No OpenCL platforms found Error #120

charith26 Jun 8, 2021

Replies: 5 comments · 9 replies

jjfumero Jun 8, 2021 Maintainer

stratika Jun 8, 2021 Maintainer

charith26 Jun 9, 2021 Author

stratika Jun 9, 2021 Maintainer

jnorthrup Jul 11, 2022

jjfumero Jul 12, 2022 Maintainer

jnorthrup Jul 12, 2022

jnorthrup Jul 13, 2022

jnorthrup Jul 13, 2022

jjfumero Jul 13, 2022 Maintainer

jjfumero Jul 13, 2022 Maintainer

jnorthrup Jul 15, 2022

jnorthrup Jul 13, 2022

charith26
Jun 8, 2021

Replies: 5 comments 9 replies

jjfumero
Jun 8, 2021
Maintainer

stratika
Jun 8, 2021
Maintainer

charith26 Jun 9, 2021
Author

stratika Jun 9, 2021
Maintainer

jnorthrup
Jul 11, 2022

jjfumero Jul 12, 2022
Maintainer

jnorthrup
Jul 12, 2022

jjfumero Jul 13, 2022
Maintainer

jjfumero Jul 13, 2022
Maintainer

jnorthrup
Jul 13, 2022