OpenCL.jl
OpenCL.jl copied to clipboard
Published
20 hours ago •
JuliaGPU
JuliaGPU
Segmentation fault in test/test_array on AMD CPU
Noticed while preparing v0.5.1 release. I am unable to catch this in RR and it occurs both in julia v0.5 and v0.6.
julia -g 2 -e 'using Base.Test; using OpenCL; include(Pkg.dir("OpenCL", "test", "test_array.jl"))'
cc: @dfdx since it is happening in a piece of code that you contributed. Any ideas?
signal (11): Segmentation fault
while loading /home/wallnuss/.julia/v0.6/OpenCL/test/test_array.jl, in expression starting on line 3
clSetKernelArg at /opt/AMDAPP/SDK/lib/sdk/libamdocl64.so (unknown line)
clSetKernelArg at /home/wallnuss/.julia/v0.6/OpenCL/src/api.jl:17
macro expansion at /home/wallnuss/.julia/v0.6/OpenCL/src/macros.jl:4 [inlined]
set_arg! at /home/wallnuss/.julia/v0.6/OpenCL/src/kernel.jl:76
unknown function (ip: 0x7f46cd180a60)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
set_args! at /home/wallnuss/.julia/v0.6/OpenCL/src/kernel.jl:106
#transpose!#145 at /home/wallnuss/.julia/v0.6/OpenCL/src/array.jl:115
unknown function (ip: 0x7f46cd1d6ed2)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_invoke at /home/wallnuss/src/julia/src/gf.c:41
transpose! at /home/wallnuss/.julia/v0.6/OpenCL/src/array.jl:109
unknown function (ip: 0x7f46cd1d6ac6)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_invoke at /home/wallnuss/src/julia/src/gf.c:41
macro expansion at /home/wallnuss/.julia/v0.6/OpenCL/test/test_array.jl:53 [inlined]
macro expansion at ./test.jl:853 [inlined]
macro expansion at /home/wallnuss/.julia/v0.6/OpenCL/test/test_array.jl:0 [inlined]
macro expansion at ./test.jl:853 [inlined]
anonymous at ./<missing> (unknown line)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_toplevel_eval_flex at /home/wallnuss/src/julia/src/toplevel.c:589
jl_parse_eval_all at /home/wallnuss/src/julia/src/ast.c:847
jl_load at /home/wallnuss/src/julia/src/toplevel.c:616
include_from_node1 at ./loading.jl:539
unknown function (ip: 0x7f46cd10c012)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
include at ./sysimg.jl:14
unknown function (ip: 0x7f48e6b8a3bb)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
do_call at /home/wallnuss/src/julia/src/interpreter.c:75
eval at /home/wallnuss/src/julia/src/interpreter.c:230
jl_interpret_toplevel_expr at /home/wallnuss/src/julia/src/interpreter.c:34
jl_toplevel_eval_flex at /home/wallnuss/src/julia/src/toplevel.c:577
jl_eval_module_expr at /home/wallnuss/src/julia/src/toplevel.c:203
jl_toplevel_eval_flex at /home/wallnuss/src/julia/src/toplevel.c:480
jl_parse_eval_all at /home/wallnuss/src/julia/src/ast.c:847
jl_load at /home/wallnuss/src/julia/src/toplevel.c:616
include_from_node1 at ./loading.jl:539
unknown function (ip: 0x7f48e6cd1dcb)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
include at ./sysimg.jl:14
unknown function (ip: 0x7f48e6b8a3bb)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
process_options at ./client.jl:305
_start at ./client.jl:371
unknown function (ip: 0x7f48e6cfdd88)
jl_call_method_internal at /home/wallnuss/src/julia/src/julia_internal.h:248 [inlined]
jl_apply_generic at /home/wallnuss/src/julia/src/gf.c:2212
jl_apply at /home/wallnuss/src/julia/ui/../src/julia.h:1410 [inlined]
true_main at /home/wallnuss/src/julia/ui/repl.c:127
main at /home/wallnuss/src/julia/ui/repl.c:264
__libc_start_main at /usr/lib/libc.so.6 (unknown line)
unknown function (ip: 0x401519)
Allocations: 21279480 (Pool: 21276613; Big: 2867); GC: 48
This is the AMD SDK for CPUs v3.0
clinfo:
Platform Name AMD Accelerated Parallel Processing
Number of devices 1
Device Name Intel(R) Xeon(R) CPU E5-2643 v3 @ 3.40GHz
Device Vendor GenuineIntel
Device Vendor ID 0x1002
Device Version OpenCL 1.2 AMD-APP (1800.8)
Driver Version 1800.8 (sse2,avx)
Device OpenCL C Version OpenCL C 1.2
Device Type CPU
Device Profile FULL_PROFILE
Device Board Name (AMD)
Device Topology (AMD) (n/a)
Max compute units 12
Max clock frequency 1399MHz
Device Partition (core, cl_ext_device_fission)
Max number of sub-devices 12
Supported partition types equally, by counts, by affinity domain
Supported affinity domains L3 cache, L2 cache, L1 cache, next partitionable
Supported partition types (ext) equally, by counts, by affinity domain
Supported affinity domains (ext) L3 cache, L2 cache, L1 cache, next fissionable
Max work item dimensions 3
Max work item sizes 1024x1024x1024
Max work group size 1024
Preferred work group size multiple 1
Preferred / native vector sizes
char 16 / 16
short 8 / 8
int 4 / 4
long 2 / 2
half 4 / 4 (n/a)
float 8 / 8
double 4 / 4 (cl_khr_fp64)
Half-precision Floating-point support (n/a)
Single-precision Floating-point support (core)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations Yes
Double-precision Floating-point support (cl_khr_fp64)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Address bits 64, Little-Endian
Global memory size 67485077504 (62.85GiB)
Error Correction support No
Max memory allocation 16871269376 (15.71GiB)
Unified memory for Host and Device Yes
Minimum alignment for any data type 128 bytes
Alignment of base address 1024 bits (128 bytes)
Global Memory cache type Read/Write
Global Memory cache size 32768
Global Memory cache line 64 bytes
Image support Yes
Max number of samplers per kernel 16
Max size for 1D images from buffer 65536 pixels
Max 1D or 2D image array size 2048 images
Max 2D image size 8192x8192 pixels
Max 3D image size 2048x2048x2048 pixels
Max number of read image args 128
Max number of write image args 64
Local memory type Global
Local memory size 32768 (32KiB)
Max constant buffer size 65536 (64KiB)
Max number of constant args 8
Max size of kernel argument 4096 (4KiB)
Queue properties
Out-of-order execution No
Profiling Yes
Prefer user sync for interop Yes
Profiling timer resolution 1ns
Profiling timer offset since Epoch (AMD) 1487579561779808809ns (Mon Feb 20 17:32:41 2017)
Execution capabilities
Run OpenCL kernels Yes
Run native kernels Yes
SPIR versions 1.2
printf() buffer size 65536 (64KiB)
Built-in kernels
Device Available Yes
Compiler Available Yes
Linker Available Yes
Device Extensions cl_khr_fp64 cl_amd_fp64 cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_device_fission cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_spir cl_khr_gl_event
On Travis we are currently using AMD-APP-SDK-linux-v2.9-1.599.381-GA-x64.tar.bz2 If I remember correctly because of bugs like these...
This might be related to this commented line - if I recall it correctly, I copied most of the code for transpose from some implementation in C where the trick with block_size + 1 came unexplained, so I just removed it. Could you please try to uncomment this line (and comment out the next one, of course) and try again?
Just did and I am seeing the same behaviour. The seqfault happens when setting the first argument of the kernel https://github.com/JuliaGPU/OpenCL.jl/blob/409cb88e0921ffa5c12e79525fb16b2ab89025da/src/kernel.jl#L76, so for buffer(B). Does anybody have an AMD gpu with the newest driver? (@SimonDanisch maybe?)
I get this only with finalizers activated for Buffer... Uncommenting: https://github.com/JuliaGPU/OpenCL.jl/blob/47fbbdda994188875d2e2895aaf00402126ce02b/src/buffer.jl#L18 Removes the error! Maybe it's time to correctly define cconvert and friends!? ;)
Since the same "fix" solves also https://github.com/JuliaGPU/CLBLAS.jl/pull/25
Uh.... Not sure on what branches I was, but master actually passes all tests locally for me...
Okay, I think the last remaining problem is, that this segfaults:
for i=1:N
device, ctx, queue = cl.create_compute_context()
# do stuff
...
end
The rest seems to work fine!
These failures might be non-deterministic. Especially when finalizers are involved.
Maybe it's time to correctly define cconvert and friends!? ;)
Definitely! Want to give this a go? ;P
for i=1:N device, ctx, queue = cl.create_compute_context() # do stuff ... endThe rest seems to work fine!
Yeah that definitely should not segfault...
@SimonDanisch can you reproduce the seqfault and send a log/backtrace/rr trace along?
Figured it is better for me to comment here than open another issue. I am also getting segfaults every time, as well as a ReadOnlyMemoryError.
julia> versioninfo()
Pkg.stJulia Version 0.6.0
Commit 903644385b (2017-06-19 13:05 UTC)
Platform Info:
OS: Linux (x86_64-linux-gnu)
CPU: AMD FX(tm)-6300 Six-Core Processor
WORD_SIZE: 64
BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Piledriver)
LAPACK: libopenblas64_
LIBM: libopenlibm
LLVM: libLLVM-3.9.1 (ORCJIT, bdver1)
julia> Pkg.status("OpenCL")
- OpenCL 0.6.1 master
julia> using OpenCL
julia> cl.platforms()
ERROR: ReadOnlyMemoryError()
Stacktrace:
[1] clGetPlatformIDs(::Int64, ::Ptr{Void}, ::Base.RefValue{UInt32}) at /home/chris/.julia/v0.6/OpenCL/src/api.jl:17
[2] macro expansion at /home/chris/.julia/v0.6/OpenCL/src/macros.jl:4 [inlined]
[3] platforms() at /home/chris/.julia/v0.6/OpenCL/src/platform.jl:21
julia> cl.platforms()
ERROR: CLError(code=-1001, CL_PLATFORM_NOT_FOUND_KHR)
Stacktrace:
[1] macro expansion at /home/chris/.julia/v0.6/OpenCL/src/macros.jl:6 [inlined]
[2] platforms() at /home/chris/.julia/v0.6/OpenCL/src/platform.jl:21
julia> exit()
signal (11): Segmentation fault
while loading no file, in expression starting on line 0
unknown function (ip: 0x7f77db4b4db3)
_ZN4llvm2cl19generic_parser_base10findOptionEPKc at /home/chris/Documents/programs/julia/usr/bin/../lib/libLLVM-3.9.so (unknown line)
unknown function (ip: 0x7f77d9e4dcfc)
_ZN4llvm19MachinePassRegistry6RemoveEPNS_23MachinePassRegistryNodeE at /usr/lib/x86_64-linux-gnu/libLLVM-5.0.so.1 (unknown line)
__run_exit_handlers at /build/glibc-mXZSwJ/glibc-2.24/stdlib/exit.c:83
exit at /build/glibc-mXZSwJ/glibc-2.24/stdlib/exit.c:105
jl_exit at /home/chris/Documents/programs/julia/src/jl_uv.c:552
exit at ./initdefs.jl:22
unknown function (ip: 0x7f77d5ca40f8)
jl_call_fptr_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/chris/Documents/programs/julia/src/gf.c:1933
do_call at /home/chris/Documents/programs/julia/src/interpreter.c:75
eval at /home/chris/Documents/programs/julia/src/interpreter.c:242
jl_interpret_toplevel_expr at /home/chris/Documents/programs/julia/src/interpreter.c:34
jl_toplevel_eval_flex at /home/chris/Documents/programs/julia/src/toplevel.c:577
jl_toplevel_eval_in at /home/chris/Documents/programs/julia/src/builtins.c:496
eval at ./boot.jl:235
unknown function (ip: 0x7f77d5b2bf3f)
jl_call_fptr_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/chris/Documents/programs/julia/src/gf.c:1933
eval_user_input at ./REPL.jl:66
unknown function (ip: 0x7f77d5b99caf)
jl_call_fptr_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/chris/Documents/programs/julia/src/gf.c:1933
macro expansion at ./REPL.jl:97 [inlined]
#1 at ./event.jl:73
unknown function (ip: 0x7f77bc9952bf)
jl_call_fptr_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/chris/Documents/programs/julia/src/gf.c:1933
jl_apply at /home/chris/Documents/programs/julia/src/julia.h:1424 [inlined]
start_task at /home/chris/Documents/programs/julia/src/task.c:267
unknown function (ip: 0xffffffffffffffff)
Allocations: 6082652 (Pool: 6081287; Big: 1365); GC: 12
Segmentation fault (core dumped)
I know this isn't totally an OpenCL.jl issue, because ArrayFire gives me the same error and segfault:
julia> Pkg.test("ArrayFire")
INFO: Computing test dependencies for ArrayFire...
INFO: No packages to install, update or remove
INFO: Testing ArrayFire
ERROR: LoadError: InitError: ReadOnlyMemoryError()
Stacktrace:
[1] __init__() at /home/chris/.julia/v0.6/ArrayFire/src/util.jl:62
[2] _include_from_serialized(::String) at ./loading.jl:157
[3] _require_from_serialized(::Int64, ::Symbol, ::String, ::Bool) at ./loading.jl:200
[4] _require(::Symbol) at ./loading.jl:491
[5] require(::Symbol) at ./loading.jl:398
[6] include_from_node1(::String) at ./loading.jl:569
[7] include(::String) at ./sysimg.jl:14
[8] process_options(::Base.JLOptions) at ./client.jl:305
[9] _start() at ./client.jl:371
during initialization of module ArrayFire
while loading /home/chris/.julia/v0.6/ArrayFire/test/runtests.jl, in expression starting on line 2
signal (11): Segmentation fault
while loading no file, in expression starting on line 0
unknown function (ip: 0x7fb32f42cdb3)
_ZN4llvm2cl19generic_parser_base10findOptionEPKc at /home/chris/Documents/programs/julia/usr/bin/../lib/libLLVM-3.9.so (unknown line)
unknown function (ip: 0x7fb32ddc5cfc)
_ZN4llvm19MachinePassRegistry6RemoveEPNS_23MachinePassRegistryNodeE at /usr/lib/x86_64-linux-gnu/libLLVM-5.0.so.1 (unknown line)
__run_exit_handlers at /build/glibc-mXZSwJ/glibc-2.24/stdlib/exit.c:83
exit at /build/glibc-mXZSwJ/glibc-2.24/stdlib/exit.c:105
jl_exit at /home/chris/Documents/programs/julia/src/jl_uv.c:552
_start at ./client.jl:419
unknown function (ip: 0x7fb329ae6208)
jl_call_fptr_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/chris/Documents/programs/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/chris/Documents/programs/julia/src/gf.c:1933
jl_apply at /home/chris/Documents/programs/julia/ui/../src/julia.h:1424 [inlined]
true_main at /home/chris/Documents/programs/julia/ui/repl.c:127
main at /home/chris/Documents/programs/julia/ui/repl.c:264
__libc_start_main at /build/glibc-mXZSwJ/glibc-2.24/csu/../csu/libc-start.c:291
unknown function (ip: 0x561bd264f5e9)
Allocations: 1521654 (Pool: 1520517; Big: 1137); GC: 0
==========================================================================================[ ERROR: ArrayFire ]==========================================================================================
failed process: Process(`/home/chris/Documents/programs/julia/usr/bin/julia -Cnative -J/home/chris/Documents/programs/julia/usr/lib/julia/sys.so --compile=yes --depwarn=yes --check-bounds=yes --code-coverage=none --color=yes --compilecache=yes /home/chris/.julia/v0.6/ArrayFire/test/runtests.jl`, ProcessSignaled(11)) [0]
========================================================================================================================================================================================================
INFO: No packages to install, update or remove
ERROR: ArrayFire had test errors
Yet I am at a loss of where to go from here. When I run clinfo from the command line, it lists "Device OpenCL C Version" as OpenCL C 1.1, but the ICD Loader Profile is listed as OpenCL 2.1 -- is this an issue? Output of clinfo:
clinfo
Number of platforms 1
Platform Name Clover
Platform Vendor Mesa
Platform Version OpenCL 1.1 Mesa 17.3.0-devel
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_icd
Platform Extensions function suffix MESA
Platform Name Clover
Number of devices 1
Device Name AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)
Device Vendor AMD
Device Vendor ID 0x1002
Device Version OpenCL 1.1 Mesa 17.3.0-devel
Driver Version 17.3.0-devel
Device OpenCL C Version OpenCL C 1.1
Device Type GPU
Device Profile FULL_PROFILE
Max compute units 20
Max clock frequency 975MHz
Max work item dimensions 3
Max work item sizes 256x256x256
Max work group size 256
Preferred work group size multiple 64
Preferred / native vector sizes
char 16 / 16
short 8 / 8
int 4 / 4
long 2 / 2
half 8 / 8 (cl_khr_fp16)
float 4 / 4
double 2 / 2 (cl_khr_fp64)
Half-precision Floating-point support (cl_khr_fp16)
Denormals No
Infinity and NANs Yes
Round to nearest Yes
Round to zero No
Round to infinity No
IEEE754-2008 fused multiply-add No
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Single-precision Floating-point support (core)
Denormals No
Infinity and NANs Yes
Round to nearest Yes
Round to zero No
Round to infinity No
IEEE754-2008 fused multiply-add No
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Double-precision Floating-point support (cl_khr_fp64)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Address bits 64, Little-Endian
Global memory size 2146324480 (1.999GiB)
Error Correction support No
Max memory allocation 1502427136 (1.399GiB)
Unified memory for Host and Device No
Minimum alignment for any data type 128 bytes
Alignment of base address 32768 bits (4096 bytes)
Global Memory cache type None
Image support No
Local memory type Local
Local memory size 32768 (32KiB)
Max constant buffer size 1502427136 (1.399GiB)
Max number of constant args 16
Max size of kernel argument 1024
Queue properties
Out-of-order execution No
Profiling Yes
Profiling timer resolution 0ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Device Available Yes
Compiler Available Yes
Device Extensions cl_khr_byte_addressable_store cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_fp64 cl_khr_fp16
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) Clover
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [MESA]
clCreateContext(NULL, ...) [default] Success [MESA]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)
Platform Name Clover
Device Name AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)
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) Success (1)
Platform Name Clover
Device Name AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.11
ICD loader Profile OpenCL 2.1
I have some computational ideas that could be extremely parallelized, so I think they have a lot of potential on a GPU, but I'd like to experiment on the one I have:
$ lspci -v -s $(lspci | grep ' VGA ' | cut -d" " -f 1)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Curacao PRO [Radeon R7 370 / R9 270/370 OEM] (prog-if 00 [VGA controller])
Subsystem: Gigabyte Technology Co., Ltd Curacao PRO [Radeon R7 370 / R9 270/370 OEM]
Flags: bus master, fast devsel, latency 0, IRQ 31
Memory at c0000000 (64-bit, prefetchable) [size=256M]
Memory at fea00000 (64-bit, non-prefetchable) [size=256K]
I/O ports at e000 [size=256]
Expansion ROM at 000c0000 [disabled] [size=128K]
Capabilities: <access denied>
Kernel driver in use: amdgpu
Kernel modules: radeon, amdgpu
before shelling out a lot of money on a new high-performing one. (As a note on the kernel driver: I tried the radeon earlier, and that didn't work either.)
All this is well outside of my experience but if anyone has suggestions on where I can go or what I can read to learn about what's going on, let me know. But struggling as much as I have to get things to work here also makes me hesitant to invest in a new GPU (and on that note, I would prefer to support OpenCL over proprietary cuda). I'd also be willing to switch Linux distros. I'm on Ubuntu now, but Arch Linux for example can use AMD Catalyst.
EDIT: ArrayFire's tests all passed:
$cp -r /usr/local/share/ArrayFire/examples .
$cd examples
$mkdir build
$cd build
$cmake ..
$make
[ 0%] Building CXX object CMakeFiles/example_basic_opencl.dir/unified/basic.cpp.o
[ 1%] Linking CXX executable unified/basic_opencl
[ 1%] Built target example_basic_opencl
[ 1%] Building CXX object CMakeFiles/example_rbm_opencl.dir/machine_learning/rbm.cpp.o
---snip---
[ 99%] Built target example_deep_belief_net_unified
[100%] Building CXX object CMakeFiles/example_geneticalgorithm_unified.dir/machine_learning/geneticalgorithm.cpp.o
[100%] Linking CXX executable machine_learning/geneticalgorithm_unified
[100%] Built target example_geneticalgorithm_unified
(Not the Julia wrapper -- that one failed as above.)
I also get:
$ make
gcc -std=gnu99 -ocl-demo cl-demo.c cl-helper.c -lrt -lOpenCL
cl-helper.c: In function ‘create_context_on’:
cl-helper.c:351:15: warning: ‘clCreateCommandQueue’ is deprecated [-Wdeprecated-declarations]
*queue = clCreateCommandQueue(*ctx, dev, qprops, &status);
^
In file included from cl-helper.h:36:0,
from cl-helper.c:26:
/usr/include/CL/cl.h:1427:1: note: declared here
clCreateCommandQueue(cl_context /* context */,
^~~~~~~~~~~~~~~~~~~~
gcc -std=gnu99 -oprint-devices print-devices.c cl-helper.c -lrt -lOpenCL
cl-helper.c: In function ‘create_context_on’:
cl-helper.c:351:15: warning: ‘clCreateCommandQueue’ is deprecated [-Wdeprecated-declarations]
*queue = clCreateCommandQueue(*ctx, dev, qprops, &status);
^
In file included from cl-helper.h:36:0,
from cl-helper.c:26:
/usr/include/CL/cl.h:1427:1: note: declared here
clCreateCommandQueue(cl_context /* context */,
^~~~~~~~~~~~~~~~~~~~
$ ./print-devices
platform 0: vendor 'Mesa'
device 0: 'AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)'
$ ./cl-demo 1000000 10
Choose platform:
[0] Mesa
Enter choice: 0
Choose device:
[0] AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)
Enter choice: 0
---------------------------------------------------------------------
NAME: AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)
VENDOR: AMD
PROFILE: FULL_PROFILE
VERSION: OpenCL 1.1 Mesa 17.3.0-devel
EXTENSIONS: cl_khr_byte_addressable_store cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_int64_base_atomics cl_khr_int64_extended_atomics cl_khr_fp64 cl_khr_fp16
DRIVER_VERSION: 17.3.0-devel
Type: GPU
EXECUTION_CAPABILITIES: Kernel
GLOBAL_MEM_CACHE_TYPE: None (0)
CL_DEVICE_LOCAL_MEM_TYPE: Local (1)
SINGLE_FP_CONFIG: 0x6
QUEUE_PROPERTIES: 0x2
VENDOR_ID: 4098
MAX_COMPUTE_UNITS: 20
MAX_WORK_ITEM_DIMENSIONS: 3
MAX_WORK_GROUP_SIZE: 256
PREFERRED_VECTOR_WIDTH_CHAR: 16
PREFERRED_VECTOR_WIDTH_SHORT: 8
PREFERRED_VECTOR_WIDTH_INT: 4
PREFERRED_VECTOR_WIDTH_LONG: 2
PREFERRED_VECTOR_WIDTH_FLOAT: 4
PREFERRED_VECTOR_WIDTH_DOUBLE: 2
MAX_CLOCK_FREQUENCY: 975
ADDRESS_BITS: 64
MAX_MEM_ALLOC_SIZE: 1502427136
IMAGE_SUPPORT: 0
MAX_READ_IMAGE_ARGS: 32
MAX_WRITE_IMAGE_ARGS: 32
IMAGE2D_MAX_WIDTH: 32768
IMAGE2D_MAX_HEIGHT: 32768
IMAGE3D_MAX_WIDTH: 4096
IMAGE3D_MAX_HEIGHT: 4096
IMAGE3D_MAX_DEPTH: 4096
MAX_SAMPLERS: 32
MAX_PARAMETER_SIZE: 1024
MEM_BASE_ADDR_ALIGN: 32768
MIN_DATA_TYPE_ALIGN_SIZE: 128
GLOBAL_MEM_CACHELINE_SIZE: 0
GLOBAL_MEM_CACHE_SIZE: 0
GLOBAL_MEM_SIZE: 2146324480
MAX_CONSTANT_BUFFER_SIZE: 1502427136
MAX_CONSTANT_ARGS: 16
LOCAL_MEM_SIZE: 32768
ERROR_CORRECTION_SUPPORT: 0
PROFILING_TIMER_RESOLUTION: 0
ENDIAN_LITTLE: 1
AVAILABLE: 1
COMPILER_AVAILABLE: 1
MAX_WORK_GROUP_SIZES: 256 256 256
---------------------------------------------------------------------
*** build of 'sum' on 'AMD Radeon(TM) HD8800 Series (PITCAIRN / DRM 3.9.0 / 4.10.0-35-generic, LLVM 5.0.0)' said:
*** (end of message)
0.000634 s
18.924809 GB/s
GOOD
Which are tests that were suggested here. So tests within Julia fail, and tests outside of it have passed.
EDIT: On a different Ubuntu machine (this one 16.4; the one from above was 17.4), I followed the instructions here (primary difference was Paulo Miguel's PPA instead of Oibaf's). I'm again able to pass tests, compile and run code on the GPU (eg, hello.c from the above link), etc. But I now get a ReadOnlyMemoryError() when precompiling CLBLAS, now without a segfault upon trying to exit. I haven't tested CLBLAS. Perhaps I should file an issue there.
I'll start reading about OpenCL in my spare time.
I see you're using Mesa's Clover. Note that Clover is pretty much undeveloped now, and in my experience, segfaults if you so much as look at it funny. I could never get it to work with OpenCL.jl properly, similar situation as you're seeing (I tried with both a R9 390 and RX 480).
What I recommend is using either AMDGPU-PRO (I know, proprietary) if you can get it to work, or go the ROCm route (which I have not tested yet, but will soon). I don't think your FX CPU is going to cut it for ROCm unfortunately, as FX CPUs use PCIe 2.0, which does not support the Atomic Completions that ROCm requires.
AMDGPU-PRO, however, works nicely if you can get it to work with your kernel. It is installable as a .deb that you can get from AMD's website, and is very compatible with OpenCL.jl and similar packages.
Okay, cool -- trying amdgpu-pro on the second computer, I'm able to build CLBLAS but ran into two issues:
- https://github.com/JuliaGPU/OpenCL.jl/issues/123
- The graphics no longer seemed to be working, and I could not log in to the GUI (screen would flash and I'd be back at the login screen). I switched to tty1 to test.
"1)" is the bigger problem because I can work around "2)" through either (a) bringing a laptop for a proper screen or (b) simply un- and re-installing amdgpu-pro as needed. Per that issue, I'll try removing all of my other OpenCL packages and try again.
EDIT: a) Forgot to mention earlier, using GPUArrays I also got an error saying no method similar(GPUArray) found when trying to follow the example in the readme: GPUArray(rand(Float32, 32, 32)) b) No luck after removing mesa and purging the repository.
Also, neither the GPU of the computer I'm on currently ( Radeon HD 8490 / R5 235X OEM ) or my home computer are listed among the supported GPUs on the AMDGPU-PRO page. I suspect that may be a problem. Both GPUs you mentioned, however, are supported.
Regarding 1), it could definitely be the choice of GPU you're using. Although if it's the case that AMDGPU-PRO doesn't support it, I unfortunately can't help you, other than recommending you upgrade to something newer, like the Polaris or Vega line.
Regarding 2), it's probably some conflict between your AMDGPU kernel driver and the AMDGPU-PRO kernel module (it only works on specific versions). This is also possibly what's causing 1), if you're lucky (since you can just downgrade/upgrade your kernel to one that it supports).
And regarding your edit a) about GPUArrays, you may want to file an issue about that. I know Simon has been doing a lot of work on GPUArrays that moves functionality out of that package into other (more specific) packages, so he probably just forgot to update the docs. EDIT: Actually that might be a real bug, not just docs. Please file a ticket either way :smile:
Either way, if you want to take this offline (since this isn't really an OpenCL.jl issue), feel free to email me at jpsamaroo -at- gmail -dot- com.
I see segmentation fault as well:
julia> Pkg.test("OpenCL")
INFO: Testing OpenCL
Test Summary: | Pass Total
layout | 2 2
Platform Info: Error During Test
Test threw an exception of type ReadOnlyMemoryError
Expression: length(cl.platforms()) == cl.num_platforms()
ReadOnlyMemoryError()
Stacktrace:
[1] clGetPlatformIDs(::Int64, ::Ptr{Void}, ::Base.RefValue{UInt32}) at /home/sambit/.julia/v0.6/OpenCL/src/api.jl:15
[2] macro expansion at /home/sambit/.julia/v0.6/OpenCL/src/macros.jl:4 [inlined]
[3] platforms() at /home/sambit/.julia/v0.6/OpenCL/src/platform.jl:21
[4] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:3 [inlined]
[5] macro expansion at ./test.jl:860 [inlined]
[6] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:2 [inlined]
[7] macro expansion at ./test.jl:860 [inlined]
[8] anonymous at ./<missing>:?
Platform Info: Error During Test
Got an exception of type OpenCL.cl.CLError outside of a @test
CLError(code=-1001, CL_PLATFORM_NOT_FOUND_KHR)
Stacktrace:
[1] macro expansion at /home/sambit/.julia/v0.6/OpenCL/src/macros.jl:6 [inlined]
[2] platforms() at /home/sambit/.julia/v0.6/OpenCL/src/platform.jl:21
[3] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:4 [inlined]
[4] macro expansion at ./test.jl:860 [inlined]
[5] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:2 [inlined]
[6] macro expansion at ./test.jl:860 [inlined]
[7] anonymous at ./<missing>:?
[8] include_from_node1(::String) at ./loading.jl:576
[9] include(::String) at ./sysimg.jl:14
[10] include_from_node1(::String) at ./loading.jl:576
[11] include(::String) at ./sysimg.jl:14
[12] process_options(::Base.JLOptions) at ./client.jl:305
[13] _start() at ./client.jl:371
Platform Equality: Error During Test
Got an exception of type OpenCL.cl.CLError outside of a @test
CLError(code=-1001, CL_PLATFORM_NOT_FOUND_KHR)
Stacktrace:
[1] macro expansion at /home/sambit/.julia/v0.6/OpenCL/src/macros.jl:6 [inlined]
[2] platforms() at /home/sambit/.julia/v0.6/OpenCL/src/platform.jl:21
[3] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:17 [inlined]
[4] macro expansion at ./test.jl:860 [inlined]
[5] macro expansion at /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl:16 [inlined]
[6] macro expansion at ./test.jl:860 [inlined]
[7] anonymous at ./<missing>:?
[8] include_from_node1(::String) at ./loading.jl:576
[9] include(::String) at ./sysimg.jl:14
[10] include_from_node1(::String) at ./loading.jl:576
[11] include(::String) at ./sysimg.jl:14
[12] process_options(::Base.JLOptions) at ./client.jl:305
[13] _start() at ./client.jl:371
Test Summary: | Error Total
OpenCL.Platform | 3 3
Platform Info | 2 2
Platform Equality | 1 1
ERROR: LoadError: LoadError: Some tests did not pass: 0 passed, 0 failed, 3 errored, 0 broken.
while loading /home/sambit/.julia/v0.6/OpenCL/test/test_platform.jl, in expression starting on line 1
while loading /home/sambit/.julia/v0.6/OpenCL/test/runtests.jl, in expression starting on line 24
signal (11): Segmentation fault
while loading no file, in expression starting on line 0
unknown function (ip: 0x7f5b9f255cb9)
_ZN4llvm2cl19generic_parser_base10findOptionEPKc at /usr/local/bin/../lib/julia/libLLVM-3.9.so (unknown line)
unknown function (ip: 0x7f5b9dc5ce8c)
_ZN4llvm19MachinePassRegistry6RemoveEPNS_23MachinePassRegistryNodeE at /usr/lib/x86_64-linux-gnu/libLLVM-5.0.so.1 (unknown line)
__run_exit_handlers at /build/glibc-itYbWN/glibc-2.26/stdlib/exit.c:83
exit at /build/glibc-itYbWN/glibc-2.26/stdlib/exit.c:105
jl_exit at /buildworker/worker/package_linux64/build/src/jl_uv.c:552
_start at ./client.jl:419
unknown function (ip: 0x7f5b99c5fd28)
jl_call_fptr_internal at /buildworker/worker/package_linux64/build/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /buildworker/worker/package_linux64/build/src/julia_internal.h:358 [inlined]
jl_apply_generic at /buildworker/worker/package_linux64/build/src/gf.c:1926
jl_apply at /buildworker/worker/package_linux64/build/ui/../src/julia.h:1424 [inlined]
true_main at /buildworker/worker/package_linux64/build/ui/repl.c:127
main at /buildworker/worker/package_linux64/build/ui/repl.c:264
__libc_start_main at /build/glibc-itYbWN/glibc-2.26/csu/../csu/libc-start.c:308
unknown function (ip: 0x4016bc)
Allocations: 2324417 (Pool: 2323188; Big: 1229); GC: 2
======================================================================================[ ERROR: OpenCL ]======================================================================================
failed process: Process(`/usr/local/bin/julia -Cx86-64 -J/usr/local/lib/julia/sys.so --compile=yes --depwarn=yes --check-bounds=yes --code-coverage=none --color=yes --compilecache=yes /home/sambit/.julia/v0.6/OpenCL/test/runtests.jl`, ProcessSignaled(11)) [0]
=============================================================================================================================================================================================
INFO: No packages to install, update or remove
ERROR: OpenCL had test errors
julia>
clinfo:
Number of platforms 1
Platform Name Clover
Platform Vendor Mesa
Platform Version OpenCL 1.1 Mesa 17.2.8
Platform Profile FULL_PROFILE
Platform Extensions cl_khr_icd
Platform Extensions function suffix MESA
Platform Name Clover
Number of devices 1
Device Name AMD Radeon (TM) R7 M360 (AMD ICELAND / DRM 3.18.0 / 4.13.0-32-generic, LLVM 5.0.0)
Device Vendor AMD
Device Vendor ID 0x1002
Device Version OpenCL 1.1 Mesa 17.2.8
Driver Version 17.2.8
Device OpenCL C Version OpenCL C 1.1
Device Type GPU
Device Profile FULL_PROFILE
Max compute units 6
Max clock frequency 980MHz
Max work item dimensions 3
Max work item sizes 256x256x256
Max work group size 256
Preferred work group size multiple 64
Preferred / native vector sizes
char 16 / 16
short 8 / 8
int 4 / 4
long 2 / 2
half 0 / 0 (n/a)
float 4 / 4
double 2 / 2 (cl_khr_fp64)
Half-precision Floating-point support (n/a)
Single-precision Floating-point support (core)
Denormals No
Infinity and NANs Yes
Round to nearest Yes
Round to zero No
Round to infinity No
IEEE754-2008 fused multiply-add No
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Double-precision Floating-point support (cl_khr_fp64)
Denormals Yes
Infinity and NANs Yes
Round to nearest Yes
Round to zero Yes
Round to infinity Yes
IEEE754-2008 fused multiply-add Yes
Support is emulated in software No
Correctly-rounded divide and sqrt operations No
Address bits 64, Little-Endian
Global memory size 4293337088 (3.998GiB)
Error Correction support No
Max memory allocation 3005335961 (2.799GiB)
Unified memory for Host and Device No
Minimum alignment for any data type 128 bytes
Alignment of base address 1024 bits (128 bytes)
Global Memory cache type None
Image support No
Local memory type Local
Local memory size 32768 (32KiB)
Max constant buffer size 2147483647 (2GiB)
Max number of constant args 16
Max size of kernel argument 1024
Queue properties
Out-of-order execution No
Profiling Yes
Profiling timer resolution 0ns
Execution capabilities
Run OpenCL kernels Yes
Run native kernels No
Device Available Yes
Compiler Available Yes
Device Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_byte_addressable_store cl_khr_fp64
NULL platform behavior
clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) Clover
clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [MESA]
clCreateContext(NULL, ...) [default] Success [MESA]
clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform
clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) Success (1)
Platform Name Clover
Device Name AMD Radeon (TM) R7 M360 (AMD ICELAND / DRM 3.18.0 / 4.13.0-32-generic, LLVM 5.0.0)
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) Success (1)
Platform Name Clover
Device Name AMD Radeon (TM) R7 M360 (AMD ICELAND / DRM 3.18.0 / 4.13.0-32-generic, LLVM 5.0.0)
ICD loader properties
ICD loader Name OpenCL ICD Loader
ICD loader Vendor OCL Icd free software
ICD loader Version 2.2.11
ICD loader Profile OpenCL 2.1
Try building Julia from source with this in your Make.user:
override USE_LLVM_SHLIB = 0
I had a lot of different types of Mesa-related segfaults that were all solved by statically linking Julia and llvm.
For example, running Pkg.test("OpenCL") from Julia built normally (dynamically linked):
julia> Pkg.test("OpenCL")
INFO: Testing OpenCL
Test Summary: | Pass Total
layout | 2 2
Cannot find target for triple amdgcn-- No available targets are compatible with this triple.
signal (11): Segmentation fault
while loading /home/celrod/.julia/v0.6/OpenCL/test/test_platform.jl, in expression starting on line 1
LLVMCreateTargetMachine at /home/celrod/Documents/prog/julia/usr/bin/../lib/libLLVM-3.9.so (unknown line)
unknown function (ip: 0x7f0707c55561)
unknown function (ip: 0x7f0707c0479a)
radeon_drm_winsys_create at /usr/lib64/gallium-pipe/pipe_radeonsi.so (unknown line)
unknown function (ip: 0x7f0707b29fa0)
unknown function (ip: 0x7f0728892600)
unknown function (ip: 0x7f0728892cff)
unknown function (ip: 0x7f073e730c12)
unknown function (ip: 0x7f073e735b69)
_dl_catch_error at /lib64/libc.so.6 (unknown line)
unknown function (ip: 0x7f073e735078)
unknown function (ip: 0x7f073de10f95)
_dl_catch_error at /lib64/libc.so.6 (unknown line)
unknown function (ip: 0x7f073de11714)
dlopen at /lib64/libdl.so.2 (unknown line)
unknown function (ip: 0x7f0728b6da81)
clGetPlatformIDs at /lib64/libOpenCL.so.1 (unknown line)
clGetPlatformIDs at /home/celrod/.julia/v0.6/OpenCL/src/api.jl:15
macro expansion at /home/celrod/.julia/v0.6/OpenCL/src/macros.jl:4 [inlined]
platforms at /home/celrod/.julia/v0.6/OpenCL/src/platform.jl:21
macro expansion at /home/celrod/.julia/v0.6/OpenCL/test/test_platform.jl:3 [inlined]
macro expansion at ./test.jl:860 [inlined]
macro expansion at /home/celrod/.julia/v0.6/OpenCL/test/test_platform.jl:2 [inlined]
macro expansion at ./test.jl:860 [inlined]
anonymous at ./<missing> (unknown line)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_toplevel_eval_flex at /home/celrod/Documents/prog/julia/src/toplevel.c:589
jl_parse_eval_all at /home/celrod/Documents/prog/julia/src/ast.c:873
jl_load at /home/celrod/Documents/prog/julia/src/toplevel.c:616
include_from_node1 at ./loading.jl:576
unknown function (ip: 0x7f07140bd8a2)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/celrod/Documents/prog/julia/src/gf.c:1926
include at ./sysimg.jl:14
unknown function (ip: 0x7f0731931abb)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/celrod/Documents/prog/julia/src/gf.c:1926
do_call at /home/celrod/Documents/prog/julia/src/interpreter.c:75
eval at /home/celrod/Documents/prog/julia/src/interpreter.c:242
jl_interpret_toplevel_expr at /home/celrod/Documents/prog/julia/src/interpreter.c:34
jl_toplevel_eval_flex at /home/celrod/Documents/prog/julia/src/toplevel.c:577
jl_eval_module_expr at /home/celrod/Documents/prog/julia/src/toplevel.c:205
jl_toplevel_eval_flex at /home/celrod/Documents/prog/julia/src/toplevel.c:480
jl_parse_eval_all at /home/celrod/Documents/prog/julia/src/ast.c:873
jl_load at /home/celrod/Documents/prog/julia/src/toplevel.c:616
include_from_node1 at ./loading.jl:576
unknown function (ip: 0x7f0731a9a10b)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/celrod/Documents/prog/julia/src/gf.c:1926
include at ./sysimg.jl:14
unknown function (ip: 0x7f0731931abb)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/celrod/Documents/prog/julia/src/gf.c:1926
process_options at ./client.jl:305
_start at ./client.jl:371
unknown function (ip: 0x7f0731aa8948)
jl_call_fptr_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:339 [inlined]
jl_call_method_internal at /home/celrod/Documents/prog/julia/src/julia_internal.h:358 [inlined]
jl_apply_generic at /home/celrod/Documents/prog/julia/src/gf.c:1926
jl_apply at /home/celrod/Documents/prog/julia/ui/../src/julia.h:1424 [inlined]
true_main at /home/celrod/Documents/prog/julia/ui/repl.c:127
main at /home/celrod/Documents/prog/julia/ui/repl.c:264
__libc_start_main at /lib64/libc.so.6 (unknown line)
unknown function (ip: 0x401519)
Allocations: 3989200 (Pool: 3987906; Big: 1294); GC: 6
=================================================================[ ERROR: OpenCL ]=================================================================
failed process: Process(`/home/celrod/Documents/prog/julia/usr/bin/julia -Cnative -J/home/celrod/Documents/prog/julia/usr/lib/julia/sys.so --compile=yes --depwarn=yes --check-bounds=yes --code-coverage=none --color=yes --compilecache=yes /home/celrod/.julia/v0.6/OpenCL/test/runtests.jl`, ProcessSignaled(11)) [0]
===================================================================================================================================================
INFO: No packages to install, update or remove
ERROR: OpenCL had test errors
and when running it with a statically linked LLVM:
julia> Pkg.test("OpenCL")
INFO: Testing OpenCL
Test Summary: | Pass Total
layout | 2 2
Test Summary: | Pass Total
OpenCL.Platform | 13 13
Test Summary: | Pass Total
OpenCL.Context | 27 27
Test Summary: | Pass Total
OpenCL.Device | 40 40
Test Summary: | Pass Total
OpenCL.CmdQueue | 21 21
Test Summary: | Pass Total
OpenCL.Minver | 10 10
Test Callback
Test Summary: | Pass Total
OpenCL.Event | 16 16
Test Summary: | Pass Total
OpenCL.Program | 22 22
OpenCL.Kernel info: Error During Test
Test threw an exception of type OutOfMemoryError
Expression: typeof(k[:attributes]) == String
OutOfMemoryError()
Stacktrace:
[1] info(::OpenCL.cl.Kernel, ::Symbol) at /home/celrod/.julia/v0.6/OpenCL/src/kernel.jl:437
[2] macro expansion at /home/celrod/.julia/v0.6/OpenCL/test/test_kernel.jl:53 [inlined]
[3] macro expansion at ./test.jl:860 [inlined]
[4] macro expansion at /home/celrod/.julia/v0.6/OpenCL/test/test_kernel.jl:39 [inlined]
[5] macro expansion at ./test.jl:860 [inlined]
[6] anonymous at ./<missing>:?
OpenCL.Kernel: Test Failed
Expression: r == [1.0f0, 2.0f0, 3.0f0, 22.0f0]
Evaluated: Float32[3.0, 0.0, 0.0, 0.0] == Float32[1.0, 2.0, 3.0, 22.0]
Stacktrace:
[1] macro expansion at /home/celrod/.julia/v0.6/OpenCL/test/test_kernel.jl:258 [inlined]
[2] macro expansion at ./test.jl:860 [inlined]
[3] anonymous at ./<missing>:?
Test Summary: | Pass Fail Error Total
OpenCL.Kernel | 37 1 1 39
OpenCL.Kernel constructor | 2 2
OpenCL.Kernel info | 4 1 5
OpenCL.Kernel mem/workgroup size | 14 14
OpenCL.Kernel set_arg!/set_args! | 9 9
OpenCL.Kernel enqueue_kernel | 7 7
ERROR: LoadError: LoadError: Some tests did not pass: 37 passed, 1 failed, 1 errored, 0 broken.
while loading /home/celrod/.julia/v0.6/OpenCL/test/test_kernel.jl, in expression starting on line 7
while loading /home/celrod/.julia/v0.6/OpenCL/test/runtests.jl, in expression starting on line 31
=================================================================[ ERROR: OpenCL ]=================================================================
failed process: Process(`/home/celrod/Documents/prog/julia-static-llvm/usr/bin/julia -Cnative -J/home/celrod/Documents/prog/julia-static-llvm/usr/lib/julia/sys.so --compile=yes --depwarn=yes --check-bounds=yes --code-coverage=none --color=yes --compilecache=yes /home/celrod/.julia/v0.6/OpenCL/test/runtests.jl`, ProcessExited(1)) [1]
===================================================================================================================================================
INFO: No packages to install, update or remove
ERROR: OpenCL had test errors