You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have tested on various environments(macOS, EC2 GPU, GPUEater).
It worked on macOS and EC2 GPU, but I got error with AMD GPU
These I faced situation.
I got error and modified source code.
/tmp/AMD_735_19/t_735_21.cl:12:13: error: access qualifier can only be used for pipe and image type
__read_only global const half *krn,
^
/tmp/AMD_735_19/t_735_21.cl:13:13: error: access qualifier can only be used for pipe and image type
__read_only global const half *bias,
^
/tmp/AMD_735_19/t_735_21.cl:14:13: error: access qualifier can only be used for pipe and image type
__read_only global const float *in,
^
/tmp/AMD_735_19/t_735_21.cl:15:13: error: access qualifier can only be used for pipe and image type
__write_only global float *out,
^
/tmp/AMD_735_19/t_735_21.cl:55:13: error: access qualifier can only be used for pipe and image type
__read_only global const float *in,
^
/tmp/AMD_735_19/t_735_21.cl:56:13: error: access qualifier can only be used for pipe and image type
__write_only global half *out
^
6 errors generated.
I could be fixed by removing __read_only , __write_only.
stopped code at Kotlin source, this is error log.
[scale2] %.1f%% - ELA: 0.0 - ETA: 00:00:03 - MEM: 596:31:23 com.jogamp.opencl.CLException$CLInvalidArgSizeException: error setting arg 7 to value java.nio.DirectByteBuffer[pos=0 lim=24 cap=24] of size 8 of CLKernel [id: 140470940142384 name: waifu2x] [error: CL_INVALID_ARG_SIZE]
at com.jogamp.opencl.CLException.newException(CLException.java:84)
at com.jogamp.opencl.CLKernel.setArgument(CLKernel.java:266)
at com.jogamp.opencl.CLKernel.setArg(CLKernel.java:194)
at com.jogamp.opencl.CLKernel.putArg(CLKernel.java:126)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2x(Kaifu2x.kt:285)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2x(Kaifu2x.kt:310)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2x(Kaifu2x.kt:327)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2x(Kaifu2x.kt:355)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2xChunkedYCbCr(Kaifu2x.kt:398)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2xChunkedRgba(Kaifu2x.kt:439)
at com.soywiz.kaifu2x.Kaifu2xOpencl.waifu2xChunkedRgbaFast(Kaifu2x.kt:455)
at com.soywiz.kaifu2x.Kaifu2x$scaleRgba$2.doResume(Kaifu2x.kt:551)
at com.soywiz.kaifu2x.Kaifu2x$scaleRgba$2.invoke(Kaifu2x.kt)
at com.soywiz.kaifu2x.Kaifu2x$scaleRgba$2.invoke(Kaifu2x.kt:513)
at com.soywiz.kaifu2x.Kaifu2xKt.processMeasurer(Kaifu2x.kt:568)
at com.soywiz.kaifu2x.Kaifu2x.scaleRgba(Kaifu2x.kt:550)
at com.soywiz.kaifu2x.Kaifu2x.scaleRgba$default(Kaifu2x.kt:545)
at com.soywiz.kaifu2x.Kaifu2xCli$main$1.doResume(Kaifu2x.kt:101)
at kotlin.coroutines.experimental.jvm.internal.CoroutineImpl.resume(CoroutineImpl.kt:54)
at kotlin.coroutines.experimental.jvm.internal.CoroutineImpl.resume(CoroutineImpl.kt:53)
at kotlin.coroutines.experimental.jvm.internal.CoroutineImpl.resume(CoroutineImpl.kt:53)
at kotlin.coroutines.experimental.SafeContinuation.resume(SafeContinuationJvm.kt:56)
at com.soywiz.korio.async.EventLoopJvmAndCSharp.step(EventLoopFactoryJvm.kt:105)
at com.soywiz.korio.async.EventLoopJvmAndCSharp.loop(EventLoopFactoryJvm.kt:122)
at com.soywiz.korio.async.EventLoop$Companion.main(EventLoop.kt:46)
at com.soywiz.korio.async.EventLoop$Companion.main(EventLoop.kt:51)
at com.soywiz.korio.KorioKt.Korio(Korio.kt:5)
at com.soywiz.kaifu2x.Kaifu2xCli.main(Kaifu2x.kt:46)
Do you think these error are can fixed?
I am hoping kaifu2x works great jobs.
The text was updated successfully, but these errors were encountered:
Hi
I have tested on various environments(macOS, EC2 GPU, GPUEater).
It worked on macOS and EC2 GPU, but I got error with AMD GPU
These I faced situation.
I could be fixed by removing
__read_only
,__write_only
.Do you think these error are can fixed?
I am hoping kaifu2x works great jobs.
The text was updated successfully, but these errors were encountered: