core dump
koen-dejonghe opened this issue · comments
Koen Dejonghe commented
Examples core dump. On windows, linux, cpu and gpu.
val t1: Tensor = Tensor(Array(Seq(1.0f, 2.0f, 3.0f), Seq(4.0f, 5.0f, 6.0f)))
println(t1)
results in
# SIGSEGV (0xb) at pc=0x00007f149ef51350, pid=109309, tid=109478
Here is the build.sbt:
import Dependencies._
ThisBuild / scalaVersion := "2.12.8"
ThisBuild / version := "0.1.0-SNAPSHOT"
ThisBuild / organization := "com.example"
ThisBuild / organizationName := "example"
lazy val root = (project in file("."))
.settings(
name := "compute",
libraryDependencies += scalaTest % Test,
libraryDependencies += "com.thoughtworks.compute" %% "cpu" % "latest.release",
libraryDependencies += "com.thoughtworks.compute" %% "gpu" % "latest.release"
)
// Platform dependent runtime of LWJGL core library
libraryDependencies += ("org.lwjgl" % "lwjgl" % "latest.release").jar().classifier {
import scala.util.Properties._
if (isMac) {
"natives-macos"
} else if (isLinux) {
"natives-linux"
} else if (isWin) {
"natives-windows"
} else {
throw new MessageOnlyException(s"lwjgl does not support $osName")
}
}
Yang, Bo commented
Latest Compute.scala is built with lwjgl 3.1.6, and latest lwjgl is not backward compatible with 3.1.6.
You can replace "org.lwjgl" % "lwjgl" % "latest.release"
to "org.lwjgl" % "lwjgl" % "3.1.6"
and it should not SIGSEGV any more
Yang, Bo commented
Another fix is explicitly using latest lwjgl-opencl instead of 3.1.6, which is compatible with latest LWJGL
platform dependent runtime, as described in 252eceb
Koen Dejonghe commented
It works. Thank you.