I cloned the vortex master branch yesterday(2024.04.09) and built it follow the README.md. The build step run properly until I run the $ ./ci/blackbox.sh --cores=2 --app=vecadd , it returned with an error as follows:
Workload size=64 invalid caps id: 6 ../common.mk:94: recipe for target 'run-simx' failed
I checked the source code in the "runtime/simx/vortex.cpp" and found that it returns the error since there is no "VX_CAPS_LOCAL_MEM_SIZE" caps support in the simx runtime. And I found that the pocl runtime lib "libpocl-devices-vortex.so" exist a "vx_dev_caps" call in the newest library download by the "ci/toolchain_install.sh" script. I wonder is there any mismatching between the simx runtime and the pocl runtime?

llvm-riscv prebuild can not run in centos 7, it reports libstdc++.so.6 not
./clang: /lib64/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by ./clang)
anyone can help?

Hello,

Sorry I don't know where to raise a query, so I just put the query here.

my query is why there are llvm-riscv and riscv-gnu-toolchain?
POCL compiler is used to compile kernel, it looks that llvm-riscv and riscv-gnu-toolchain are not needed.
however, I see their headers and libs are used in APP Makefile (such as vecadd).
Could you help to comment on my understanding?

The prebuilt llvm-pocl is consists of 7 different compressed file (from partaa,partab ..... partag). But the ci/toolchain_install.sh only downloads parts a and b which leads to the following error.

bzip2: Compressed file ends unexpectedly;
	perhaps it is corrupted?  *Possible* reason follows.
bzip2: Inappropriate ioctl for device
	Input file = (stdin), output file = (stdout)

It is possible that the compressed file(s) have become corrupted.
You can use the -tvv option to test integrity of such files.

You can use the `bzip2recover' program to attempt to recover
data from undamaged sections of corrupted files.

tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now

The solution is to modify the script ci/toolchain_install.sh script's llvm-pocl subroutine/function to the following

llvm-pocl()
{
    case $OSDIR in
   "centos/7") parts=$(eval echo {a..g}) ;;
    *)          parts=$(eval echo {a..b}) ;;
    esac
    echo $parts
    rm -f llvm-pocl.tar.bz2.parta*
    for x in $parts
    do
        wget $REPOSITORY/llvm-pocl/$OSDIR/llvm-pocl.tar.bz2.parta$x
    done

The change is in the 4th line. Changing from
"centos/7") parts=$(eval echo {a..b}) ;;
to
"centos/7") parts=$(eval echo {a..g}) ;;

Hope this helps

Hi,
I would like to use my custom-built POCL with Vortex instead of pre-built binaries. However, when I build POCL, I don't get the same output directory structure as in the pre-built POCL (pocl/compiler and pocl/runtime) and Vortex Makefiles seem to assume this particular directory layout.
How do I go about getting Vortex to build with my POCL build?

Thanks,
Tharindu

on wget riscv toolchain, minor updated :), tar them I got EOF error.
Error message:
bzip2: Data integrity error when decompressing.
Input file = (stdin), output file = (stdout)

It is possible that the compressed file(s) have become corrupted.
You can use the -tvv option to test integrity of such files.

You can use the `bzip2recover' program to attempt to recover
data from undamaged sections of corrupted files.

tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now