OpenCL + Cyclone V SoC

--- Quote Start ---

Thanks for your suggestions eyguacel,

After a rather frustrating Saturday afternoon I now have yet another flavour of embedded ARM Linux on my hard drive! The RocketBoards Linaro SD image seems to function OK on my SoCKit, but their git server also seems to be non-functional at the moment. I've been using the toolchain that came with the Altera version of ARM DS-5 from /opt/altera/13.1/embedded/DS-5/bin, although I had to build my own mkimage. I eventually found what looks to be a clone of the RocketBoards repo, and did this:

git clone git://support.criticallink.com/home/git/linux-socfpga.git
cd linux-socfpga
git checkout socfpga-3.8
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- socfpga_defconfig
make ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- LOADADDR=0x8000 uImage

I replaced the uImage on the SD card with my new one (which did of course kill my nice LXDE desktop), and built the driver and hello_world following your instructions. Where is pkg_editor.h hiding do you suppose?

Transferring that lot to the SD card as well revealed this:

root@localhost:~# ./aclsoc_load
./aclsoc_load: 21: [: 251: unexpected operator
root@localhost:~# ls /dev/acl
/dev/acl
root@localhost:~# ./hello_world
Reprogramming was successful!
FPGA is in user mode. Enabling bridges
sh: 1: cannot create /sys/class/fpga-bridge/fpga2hps/enable: Directory nonexistent
sh: 1: cannot create /sys/class/fpga-bridge/hps2fpga/enable: Directory nonexistent
sh: 1: cannot create /sys/class/fpga-bridge/lwhps2fpga/enable: Directory nonexistent
PCIe-MMD Fatal: Version mismatch, expected a0c7c1e0 but read 0
hello_world: acl_pcie_device.cpp:217: ACL_PCIE_DEVICE::ACL_PCIE_DEVICE(ACL_PCIE_BOARD*, int): Assertion `.
Aborted
root@localhost:~#

Does anyone from Altera pop in here on a regular basis to assist? I don't know if this is good form or not, but I've sent them a "Service Request". I have yet to receive a reply however.

As I mentioned at the top of the thread, I'm new to all things Altera. I've seen the SoC publicity, and gained the impression that it's possible to evaluate all their development tools free of charge on "entry level" hardware. Maybe that doesn't apply to OpenCL, or maybe it's an inadvertent "feature" that's easily fixed?

--- Quote End ---

Hi Jim,

I encountered the same problem.

pcie-mmd fatal: version mismatch, expected a0c7c1e0 but read 0

hello_world: acl_pcie_device.cpp:217: acl_pcie_device::acl_pcie_device(acl_pcie_board*, int): assertion `.

aborted

How did you solve that? Thanks.