"Error: Invalid argument type" error


#1

Hi,

Can anyone provide some pointers to tracking down a “Error: Invalid argument type” error?

The last few lines in the terminal of the simulation look like this:

INFO: [XOCC 60-629] Linking for hardware emulation target
INFO: [XOCC 60-895]    Target platform: /opt/Xilinx/SDx/2017.1.op/platforms/xilinx_aws-vu9p-f1_4ddr-xpr-2pr_4_0/xilinx_aws-vu9p-f1_4ddr-xpr-2pr_4_0.xpfm
INFO: [XOCC 60-423]   Target device: xilinx:aws-vu9p-f1:4ddr-xpr-2pr:4.0
INFO: [XOCC 60-251]   Hardware accelerator integration...
 ... make: *** ["/mnt/.reco-work/sdaccel/dist/xclbin"/"kernel_test".hw_emu.xilinx_aws-vu9p-f1_4ddr-xpr-2pr_4_0.xclbin] Terminated
/opt/sdaccel-builder/sdaccel-builder.mk:51: recipe for target '"/mnt/.reco-work/sdaccel/dist/xclbin"/"kernel_test".hw_emu.xilinx_aws-vu9p-f1_4ddr-xpr-2pr_4_0.xclbin' failed
  usage: aws s3 cp <LocalPath> <S3Uri> or <S3Uri> <LocalPath> or <S3Uri> <S3Uri>
Error: Invalid argument type
32de8d9b-d620-446c-ac89-89f9aa291134

thanks.


#2

The message is confusing there. It looks like your simulation timed out when trying to build, causing everything else to fail.

We have a limit of 30 minutes total just in case of infinite loops in kernels, for example. This seems to not have even started the simulation, though. Are you working on a particularly large kernel?


#3

Nope…quite small kernel that does SHA256. I’ll try and time how long it takes before I see the error message.


#4

Have you successfully simulated this code before? We just released a new version of the compiler with new latch reduction. I want to be sure that this isn’t because of that.

We do require our MD5 kernel to also compile & simulate in this time frame before release, and it’s well within the 30 minute limit.


#5

No, it hasn’t simulated before. I’ve tried to copy the structure of the MD5 design as much as possible, but SHA256 is somewhat “heavier”…it has 64 rounds of data “mangling” before producing the 256bit hash.


#6

If this happens again, could you possibly share your code? I’d like to add it as a regression case to ensure we don’t hit this issue again.


#7

It errors out after approx 35 mins. Can I email you a zip with the code?


#8

Yes. josh.bohde@reconfigure.io


#9

Should be on its way to you now.


#10

I think there’s definitely a regression on our part: compiling this is taking way longer than our other test cases. I think this is hitting some pathological behavior in our optimization rules.

I’ll have to get back to you with a fix for this.