search

EPFL-LAP / dynamatic

DHLS (Dynamic High-Level Synthesis) compiler based on MLIR
Other
49 stars 14 forks source link

Wrong result used in integration tests #79

Closed chatpion closed 3 months ago

chatpion commented 3 months ago

Hi,

The gcd example of the integration tests seems to expect wrong values.

It is supposed to compute the GCD of 7966496 and 314080416, which is 32. Running the C code returns the correct result, but the generated circuit returns instead 7966496. However, it seems that the integration test expects this value, instead of the correct one.

I'm working in the VM provided in the README.

lucas-rami commented 3 months ago

Hi! Thanks for the bug report :)

It looks like there is in fact an issue with our gcd algorithm. Both the C code and the circuit return 7966496 when running on my machine. I will soon push a fix.

chatpion commented 3 months ago

Hi, does the circuit generated from your fixed code manage to return a value? On my end, the simulation was still not done after running for a few hours. This also happens when using smaller values for a and b.