Skip to content

depth-clear-precision-check: Assert-fails on 32-bit platforms

The test introduced in !98 (merged) fails an assertion which tries to confirm that the floating-point math it's doing is correct. This failure has been found to occur on 32-bit builds in the Intel CI (for gen9 and ICL).

assert(maxStoreableValue * fdepth == depth);

For context, the variables are initialized as follows:

uint32_t maxStoreableValue = (1l << numOfBits) - 1;
uint64_t depth = 0;
float fdepth = depth / (float)maxStoreableValue;

This was reported by @majanes.

/cc @illia.abernikhin @yk

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information