"Hammer Test" == "Rowhammer Test"?

[BoMbY]

Does the Hammer Test in MemTest86 6.0.0 use the same patterns used in this new Rowhammer Bug scenario, uncovered by Google (See: Exploiting the DRAM rowhammer bug to gain kernel privileges)?

Would this test be a good indicator for a safe system?

[keith]

The work by the Project Zero team from Google (as well as the implementation of row hammer test in MemTest86 v6) was actually based on the research and findings of Flipping Bits in Memory Without Accessing Them: An Experimental Study of DRAM Disturbance Errors by Yoongu Kim et al. We have seen many users (as well as our own systems) report multiple 1-bit errors after running Test 13 (row hammer test) in MemTest86 v6.

Having said that, however, our algorithm could probably be optimized further by taking into account the addressing scheme of each CPU, the channels in use and ram timings, etc. Also, "double-sided hammering" could be used to induce more errors. Our current algorithm is fairly general and not targeted at any particular RAM setup or CPU.

As for how safe your system is, it depends on how you use the system (eg. home computing vs flight control system), and how willing are you to accept the risk of failures as a result of the errors (see this post)