Kernel to estimate DRAM latency

[vb000]

Approach:

• Flush the Vcache w/ bank 0:
  • Implemented a function to compute eva from block index and vcache inverse hash function.
  • Using the block index --> eva function, load from all blocks in vcache associated with bank 0.
• Load from the page starting right after the Vcache size with print stats enabled.

It's unclear how we can test this test. Two ideas I'm considering:

• Look in the waveform to make sure load requests are only to vcache 0.
• Probably, I can confirm this in vcache stats output.