)]}'
{
  "commit": "1c026d6659bc16ad56fde46fac3e2ed98ca6b603",
  "tree": "869b0c193939a8d89c815b4d9eaa82fc65e6ec76",
  "parents": [
    "1330f8548ed014a3ef5ff9a051bf11a78c5b520b"
  ],
  "author": {
    "name": "Rupert Swarbrick",
    "email": "rswarbrick@lowrisc.org",
    "time": "Tue Nov 16 14:18:48 2021 +0000"
  },
  "committer": {
    "name": "Rupert Swarbrick",
    "email": "rswarbrick@gmail.com",
    "time": "Tue Nov 16 16:56:07 2021 +0000"
  },
  "message": "[util] Remove debug prints to speed up secded_gen.py\n\nNow that we\u0027re adding ECC check bits to flash images, a significant\nfraction of SW build time is spent running gen-flash-img.py. Profiling\na run of:\n\n    ../util/design/gen-flash-img.py \\\n      sw/device/tests/uart_smoketest_sim_dv.64.vmem \\\n      sw/device/tests/uart_smoketest_sim_dv.64.scr.vmem\n\n(run this in the build-out directory; choice of test completely\narbitrary!) with py-spy, I realised that we were spending almost half\nour time constructing debug messages that we didn\u0027t write out.\n\nI\u0027m sure there are cleverer things we can do, but it looks like we\nprobably don\u0027t really need these so let\u0027s just drop them for now.\n\nAfter these changes, here are the big contributors to the total\nruntime for the script:\n\n    38%    load up secded_cfg.hjson\n    36%    pick an ECC code (we choose it on-the-fly)\n   ~15%    actually computing ECC bits\n\nThis still seems a little silly, but this patch gets us a factor of 2,\nso maybe it\u0027s a good start!\n\nSigned-off-by: Rupert Swarbrick \u003crswarbrick@lowrisc.org\u003e\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "9771705a8c04650fc717c2a766999a09b711d35e",
      "old_mode": 33261,
      "old_path": "util/design/secded_gen.py",
      "new_id": "f2c597b751fea2c4ab10e930973bb548c9338ad1",
      "new_mode": 33261,
      "new_path": "util/design/secded_gen.py"
    }
  ]
}