)]}'
{
  "commit": "e1d5117ef1691a9d840aafc1ad8f7ca2e2185b31",
  "tree": "89e1f5e9f4c1c2724882b2f96b9889c9c155a5ef",
  "parents": [
    "86be10dfb67e80788634240aa32f6edd9945e8d0"
  ],
  "author": {
    "name": "Abseil Team",
    "email": "absl-team@google.com",
    "time": "Thu Nov 12 22:50:13 2026 -0800"
  },
  "committer": {
    "name": "Copybara-Service",
    "email": "copybara-worker@google.com",
    "time": "Thu Nov 12 22:52:46 2026 -0800"
  },
  "message": "Draw exact memoryless geometric skip counts in ExponentialBiased.\n\n`ExponentialBiased::GetSkipCount(mean)` previously approximated a geometric distribution by rounding `-ln(u) * mean` to the nearest integer (`std::rint`) and carrying a cumulative rounding-error term (`bias_`):\n- Nearest-integer rounding gives `0` (`GetStride(period) \u003d\u003d 1`) a half-width bin `[0, 0.5)` while every `k \u003e\u003d 1` gets width `1.0`. As a result, `P(GetSkipCount(m) \u003d\u003d 0)` was `1 / (2*m + 2/3) ~\u003d 0.5 / (m + 1)` instead of `1 / (m + 1)`, under-sampling the first event on every new thread by ~2x, introducing a ~2x refractory dip immediately after each sampled event, and negatively correlating consecutive strides through `bias_`.\n- Using only 26 PRNG bits with `+ 1.0` also floored `P(GetSkipCount(m) \u003d\u003d 0)` at `2^-26` for `m \u003e\u003e 2^26` and capped strides at `~18 * period`.\n\nEvaluate the exact `Geometric(1 / (mean + 1))` inverse CDF `floor(ln(u) / ln(1 - 1 / (mean + 1)))` using all 48 PRNG bits mapped to bin midpoints `(rng_ + 0.5) * 2^-48`, and drop `bias_` (shrinking `ExponentialBiased` from 24 B to 16 B). Every integer `k \u003e\u003d 0` now has exact geometric probability, exact expectation `mean`, and independent memoryless draws from the first call onward.\n\nPiperOrigin-RevId: 1020748224\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "8682ca15e7b67898819ce5b2fbe551aff3cb4886",
      "old_mode": 33188,
      "old_path": "absl/profiling/internal/exponential_biased.cc",
      "new_id": "b843136684a1e1390b60658df956f153d9f0aa3c",
      "new_mode": 33188,
      "new_path": "absl/profiling/internal/exponential_biased.cc"
    },
    {
      "type": "modify",
      "old_id": "d31f7782e83dca4461073c443b9b04582a5921df",
      "old_mode": 33188,
      "old_path": "absl/profiling/internal/exponential_biased.h",
      "new_id": "127358b850c06184f49b8937f61cedf5a005a033",
      "new_mode": 33188,
      "new_path": "absl/profiling/internal/exponential_biased.h"
    },
    {
      "type": "modify",
      "old_id": "3471b071064612945a65851eb5cc92e5a95eb982",
      "old_mode": 33188,
      "old_path": "absl/profiling/internal/exponential_biased_test.cc",
      "new_id": "edcd9f0e944510811956b4b3747f6416e413b8ae",
      "new_mode": 33188,
      "new_path": "absl/profiling/internal/exponential_biased_test.cc"
    }
  ]
}
