)]}'
{
  "commit": "10c054fb12d255226c2f51efc16e3de1f30f37f2",
  "tree": "4fb76507f02c857bc7a287b77721458f2b6da62d",
  "parents": [
    "d5dd40709cd47d4227ab10ac5c3c80c8375f25ae"
  ],
  "author": {
    "name": "Abseil Team",
    "email": "absl-team@google.com",
    "time": "Fri Nov 13 00:11:11 2026 -0800"
  },
  "committer": {
    "name": "Copybara-Service",
    "email": "copybara-worker@google.com",
    "time": "Mon Nov 16 22:15:34 2026 -0800"
  },
  "message": "Replace stateful rounding bias in ExponentialBiased with `mean + 0.5` floor scaling.\n\n`ExponentialBiased::GetSkipCount(mean)` previously compensated for the `-0.5` truncation offset of `floor(-ln(u) * mean)` by rounding 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` has width `1.0`, halving `P(GetSkipCount(m) \u003d\u003d 0)` to `1 / (2*m + 2/3) ~\u003d 0.5 / (m + 1)`. This under-samples the first event on every new thread by ~2x, introduces a ~2x post-sample refractory dip, and negatively correlates consecutive draws via `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\nPre-adding `+0.5` to `mean` before truncating (`floor(-ln(u) * (mean + 0.5))`, using all 48 PRNG bits mapped to bin midpoints `(rng_ + 0.5) * 2^-48`) cancels `floor`\u0027s `-0.5` fractional offset in expectation on every draw (`E[stride] \u003d period + 1 / (12 * period) + O(period^-3)`) while producing an exact memoryless geometric distribution with full-width bins from `0` upward, lower ALU/FP cost (no `std::rint` or `bias_` load/store), and smaller TLS state (`16` B instead of `24` B). Also replace the assertion-less `printf` demo tests with non-flaky distribution shape, mean, first-draw, and lag-1 independence tests.\n\nPiperOrigin-RevId: 1020795145\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "8682ca15e7b67898819ce5b2fbe551aff3cb4886",
      "old_mode": 33188,
      "old_path": "absl/profiling/internal/exponential_biased.cc",
      "new_id": "6f5125b971564082bd8301da189c5da2bc3a11b5",
      "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": "8823598b4820bdbc6280ca5dcfd67df687dfcfa9",
      "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": "1961b3f32bc92252f6062a3f130c91cf720cacd9",
      "new_mode": 33188,
      "new_path": "absl/profiling/internal/exponential_biased_test.cc"
    }
  ]
}
