leonarski_fandClaude Opus 5.5
87f0bdc870
FitErrorModelGPU puts the samples in the host's rank order with four stable radix passes, so the equal-count bins hold the same samples; computes every normalised deviation with the host's rounding, so every bin median is the host's to the bit; and hands the per-bin terms to the host's own update step, now ErrorModelUpdate (a pure extraction from FitErrorModel). The one thing that differs is the order each bin's two sums are added in: a fixed tree on the device, the order nth_element happened to leave the bin in on the host. a and b^2 can therefore differ from the host fit in their last bits - measured on two synthetic pools: relative difference 2e-16 and 3e-13 in a, 5e-16 in b^2. The device fit is deterministic (same pool, same bits). p.mtz md5 nevertheless unchanged on myob, cytc, 8a1a and 8qaw (GPU build): the difference does not survive the float rounding of the merged sigmas there. A knife-edge decision elsewhere could still see it, which is why this is kept as its own commit. The host fit took 0.4-0.6 s per fit on the large merges, about half of it the initial equal-count split, with only 16-way parallelism in the iterations. Device peak about 72 bytes per sample (the rank sort; each stage's buffers are freed when it is done); the free memory is checked first and a device without it stops with a message naming the CPU build. Kept last on the branch so that it can be dropped on its own. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01SVmAWnzCmRKAXVUCdc4iNi
17 lines
863 B
C++
17 lines
863 B
C++
// SPDX-FileCopyrightText: 2026 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
|
|
// SPDX-License-Identifier: GPL-3.0-only
|
|
|
|
#pragma once
|
|
|
|
#include <vector>
|
|
|
|
#include "ErrorModel.h"
|
|
|
|
// FitErrorModel (ErrorModel.h) on the GPU. The bins hold the same samples, every normalised deviation
|
|
// and so every bin median is the host's to the bit, and the update step is the host's own
|
|
// (ErrorModelUpdate). What differs is the order each bin's two sums are added in - a fixed tree here,
|
|
// the order nth_element left the bin in on the host - so a and b^2 can differ from the host fit in
|
|
// their last bits. Deterministic: the same pool gives the same fit on every run. Only built with CUDA;
|
|
// a device without room for about 72 bytes a sample throws (run on the CPU build then).
|
|
ErrorModelFit FitErrorModelGPU(const std::vector<ErrorModelSample> &pool);
|