<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Cuckoo-GPU: Accelerating Cuckoo Filters on Modern GPUs · HackerPapers discussions</title><link>https://hackerpaper.kaipfister.de/p/2603.15486/</link><description>Collected HN comments and submissions linking this paper.</description><language>en</language><lastBuildDate>Mon, 02 Nov 2026 11:16:22 GMT</lastBuildDate><atom:link href="https://hackerpaper.kaipfister.de/p/2603.15486/feed.xml" rel="self" type="application/rss+xml" /><item><title>Cuckoo-GPU: Accelerating Cuckoo Filters on Modern GPUs — mentioned by tdortman</title><link>https://news.ycombinator.com/item?id=47414373</link><guid isPermaLink="false">urn:hackerpapers:mention:2603.15486:47414373</guid><pubDate>Tue, 17 Mar 2026 15:51:00 GMT</pubDate><description>&lt;p&gt;Hi HN, I&amp;#x27;m Tim, the primary author of Cuckoo-GPU. We&amp;#x27;re open-sourcing a header-only CUDA library that brings high-performance Cuckoo filters to modern GPUs. You can read the full arXiv preprint here: https://arxiv.org/abs/2603.15486 The Problem: Approximate Membership Query (AMQ) data structures like Bloom filters are heavily used to filter out non-members before expensive I/O operations, but they often don&amp;#x27;t support deleting items. While dynamic alternatives like Cuckoo filters solve this, porting them to GPUs has historically been a nightmare. The eviction chains cause massive thread…&lt;/p&gt;&lt;p&gt;&lt;a href="https://hackerpaper.kaipfister.de/p/2603.15486/"&gt;Paper and discussion history&lt;/a&gt;. A mention is not an endorsement.&lt;/p&gt;</description><category>cs.DC</category></item></channel></rss>
