The latest articles on DEV Community by rehantheorylab-pixel (@rehantheorylabpixel). https://dev.to/rehantheorylabpixel https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3953038%2Fe49997cb-8f34-435d-b591-7dd2747aaee3.png https://dev.to/rehantheorylabpixel en rehantheorylab-pixel Tue, 26 May 2026 18:04:02 +0000 https://dev.to/rehantheorylabpixel/i-built-a-subset-sum-solver-35000x-faster-than-previous-approaches-4aki https://dev.to/rehantheorylabpixel/i-built-a-subset-sum-solver-35000x-faster-than-previous-approaches-4aki <p>The subset sum problem is NP-complete. Given integers, find a subset that sums to a target. Simple to state -- exponentially hard to solve.</p> <p>I built a solver that's 35,000x faster than previous approaches. It holds world records across all 65 test categories. First solver to reach n=70 with 1 quadrillion values.</p> <h2> How It Works </h2> <p>23 parallel engines attack the problem simultaneously. The moment any engine finds the answer, all others stop. You fire everything at once and the best one wins.</p> <p><strong>Sum-Range Partitioning</strong>: The key breakthrough. Split the target range into 8 slices, each on its own thread with zero shared state. 6.6x speedup on 8 cores.</p> <p><strong>GDEP Engine</strong>: My newest engine. After picking an element, dynamically restricts the pool to only values smaller than the remainder. Shrinks both goal AND element set simultaneously.</p> <p><strong>Performance</strong>:</p> <ul> <li>n=40: 0.1s</li> <li>n=60: 24s (BCJ would take ~240 hours -- 35,000x faster)</li> <li>n=70: 417s (world record)</li> </ul> <h2> Try It </h2> <p>Live demo: <a href="https://subsetsolver.netlify.app" rel="noopener noreferrer">https://subsetsolver.netlify.app</a><br> GitHub: <a href="https://github.com/rehantheorylab-pixel/35000x-faster-subset-sum-algorithm-n70" rel="noopener noreferrer">https://github.com/rehantheorylab-pixel/35000x-faster-subset-sum-algorithm-n70</a><br> Zenodo DOI: 10.5281/zenodo.20399806</p> <p>Open source. MIT license. Built by Rehan.</p> rehantheorylab-pixel Tue, 26 May 2026 18:04:02 +0000 https://dev.to/rehantheorylabpixel/i-built-a-subset-sum-solver-35000x-faster-than-previous-approaches-31ic https://dev.to/rehantheorylabpixel/i-built-a-subset-sum-solver-35000x-faster-than-previous-approaches-31ic <p>The subset sum problem is NP-complete. Given integers, find a subset that sums to a target. Simple to state -- exponentially hard to solve.</p> <p>I built a solver that's 35,000x faster than previous approaches. It holds world records across all 65 test categories. First solver to reach n=70 with 1 quadrillion values.</p> <h2> How It Works </h2> <p>23 parallel engines attack the problem simultaneously. The moment any engine finds the answer, all others stop. You fire everything at once and the best one wins.</p> <p><strong>Sum-Range Partitioning</strong>: The key breakthrough. Split the target range into 8 slices, each on its own thread with zero shared state. 6.6x speedup on 8 cores.</p> <p><strong>GDEP Engine</strong>: My newest engine. After picking an element, dynamically restricts the pool to only values smaller than the remainder. Shrinks both goal AND element set simultaneously.</p> <p><strong>Performance</strong>:</p> <ul> <li>n=40: 0.1s</li> <li>n=60: 24s (BCJ would take ~240 hours -- 35,000x faster)</li> <li>n=70: 417s (world record)</li> </ul> <h2> Try It </h2> <p>Live demo: <a href="https://subsetsolver.netlify.app" rel="noopener noreferrer">https://subsetsolver.netlify.app</a><br> GitHub: <a href="https://github.com/rehantheorylab-pixel/35000x-faster-subset-sum-algorithm-n70" rel="noopener noreferrer">https://github.com/rehantheorylab-pixel/35000x-faster-subset-sum-algorithm-n70</a><br> Zenodo DOI: 10.5281/zenodo.20399806</p> <p>Open source. MIT license. Built by Rehan.</p>