[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-math-pins-down-gpu-ratios-for-split-llm-serving":10,"sections":45},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":35,"tags":36,"sources":40,"feedback":44,"feedback_at":22,"cost_usd":44,"total_tokens":44},5579,"new-math-pins-down-gpu-ratios-for-split-llm-serving","New Math Pins Down GPU Ratios for Split LLM Serving","A new arXiv paper offers a closed-form rule for balancing Attention and FFN GPUs in disaggregated LLM serving, cutting guesswork and idle time.","A paper posted to arXiv, [arXiv:2601.21351](https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.21351), gives LLM infrastructure teams a formula for one of inference serving's most expensive guessing games: how many GPUs to spend on memory versus compute.\n\nThe paper studies attention-FFN disaggregation, an architecture that splits LLM decoding into two separate jobs. Attention workers hold the growing KV cache and handle state-heavy lookups; FFN workers do the stateless, compute-heavy math; the two communicate every step. That split lets operators scale memory and compute independently, but performance hinges on getting the ratio of attention to FFN workers right, since a bad ratio leaves workers idle waiting on the slowest step. The authors model the randomness in prompt lengths, decode lengths, and cross-worker synchronization, then derive a closed-form rule for the optimal ratio.\n\nCompanies running inference at scale usually tune these ratios by trial and error, burning GPU-hours on test runs or over-provisioning as insurance. A formula that predicts the right split from a single measurable workload statistic could turn that guesswork into a calculation, which matters when idle GPU time is a direct cost.\n\nWorth noting: the validation here comes from a trace-calibrated simulator that matched the true optimum within 10 percent, not a live production fleet, so the real test is whether the math holds up against the messiness of an actual serving cluster.","[\"llm-serving\",\"gpu-scaling\",\"ai-infrastructure\"]","2026-08-18T04:00:00.000Z","2026-08-19T02:10:46.306Z","2026-08-19T02:10:57.952Z","published",null,[24,30],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"Remove or verify the specific claim that the paper was 'updated August 18, 2026' — no such date appears in the source material, only the fact that it's v4 of the arXiv listing, so this specific date is an unsupported invented detail.","resolved",{"id":31,"reviewer":32,"round":33,"reason":34,"status":29},"publisher-r2","publisher",2,"The article never names the paper, its authors, or a source\u002Flink, so the central claims cannot be verified as attributable and factual.","ai",[37,38,39],"llm-serving","gpu-scaling","ai-infrastructure",[41],{"name":42,"url":43},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2601.21351",0,{"sections":46},[47,51,55,60,65,70,75,80,85,89,94,99,104,109],{"name":48,"slug":35,"count":49,"latest_published_at":50},"AI",3293,"2026-08-20T04:00:00.000Z",{"name":52,"slug":53,"count":54,"latest_published_at":50},"Security","security",435,{"name":56,"slug":57,"count":58,"latest_published_at":59},"Policy","policy",210,"2026-08-19T09:32:27.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Hardware","hardware",140,"2026-08-19T18:25:42.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Consumer Tech","consumer-tech",95,"2026-08-18T16:05:00.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Science","science",90,"2026-08-19T18:41:02.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Software","software",73,"2026-08-18T07:51:50.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":18},"Dev Tools","dev-tools",69,{"name":90,"slug":91,"count":92,"latest_published_at":93},"Startups","startups",47,"2026-08-19T19:13:46.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"General","general",33,"2026-08-18T22:18:13.000Z",{"name":105,"slug":106,"count":107,"latest_published_at":108},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":110,"slug":111,"count":112,"latest_published_at":113},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]