{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2XD6IYPJEFF2VXT462REPMZOZF","short_pith_number":"pith:2XD6IYPJ","schema_version":"1.0","canonical_sha256":"d5c7e461e9214baade7cf6a247b32ec95c5d70b08586b611863c41278fb9f618","source":{"kind":"arxiv","id":"2504.01994","version":1},"attestation_state":"computed","paper":{"title":"PIM-LLM: A High-Throughput Hybrid PIM Architecture for 1-bit LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.AR","authors_text":"Jinendra Malekar, Md Hasibul Amin, Mohammed E. Elbtity, Peyton Chandarana, Ramtin Zand","submitted_at":"2025-03-31T21:42:43Z","abstract_excerpt":"In this paper, we propose PIM-LLM, a hybrid architecture developed to accelerate 1-bit large language models (LLMs). PIM-LLM leverages analog processing-in-memory (PIM) architectures and digital systolic arrays to accelerate low-precision matrix multiplication (MatMul) operations in projection layers and high-precision MatMul operations in attention heads of 1-bit LLMs, respectively. Our design achieves up to roughly 80x improvement in tokens per second and a 70% increase in tokens per joule compared to conventional hardware accelerators. Additionally, PIM-LLM outperforms previous PIM-based LL"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.01994","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AR","submitted_at":"2025-03-31T21:42:43Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"5c69cf891cfb8cb73d13bb5abead3417443de628d24a5bad5ab263116473d81e","abstract_canon_sha256":"864b6b7367ad5f43a17185cdd706c894245e1996b9dbaf585113d627140f57d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:43:48.499219Z","signature_b64":"57rQjcMsg28oH8XOw1QwoeNZfEJb4NeSuhCoEMLIL4m8fkcN6CC6X+fCmyTQw1jZH4KYIAB/DRoD0qWx8SyUCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d5c7e461e9214baade7cf6a247b32ec95c5d70b08586b611863c41278fb9f618","last_reissued_at":"2026-07-05T10:43:48.498751Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:43:48.498751Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PIM-LLM: A High-Throughput Hybrid PIM Architecture for 1-bit LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.AR","authors_text":"Jinendra Malekar, Md Hasibul Amin, Mohammed E. Elbtity, Peyton Chandarana, Ramtin Zand","submitted_at":"2025-03-31T21:42:43Z","abstract_excerpt":"In this paper, we propose PIM-LLM, a hybrid architecture developed to accelerate 1-bit large language models (LLMs). PIM-LLM leverages analog processing-in-memory (PIM) architectures and digital systolic arrays to accelerate low-precision matrix multiplication (MatMul) operations in projection layers and high-precision MatMul operations in attention heads of 1-bit LLMs, respectively. Our design achieves up to roughly 80x improvement in tokens per second and a 70% increase in tokens per joule compared to conventional hardware accelerators. Additionally, PIM-LLM outperforms previous PIM-based LL"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.01994","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.01994/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2504.01994","created_at":"2026-07-05T10:43:48.498807+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.01994v1","created_at":"2026-07-05T10:43:48.498807+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.01994","created_at":"2026-07-05T10:43:48.498807+00:00"},{"alias_kind":"pith_short_12","alias_value":"2XD6IYPJEFF2","created_at":"2026-07-05T10:43:48.498807+00:00"},{"alias_kind":"pith_short_16","alias_value":"2XD6IYPJEFF2VXT4","created_at":"2026-07-05T10:43:48.498807+00:00"},{"alias_kind":"pith_short_8","alias_value":"2XD6IYPJ","created_at":"2026-07-05T10:43:48.498807+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2505.05772","citing_title":"Sparse Attention Remapping with Clustering for Efficient LLM Decoding on PIM","ref_index":35,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF","json":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF.json","graph_json":"https://pith.science/api/pith-number/2XD6IYPJEFF2VXT462REPMZOZF/graph.json","events_json":"https://pith.science/api/pith-number/2XD6IYPJEFF2VXT462REPMZOZF/events.json","paper":"https://pith.science/paper/2XD6IYPJ"},"agent_actions":{"view_html":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF","download_json":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF.json","view_paper":"https://pith.science/paper/2XD6IYPJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.01994&json=true","fetch_graph":"https://pith.science/api/pith-number/2XD6IYPJEFF2VXT462REPMZOZF/graph.json","fetch_events":"https://pith.science/api/pith-number/2XD6IYPJEFF2VXT462REPMZOZF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF/action/storage_attestation","attest_author":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF/action/author_attestation","sign_citation":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF/action/citation_signature","submit_replication":"https://pith.science/pith/2XD6IYPJEFF2VXT462REPMZOZF/action/replication_record"}},"created_at":"2026-07-05T10:43:48.498807+00:00","updated_at":"2026-07-05T10:43:48.498807+00:00"}