{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/integrated-hardware-architecture-and-device","title":"Integrated Hardware Architecture and Device Placement Search","arxiv_id":"2407.13143","date":"2024-07-18","proceeding":null,"authors":["Irene Wang","Jakub Tarnawski","Amar Phanishayee","Divya Mahajan"],"abstract":"Distributed execution of deep learning training involves a dynamic interplay between hardware accelerator architecture and device placement strategy. This is the first work to explore the co-optimization of determining the optimal architecture and device placement strategy through novel algorithms, improving the balance of computational resources, memory usage, and data distribution. Our architecture search leverages tensor and vector units, determining their quantity and dimensionality, and on-chip and off-chip memory configurations. It also determines the microbatch size and decides whether to recompute or stash activations, balancing the memory footprint of training and storage size. For each explored architecture configuration, we use an Integer Linear Program (ILP) to find the optimal schedule for executing operators on the accelerator. The ILP results then integrate with a dynamic programming solution to identify the most effective device placement strategy, combining data, pipeline, and tensor model parallelism across multiple accelerators. Our approach achieves higher throughput on large language models compared to the state-of-the-art TPUv4 and the Spotlight accelerator search framework. The entire source code of PHAZE is available at https://github.com/msr-fiddle/phaze.","url_abs":"https://arxiv.org/abs/2407.13143v1","url_pdf":"https://arxiv.org/pdf/2407.13143v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"integrated-hardware-architecture-and-device","repo_url":"https://github.com/msr-fiddle/phaze","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2407.13143","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2407.13143"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/msr-fiddle/phaze","reach":null}],"summary":{"ran_honours":2,"ran":1,"ran_draft_wrong":1,"unverified":12},"by_repo_kind":{"official":{"samples":16,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"ec69c754f196e872","entry":"create_core_config","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ec69c754f196e872"}},{"code_sha256_prefix":"245dff1bccbff4b8","entry":"find_config_across_models","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"245dff1bccbff4b8"}},{"code_sha256_prefix":"73fb0cda316ba202","entry":"generate_out_filename","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"73fb0cda316ba202"}},{"code_sha256_prefix":"a2b8a47f1a78cbfb","entry":"get_num_cores","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a2b8a47f1a78cbfb"}},{"code_sha256_prefix":"88f36ee3e7356cdc","entry":"device_placement","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"88f36ee3e7356cdc"}},{"code_sha256_prefix":"726c74abb2a2fd17","entry":"extract_op_graph_ilp","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"726c74abb2a2fd17"}},{"code_sha256_prefix":"045095c47c56ff37","entry":"get_configs_to_explore","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"045095c47c56ff37"}},{"code_sha256_prefix":"53484ad987afc3c9","entry":"get_next_acc_config","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"53484ad987afc3c9"}},{"code_sha256_prefix":"527a9f1d884b10d7","entry":"ilp","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"527a9f1d884b10d7"}},{"code_sha256_prefix":"1185447c7929792e","entry":"latency_per_layer","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1185447c7929792e"}},{"code_sha256_prefix":"54176a7933685d40","entry":"optimize_latency","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"54176a7933685d40"}},{"code_sha256_prefix":"0f6fdba25526d18b","entry":"run_solver","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0f6fdba25526d18b"}},{"code_sha256_prefix":"2f8e76fed12d4260","entry":"schedule_using_ilp_x_values","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2f8e76fed12d4260"}},{"code_sha256_prefix":"48468011bfdc665a","entry":"solve","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"48468011bfdc665a"}},{"code_sha256_prefix":"b3f18a53011bc087","entry":"sort_acc_configs_by_area","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b3f18a53011bc087"}},{"code_sha256_prefix":"5594ad139855a245","entry":"write_explored_configs_to_file","repo":"msr-fiddle/phaze","repo_kind":"official","path":"Solver/solver.py","file_url":"https://github.com/msr-fiddle/phaze/blob/HEAD/Solver/solver.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"5594ad139855a245"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}