ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation | Paper Citation Record

Observation 2fe4b56d-d998-43f7-bf3b-575b126d460a · outbound

This paper cites OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.534575Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5190c8f8-1df2-482f-8c9c-9580430dcdab · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.116087Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0b223bd6-5af5-4627-80f5-5cc7748656c7 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.296539Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 205de562-cada-42a8-a44d-1e83e3daaa55 · outbound

This paper cites If any expected target file cannot be recovered or the extracted content is empty after all tiers, the task is classified as EXTRACT_FAIL.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation If any expected target file cannot be recovered or the extracted content is empty after all tiers, the task is classified as EXTRACT_FAIL

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.427678Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bb75f3b0-cd6c-4141-9189-1fbc2511cc42 · outbound

This paper cites Elias Konstantinidis and Yiannis Cotronis.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Elias Konstantinidis and Yiannis Cotronis

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.688796Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 17169d96-ea01-42ac-8ad6-158118db3c67 · outbound

This paper cites Code files are included up to a cumulative 50,000-character limit.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Code files are included up to a cumulative 50,000-character limit

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.034522Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8699f4df-2d4c-4152-b934-380836c30b74 · outbound

This paper cites This is the highest-confidence tier.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation This is the highest-confidence tier

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.961734Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5c79254e-e438-4bc7-bbb6-f38de146c0ef · outbound

This paper cites H.3 Valid Claims The following conclusions can be drawn from PARBENCHresults when accompanied by the stated conditions:.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation H.3 Valid Claims The following conclusions can be drawn from PARBENCHresults when accompanied by the stated conditions:

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.333194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ddaacc60-6a2d-410f-87ef-1d9705e0bb53 · outbound

This paper cites the model understands parallel programming.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation the model understands parallel programming

Reference 9

Resolution: malformed identifier; no resolver link, observed 2026-08-02T11:55:24.720497Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cc83510e-8000-410b-ab1f-04986fd49145 · outbound

This paper cites While repository- level co-occurrence with CUDA and OpenMP is high, the kernel-level material is insufficient for the multi-direction evaluation design that PARBENCHrequires.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation While repository- level co-occurrence with CUDA and OpenMP is high, the kernel-level material is insufficient for the multi-direction evaluation design that PARBENCHrequires

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.558812Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b9ef4701-c0dd-49e3-aa12-8abcdbf0b336 · outbound

This paper cites Including both would provide diminishing returns in programming-model diversity.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Including both would provide diminishing returns in programming-model diversity

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.736325Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b9100052-8288-4660-a581-451cec8279dd · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.872604Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 24d85293-f70d-4fa5-b594-5694c4210cf8 · outbound

This paper cites Default GCC and Clang installations support only CPU-threaded OpenMP; GPU offloading requires custom builds with target-offload support enabled.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Default GCC and Clang installations support only CPU-threaded OpenMP; GPU offloading requires custom builds with target-offload support enabled

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:18.996248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a90be6c7-16cd-484e-bd65-d97d62d998c0 · outbound

This paper cites Within PARBENCH’s curated corpus, OpenMP target implementations are available for 12 kernels: 10 from HeCBench, 1 from XSBench, and 1 from RSBench.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Within PARBENCH’s curated corpus, OpenMP target implementations are available for 12 kernels: 10 from HeCBench, 1 from XSBench, and 1 from RSBench

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.089166Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 67b7b7f8-f4eb-4916-8139-880110dad00c · outbound

This paper cites available parallel code.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation available parallel code

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.162112Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 47b6aebb-5ba2-4c4c-a4c7-e8daf43de173 · outbound

This paper cites This is the dominant build-time failure mode for CUDA-targeting directions across all three models.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation This is the dominant build-time failure mode for CUDA-targeting directions across all three models

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.237246Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7848378b-7c4d-4981-9835-212f56aac835 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.313659Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 783e363f-b499-41fd-91c6-3d85247fa567 · outbound

This paper cites CUDA”, “OpenMP.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation CUDA”, “OpenMP

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.419837Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f5d6c1cf-fb8a-425d-a40e-aa50c5c947cc · outbound

This paper cites The kernel name and benchmark description arenotincluded (anonymization).

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The kernel name and benchmark description arenotincluded (anonymization)

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.575344Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c051e371-4501-4117-b8a2-33da05abd59f · outbound

This paper cites When target infrastructure context is provided, an explanatory note clarifies that only these files replace existing project files.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation When target infrastructure context is provided, an explanatory note clarifies that only these files replace existing project files

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.647006Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ecf884d8-0bcc-4bbe-b29c-c6eebd40672f · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.724997Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c2e0565a-748e-49f7-a874-b3e32af3811d · outbound

This paper cites CUDA Toolkit ≥ 11.0.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation CUDA Toolkit ≥ 11.0

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.812312Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fb5c2cfe-0a62-4ba1-b15a-931a11e4344b · outbound

This paper cites Source Code (CUDA).

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Source Code (CUDA)

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:19.923648Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 50c466a6-a12f-4ee5-a87e-4e096925aca0 · outbound

This paper cites DO NOT MODIFY – for reference only.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation DO NOT MODIFY – for reference only

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.149419Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 97f89d0a-ba66-4a58-b86a-33a0cd8b40ac · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.316635Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d0910c8a-dded-4550-b00a-9b64078e45a0 · outbound

This paper cites The stripper uses a state-machine parser that preserves string, character, and raw string literals.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The stripper uses a state-machine parser that preserves string, character, and raw string literals

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.412215Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation feca387b-a762-4fbd-9330-d336caaac09b · outbound

This paper cites Support files are labeled Header FileN or Code FileN by extension.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Support files are labeled Header FileN or Code FileN by extension

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.486710Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 948535bc-f413-49d6-8302-a63583b68e24 · outbound

This paper cites An internal mapping restores original filenames when writing LLM output to disk, including any cross-file#includereferences the LLM emits using the generic names.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation An internal mapping restores original filenames when writing LLM output to disk, including any cross-file#includereferences the LLM emits using the generic names

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.566026Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ad793202-62c2-4a3a-8fce-5cfa7db40b83 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.673751Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bb4bd823-80a1-4064-9790-c5ee32ff992a · outbound

This paper cites CUDA”, “OpenMP.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation CUDA”, “OpenMP

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.746659Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9bbc56c4-4e17-45ab-926e-7e0407095799 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.798866Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f1dad8c2-ce97-425e-9485-0b48e145fb0c · outbound

This paper cites The evaluation matrix covers 142 unique source–target pairs (Section 2.3).

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The evaluation matrix covers 142 unique source–target pairs (Section 2.3)

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.863342Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 18c7562c-4acb-49ae-afc7-484cc20bba00 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:20.959789Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 669e0a64-d1cc-43cb-91ab-68797e3ecb0b · outbound

This paper cites The augmentation campaign applies only to direction–kernel pairs that qualify under the L0-conditional filter.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The augmentation campaign applies only to direction–kernel pairs that qualify under the L0-conditional filter

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.070692Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 559f60c4-4c4e-4efa-aa8f-79424fea03ee · outbound

This paper cites CUDA-to-OpenMP consistently passes at higher rates than OpenCL-to-CUDA across all three evaluated models (Section 5.3).

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation CUDA-to-OpenMP consistently passes at higher rates than OpenCL-to-CUDA across all three evaluated models (Section 5.3)

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.146536Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b61208cf-0c78-42bb-8702-75ff15cbe972 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.221596Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8c17564b-7f98-4a60-947a-0603c7c8224f · outbound

This paper cites H.2 What PARBENCHDoes Not Measure The following properties are explicitly outside PARBENCH’s measurement scope:.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation H.2 What PARBENCHDoes Not Measure The following properties are explicitly outside PARBENCH’s measurement scope:

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.300629Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6a3ec0e3-c06a-40f7-950b-f1af01b73bc0 · outbound

This paper cites Wall-clock time is captured but is un- reliable for sub-millisecond baselines.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Wall-clock time is captured but is un- reliable for sub-millisecond baselines

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.365909Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b0176628-e9ce-42fc-b01b-f744d4b172f6 · outbound

This paper cites A harness-passing but unreadable translation receives PASS.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation A harness-passing but unreadable translation receives PASS

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.442146Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d267ff2-c598-4c9e-92dd-ad0a996d6407 · outbound

This paper cites There is no partial-credit classifica- tion.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation There is no partial-credit classifica- tion

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.506820Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 60b69df6-87d3-4708-8ce1-694f5841ec82 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.635694Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8b932632-295b-4a2f-8979-bcfa44b0dd69 · outbound

This paper cites The evaluation pipeline implements iterative repair (multi-turn error feedback with linker analysis via --max-retries), but all reported results use max_retries=1 (zero-shot).

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The evaluation pipeline implements iterative repair (multi-turn error feedback with linker analysis via --max-retries), but all reported results use max_retries=1 (zero-shot)

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.784948Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5b734ac0-448c-4f4f-ba4a-351c89619170 · outbound

This paper cites No testing on other GPU architectures (A100, H100), CPU-only environments, or other OS configurations.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation No testing on other GPU architectures (A100, H100), CPU-only environments, or other OS configurations

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:21.942691Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ccf7ce7c-4fed-4e9d-9f65-f3642352f134 · outbound

This paper cites Host code, Makefiles, and I/O routines remain fixed.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Host code, Makefiles, and I/O routines remain fixed

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.134566Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 10c2698a-aeb0-4467-9320-de33d45fe7ed · outbound

This paper cites Model X achieves Y% pass@k on kernel-centric parallel code translation across Z directions.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X achieves Y% pass@k on kernel-centric parallel code translation across Z directions

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.488911Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e042c835-1c78-4827-98da-753e67ce1366 · outbound

This paper cites Build-stage failure is the dominant failure mode for Model X.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Build-stage failure is the dominant failure mode for Model X

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.577667Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb0e260d-9bd8-4b37-8439-8c09bf1d25ff · outbound

This paper cites Direction A→B is harder/easier than Direction C→D for Model X.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Direction A→B is harder/easier than Direction C→D for Model X

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.737640Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 434f465b-094e-4128-a8a1-5f09e9186f8f · outbound

This paper cites Model X maintains declared-oracle pass rates across L1–L4 on the L0-conditional subset.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X maintains declared-oracle pass rates across L1–L4 on the L0-conditional subset

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.834685Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2b5bf37c-b286-4f42-8cc2-e324b1adee29 · outbound

This paper cites Model X discriminates from Model Y on PARBENCH.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X discriminates from Model Y on PARBENCH

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.927867Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 164624fa-f6c5-483f-b5cd-dfd71d2fa76a · outbound

This paper cites Kernel K is harder to translate than Kernel J.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Kernel K is harder to translate than Kernel J

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:22.998703Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 28bb6ee6-1747-41ac-8f24-19a769ba0aba · outbound

This paper cites Model X produces efficient/fast parallel code.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X produces efficient/fast parallel code

Reference 54

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.100836Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 74871600-3aaa-479d-a6c0-7f0ab8bb3b36 · outbound

This paper cites Model X’s translations are production-ready.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X’s translations are production-ready

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.166504Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 971b2d41-1982-4229-a87a-652b6f68610f · outbound

This paper cites Model X understands parallel programming.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X understands parallel programming

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.328548Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8de83f10-ff34-4457-8129-f72f5e95c3e4 · outbound

This paper cites PARBENCHpass rate generalizes to all HPC translation tasks.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation PARBENCHpass rate generalizes to all HPC translation tasks

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.469755Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0eb0c4b1-f487-4ca6-802f-c396590896e6 · outbound

This paper cites Model X is definitively better than Model Y at parallel translation.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Model X is definitively better than Model Y at parallel translation

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.601797Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3cac358f-b1e0-48f7-a1fe-778f85eed179 · outbound

This paper cites Surface-form robustness proves the model is not memorizing.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Surface-form robustness proves the model is not memorizing

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.755954Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 343d7832-8109-416a-a967-d0dbae300739 · outbound

This paper cites A PASSmeans the translation is numerically faithful.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation A PASSmeans the translation is numerically faithful

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:23.943815Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation db3428d0-2d76-4a5f-b33d-55e1e43fb5aa · outbound

This paper cites OpenCL→CUDA is impossible for LLMs.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation OpenCL→CUDA is impossible for LLMs

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.158881Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 461c9b99-6cb8-4871-94ea-41eb6b3ff6ce · outbound

This paper cites If a source implemen- tation has a latent bug, faithful translations of that bug would pass the declared oracle.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation If a source implemen- tation has a latent bug, faithful translations of that bug would pass the declared oracle

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.251473Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 185bbd3c-791d-4249-9479-f7f1ab9f58a4 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.313425Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5df7baf1-ae1d-4502-ba25-dcc2111efc92 · outbound

This paper cites Different GPU architectures may have different compiler behavior, runtime characteristics, or memory limits.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Different GPU architectures may have different compiler behavior, runtime characteristics, or memory limits

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.373518Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 18a2572c-ecc2-44a6-b9b1-3b43ecf3e272 · outbound

This paper cites The pass@1-to-pass@3 gap (11.3 pp for Qwen 3.5, 7.0 pp for GPT-5.4, 5.6 pp for GPT-5.3-codex) suggests moderate within-task variance.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation The pass@1-to-pass@3 gap (11.3 pp for Qwen 3.5, 7.0 pp for GPT-5.4, 5.6 pp for GPT-5.3-codex) suggests moderate within-task variance

Reference 65

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.419476Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 77d9061f-1496-4e56-a7a7-40f716bfb7f2 · outbound

This paper cites ParEval-Repo’s 0% at>133 SLoC Davis et al.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation ParEval-Repo’s 0% at>133 SLoC Davis et al

Reference 66

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.471586Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 14570b89-f011-413e-86b9-6e53c70dbff2 · outbound

This paper cites an unresolved cited work.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Unresolved cited work

Reference 67

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.525577Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a1128190-80b5-463c-9ab8-7f71491ad546 · outbound

This paper cites [2009], HeCBench: 2023 Jin and Vetter [2023], XSBench: ANL Tramm et al.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation [2009], HeCBench: 2023 Jin and Vetter [2023], XSBench: ANL Tramm et al

Reference 68

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.592724Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8807c898-22e8-4cbe-a543-193e04205f95 · outbound

This paper cites 80 of 87 specs pass all L1–L4 augmented baseline verification.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation 80 of 87 specs pass all L1–L4 augmented baseline verification

Reference 69

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:24.663504Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 551e368f-295a-48aa-9455-2ed9748943f7 · outbound

This paper cites ide nt it y.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation ide nt it y

Reference 2015

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.818807Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 561fea2d-8bf6-42db-8d3d-f2efaa613cd4 · outbound

This paper cites ISBN 978-3-031-69576-6.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation ISBN 978-3-031-69576-6

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.110434Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation afd5e86c-e6e7-4dc2-962a-53077a875bef · outbound

This paper cites ISBN 9798400720741.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation ISBN 9798400720741

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.193540Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 215744db-5338-46ae-a023-4dcd22c19d6c · outbound

This paper cites Carlos E Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik R Narasimhan.

ParBench: A Benchmark for Reliable Evaluation of LLM Parallel Code Translation Carlos E Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik R Narasimhan

Reference 2026

Resolution: unresolved; no resolver link, observed 2026-08-02T11:55:17.376460Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers