Research map / Large language models / Agents, code and math
QuArch: A Benchmark for Evaluating LLM Reasoning in Computer Architecture
Shvetank Prakash, Andrew Cheng, Mark Mazumder, Arya Tschand, Varun Gohil, Jeffrey Jian Ma, Jason Yik, Zishen Wan, Jessica Quaye, Elisavet Lydia Alvanaki, Avinash Kumar, Chandrashis Mazumdar and 15 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use reasoning model (GPT-5.2) instead of non-reasoning GPT-5 | GPT · QuArch-Reasoning | accuracy (%): 48.4 -> 73.2 | Improved |
| use reasoning model instead of non-reasoning | GPT · QuArch-Recall | accuracy (%): 86.4 -> 88.4 | No clear effect |
| fine-tune on QuArch memory QAs (3 epochs) | Gemma-3-27B-it · memory hierarchy design task | trials meeting area and energy budget: 6/20 (30%) -> 14/20 (70%) | Improved |
| fine-tune on QuArch memory QAs (3 epochs) | Llama-3.3-70B-Instruct · memory hierarchy design task | trials meeting area and energy budget: 11/20 (55%) -> 19/20 (95%) | Improved |
| fine-tune with more aggressive config (4 epochs, higher LR) | Llama-3.3-70B-Instruct · memory hierarchy design task | best area (um^2): 7246.78 -> 3637.89 | Improved |
| fine-tune with more aggressive config (4 epochs, higher LR) (ablation) | Llama-3.3-70B-Instruct · memory hierarchy design task | success rate: 95% -> 35% (vs Config 1) | Hurt |
| fine-tune with more aggressive config (4 epochs, higher LR) | Gemma-3-27B-it · memory hierarchy design task | best area (um^2): 6728.49 -> 3608.89 | Improved |
| fine-tune on all unfiltered FRQs instead of memory-focused subset (ablation) | Gemma/Llama SFT · memory hierarchy design task | success rate | No clear effect |
| add images (multimodal questions vs text-only) | multimodal frontier models · QuArch | accuracy: about 5% lower on average | Hurt |
| use LLM judge instead of human grading | Claude-3.7-Sonnet-Thinking · 100 sampled FRQs, 925 responses | agreement rate (%): 89.41 vs human-human 90.92 | No clear effect |
Closest papers by meaning
- PCEval: A Benchmark for Evaluating Physical Computing Capabilities of Large Language Models (NeurIPS 2026 · similarity 0.88)
- CS-Bench: A Comprehensive Benchmark for Large Language Models towards Computer Science Mastery (ICLR 2025 · similarity 0.87)
- ArborKV: Structure-Aware KV Cache Management for Scaling Tree-based LLM Reasoning (ICML 2026 · similarity 0.86)
- ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning (ICML 2025 · similarity 0.86)
- Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference (NeurIPS 2025 · similarity 0.86)
- REASONING COMPILER: LLM-Guided Optimizations for Efficient Model Serving (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
