List your company
ResearchFormerly WizWand

Efficiency Evaluation on Efficiency Profiling Workload G.7 Detailed Efficiency Results

13.75End-to-End Latency (s)LLMBoost
13.393615.799318.20520.6107Dec 26, 2025
Updated 8mo ago

Evaluation results

MethodLinks
LLMBoostModel Composition=7B+7B, Execution Strategy=near-parallel, Number of GPUs=22025.1213.750.07533.1820.72
LLMBoostModel Composition=7B+7B, Execution Strategy=near-parallel, Number of GPUs=12025.1214.870.08154.1139.75
T-CopilotModel Composition=7B+7B2025.1217.210.09764.4142.05
Qwen14BModel Composition=14B2025.1217.40.09856.7436.97
LLMBoostModel Composition=7B+7B+7B, Execution Strategy=near-parallel, Number of GPUs=32025.1218.880.10333.1820.13
LLMBoostModel Composition=7B+7B+7B, Execution Strategy=near-parallel, Number of GPUs=12025.1221.670.11873.4258.48
LLMBoostModel Composition=7B+7B, Execution Strategy=sequential2025.1222.530.12560.2938.72
VOTEModel Composition=7B+7B2025.1222.610.12753.3939.14
UNITEModel Composition=7B+7B2025.1222.660.12853.3939.12