AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

MegBA: A GPU-Based Distributed Library for Large-Scale Bundle Adjustment

Jie Ren, Wenteng Liang, Ran Yan, Luo Mai, Shiwen Liu, Xiao Liu

ECCV 2022 · Architectures and efficiency › RNNs and normalization · approach: Distributed LLM inference systems · 26 citations (OpenAlex)

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 changedWhereMeasuredEffect
increase GPUs from 1 to 8 (ablation)MegBA (analytic) · BAL Venice-1778BA time (s): 11.96 -> 3.340Improved
increase GPUs from 4 to 8 (ablation)MegBA (analytic) · BAL Final-13682BA time (s): 28.70 -> 22.10Improved
use MegBA 8 GPUs instead of CeresCeres-16 · BAL Final-13682BA time (s): 916.0 -> 22.10Improved
use MegBA 8 GPUs instead of DeepLMDeepLM · BAL Final-13682BA time (s): 149.6 -> 22.10Improved
use MegBA instead of RootBARootBA-16 · BAL Final-13682BA time (s): 1,427 -> 22.10Improved
use MegBA single GPU instead of DeepLMDeepLM · BAL Venice-1778BA time (s): 24.44 -> 11.96Improved
use 32-bit instead of 64-bit floating point (ablation)MegBA-4-a · BAL Final-13682BA time (s): 28.70 -> 4.804Improved
use 32-bit instead of 64-bit floating point (ablation)MegBA-1-a · BAL Venice-1778BA time (s): 11.96 -> 2.620Improved
use 32-bit instead of 64-bit floating point (ablation)MegBA-8-a · BAL Venice-1778BA time (s): 3.340 -> 1.622Improved
use 32-bit instead of 64-bit floating point (ablation)MegBA-8-m · BAL Final-13682BA time (s): 20.68 -> 7.582Improved
use 32-bit instead of 64-bit floating point (ablation)MegBA · BAL Venice-1778 and Final-13682MSE: 0.333 -> 0.333No clear effect
MegBA 8 GPUs on 80M-observation synthetic dataset instead of CeresCeres · Synthesised-20000BA time (s): 216.26Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy