Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size
This work brings the leading accuracy, sample efficiency, and robustness of deep equivariant neural networks to the extreme computational scale. This is achieved through a combination of innovative model architecture, massive parallelization, and models and implementations optimized for efficient GPU utilization. The resulting Allegro architecture bridges the accuracy-speed tradeoff of atomistic simulations and enables description of dynamics in structures of unprecedented complexity at quantum fidelity. To illustrate the scalability of Allegro, we perform nanoseconds-long stable simulations of protein dynamics and scale up to a 44-million atom structure of a complete, all-atom, explicitly solvated HIV capsid on the Perlmutter supercomputer. We demonstrate excellent strong scaling up to 100 million atoms and 70% weak scaling to 5120 A100 GPUs..
| Medienart: |
|
|---|
Preprint| Erscheinungsjahr: |
2023 |
|---|---|
| Erschienen: |
2023 |
| Enthalten in: |
arXiv.org - (2023) vom: 19. Apr. Zur Gesamtaufnahme - year:2023 |
|---|
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Musaelian, Albert [VerfasserIn] |
|---|
| Links: |
Volltext [kostenfrei] |
|---|
| Themen: |
000 |
|---|
| Förderinstitution / Projekttitel: |
|
|---|
| PPN (Katalog-ID): |
XAR039317420 |
|---|
| LEADER | 01000naa a22002652 4500 | ||
|---|---|---|---|
| 001 | XAR039317420 | ||
| 003 | DE-627 | ||
| 005 | 20230421081118.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 230421s2023 xx |||||o 00| ||eng c | ||
| 035 | |a (DE-627)XAR039317420 | ||
| 035 | |a (arXiv)2304.10061 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Musaelian, Albert |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Scaling the leading accuracy of deep equivariant models to biomolecular simulations of realistic size |
| 264 | 1 | |c 2023 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a This work brings the leading accuracy, sample efficiency, and robustness of deep equivariant neural networks to the extreme computational scale. This is achieved through a combination of innovative model architecture, massive parallelization, and models and implementations optimized for efficient GPU utilization. The resulting Allegro architecture bridges the accuracy-speed tradeoff of atomistic simulations and enables description of dynamics in structures of unprecedented complexity at quantum fidelity. To illustrate the scalability of Allegro, we perform nanoseconds-long stable simulations of protein dynamics and scale up to a 44-million atom structure of a complete, all-atom, explicitly solvated HIV capsid on the Perlmutter supercomputer. We demonstrate excellent strong scaling up to 100 million atoms and 70% weak scaling to 5120 A100 GPUs. | ||
| 650 | 4 | |a Physics - Computational Physics |7 (dpeaa)DE-84 | |
| 650 | 4 | |a Computer Science - Machine Learning |7 (dpeaa)DE-84 | |
| 650 | 4 | |a Physics - Chemical Physics |7 (dpeaa)DE-84 | |
| 650 | 4 | |a Quantitative Biology - Biomolecules |7 (dpeaa)DE-84 | |
| 650 | 4 | |a 530 |7 (dpeaa)DE-84 | |
| 650 | 4 | |a 000 |7 (dpeaa)DE-84 | |
| 650 | 4 | |a 570 |7 (dpeaa)DE-84 | |
| 700 | 1 | |a Johansson, Anders |4 aut | |
| 700 | 1 | |a Batzner, Simon |4 aut | |
| 700 | 1 | |a Kozinsky, Boris |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t arXiv.org |g (2023) vom: 19. Apr. |
| 773 | 1 | 8 | |g year:2023 |g day:19 |g month:04 |
| 856 | 4 | 0 | |u https://arxiv.org/abs/2304.10061 |z kostenfrei |3 Volltext |
| 912 | |a GBV_XAR | ||
| 951 | |a AR | ||
| 952 | |j 2023 |b 19 |c 04 | ||