More Like this

1. Extending MAGMA Portability with OneAPI

   https://doi.org/10.1109/WACCPD56842.2022.00008  

   Fortenberry, Anna; Tomov, Stanimire (November 2022, SC 2022 Workshop on Accelerator Programming Using Directives (WACCPD))
2. Pressurizing Magma Within Heterogeneous Crust: A Case Study at the Socorro Magma Body, New Mexico, USA

   https://doi.org/10.1029/2023GL105689  

   Block, Grant A; Roy, Mousumi; Graves, Emily; Grapenthin, Ronni (October 2023, Geophysical Research Letters)

   Abstract Surface deformation plays a key role in illuminating magma transport at active volcanoes, however, unambiguous separation of deep and shallow transport remains elusive. The Socorro Magma Body (SMB) lacks an upper crustal magma transport system, allowing us to link geodetic measurements with predictions of numerical models investigating rheologic heterogeneities and magma‐mush interaction in the mid‐/lower crust. New InSAR observations confirm that a pattern of central surface uplift surrounded by a region of subsidence (previously coined “sombrero” deformation) has persisted over >100 years at the SMB. Our models suggest this pattern may reflect the presence of a large (>100 km width), weaker‐than‐ambient, compliant region (CR) surrounding the mid‐crustal magma body. Interactions between a pressurizing (e.g., due to melt injection and/or volatile exsolution) sill‐like magma body and CR drive the sombrero pattern, depending on both viscoelastic relaxation and pressurization timescales, explaining its rare observation and transient nature. 

   more » « less
3. Seismicity from the deep magma system

   https://doi.org/10.1126/science.abc2452  

   Matoza, Robin S. (May 2020, Science)
4. MagmaDNN: Accelerated Deep Learning Using MAGMA

   https://doi.org/10.1145/3332186.3333047  

   Nichols, Daniel; Wong, Kwai; Tomov, Stan; Ng, Lucien; Chen, Sihan; Gessinger, Alex (January 2019, PEARC19)

   MagmaDNN [17] is a deep learning framework driven using the highly optimized MAGMA dense linear algebra package. The library offers comparable performance to other popular frameworks, such as TensorFlow, PyTorch, and Theano. C++ is used to implement the framework providing fast memory operations, direct cuda access, and compile time errors. Common neural network layers such as Fully Connected, Convolutional, Pooling, Flatten, and Dropout are included. Hyperparameter tuning is performed with a parallel grid search engine. MagmaDNN uses several techniques to accelerate network training. For instance, convolutions are performed using the Winograd algorithm and FFTs. Other techniques include MagmaDNNs custom memory manager, which is used to reduce expensive memory transfers, and accelerated training by distributing batches across GPU nodes. This paper provides an overview of the MagmaDNN framework and how it leverages the MAGMA library to attain speed increases. This paper also addresses how deep networks are accelerated by training in parallel and further challenges with parallelization. 

   more » « less
5. Seismically anisotropic magma reservoirs underlying silicic calderas

   https://doi.org/10.1130/G45104.1  

   Jiang, Chengxin; Schmandt, Brandon; Farrell, Jamie; Lin, Fan-Chi; Ward, Kevin M. (July 2018, Geology)