Search NASA⌕ Search

Engineering topics

Wang, Tianqi

Publications and source records attributed to Wang, Tianqi.

O3BNN-R: An Out-Of-Order Architecture for High-Performance and Regularized BNN Inference

Binarized Neural Networks (BNN) have drawn tremendous attention due to significantly reduced computational complexity and memory demand. They have especially shown great potential in cost- and power-restricted domains, such as IoT and smart edge-devices, where reaching a certain accuracy bar is often sufficient, and real-time is highly desired.In this work, we demonstrate that the highly-condensed BNN model can be shrunk significantly further by dynamically pruning irregular redundant edges. Based on two new observations on BNN-specific properties, an out-of-order (OoO) architecture – O3BNN-R, can curtail edge evaluation in cases where the binary output of a neuron can be determined early. Similar to Instruction-Level-Parallelism(ILP), these fine-grained, irregular, runtime pruning opportunities are traditionally presumed to be difficult to exploit. In order to increase the pruning opportunities, we also optimize the training process by adding 2 regularization items in the loss function (1) for pooling pruning and (2) for threshold pruning. We evaluate our design on an FPGA platform using three well-known networks, including VggNet-16, AlexNet for ImageNet, and a VGG-like network for Cifar-10.

Geng, Tong↗

Self-Assembled Periodic Nanostructures Using Martensitic Phase Transformations

We describe a novel approach for the rational design and synthesis of self-assembled periodic nanostructures using martensitic phase transformations. We demonstrate this approach in a thin film of perovskite SrSnO 3 with reconfigurable periodic nanostructures consisting of regularly spaced regions of sharply contrasted dielectric properties. The films can be designed to have different periodicities and relative phase fractions via chemical doping or strain engineering. The dielectric contrast within a single film can be tuned using temperature and laser wavelength, effectively creating a variable photonic crystal. Our results show the realistic possibility of designing large-area self-assembled periodic structures using martensitic phase transformations with the potential of implementing "built-to-order" nanostructures for tailored optoelectronic functionalities.

77 NANOSCIENCE AND NANOTECHNOLOGY↗

Precursor selection in hybrid molecular beam epitaxy of alkaline-earth stannates

One of the challenges of oxide molecular beam epitaxy (MBE) is the synthesis of oxides containing metals with high electronegativity (metals that are hard to oxidize). The use of reactive organometallic precursors can potentially address this issue. To investigate the formation of radicals in MBE, we explored three carefully chosen metal-organic precursors of tin for SnO 2 and BaSnO 3 growth: tetramethyltin (TMT), tetraethyltin (TET), and hexamethylditin (HMDT). All three precursors produced single-crystalline, atomically smooth, and epitaxial SnO 2 (101) films on r-Al 2 O 3 (101¯2) in the presence of oxygen plasma. The study of growth kinetics revealed reaction-limited and flux-limited regimes except for TET, which also exhibited a decrease in the deposition rate with increasing temperature above ~800 °C. Contrary to these similarities, the performance of these precursors was dramatically different for BaSnO 3 growth. TMT and TET were ineffective in supplying adequate tin, whereas HMDT yielded phase-pure, stoichiometric BaSnO 3 films. Significantly, HMDT resulted in phase-pure and stoichiometric BaSnO 3 films even without the use of an oxygen plasma (i.e., with molecular oxygen alone). Furthermore, these results are discussed using the ability of HMDT to form tin radicals and therefore assisting with Sn → Sn 4+ oxidation reaction. Structural and electronic transport properties of films grown using HMDT with and without oxygen plasma are compared. This study provides guideline for the choice of precursors that will enable the synthesis of metal oxides containing hard-to-oxidize metals using reactive radicals in MBE.

36 MATERIALS SCIENCE↗

AWB-GCN: A Graph Convolutional Network Accelerator with Runtime Workload Rebalancing

The recent development of deep learning has been mostly focusing on Euclidean data, such as images, videos, audios, etc. However, most real-world information and relation are often expressed as graphs. To efficiently learn from graph data, graph convolutional networks (GCNs) emerge as a promising approach, showing advantages in several practical applications such as social network analysis, knowledge discovery, 3D modeling, motion capturing, etc. Real-world graphs are usually extremely large and imbalanced, posting significant performance demand and design challenges on the hardware dedicated for GCN inference. In this paper, we propose an architecture design called UW-GCN to accelerate graph convolutional network inference. To tackle the major performance bottleneck from workload imbalance, we propose dynamic neighborhood stealing and remote chunk shuffling techniques, relying on hardware flexibility to achieve hardware auto-tuning under negligible area or delay overhead. Specifically, UW-GCN is able to smartly profile the sparse graph pattern while continuously adjusting the workload distribution via routing reconfiguration among parallel processing elements (PEs). The ideal configuration is then reused in the remaining iterations. To the best of our knowledge, this is the first accelerator design particularly for GCN and the first work relying on hardware auto-tuning, which is normally based on software, to achieve near-optimal workload balance in processing sparse structures.

Geng, Tong↗

FPDeep: Scalable Acceleration of CNN Training on Deeply-Pipelined FPGA Clusters

In this paper, we propose a framework called FPDeep, which uses a hybrid of model and layer paral- lelism to configure distributed reconfigurable clusters to train DNNs. This approach has numerous benefits. First, the design does not suffer from batch size growth. Second, novel workload and weight partitioning leads to balanced loads of both among nodes. And third, the entire system is fine-grained pipeline. This leads to high parallelism and utilization and also minimizes the time features need to be cached while waiting for back-propagation.

Wang, Tianqi↗

BSTC: A Novel Binarized-Soft-Tensor-Core Design for Accelerating Bit-Based Approximated Neural Nets

Binarized neural network (or BNN) promises tremendous performance improvement over traditional DNNs through simplified bit-level computation and significantly reduced memory access/storage cost. In addition, it has other advantages of low-cost, low-energy and high-robustness, showing great utilization potential in resources-constrained, volatile and latency-critical applications, which are critical for future HPC execution. However, the promised significant performance gain of BNN inference has never been fully demonstrated on general-purpose processors, particularly on GPUs, due to: (i) the challenge to extract and leverage sufficient fine-grained bit-level-parallelism to saturate GPU cores when batch is small; (ii) the fundamental design conflict between bit-based BNN algorithm and underlying word-based architecture; (iii) architecture \& performance unfriendly BNN network design. To address (i) and (ii), we propose binarized-soft-tensor-core as a software-hardware codesign approach to construct bit-manipulation capability for modern GPUs to effectively harvest the emerging bit-level-parallelism. To tackle (iii), we propose intra- and inter-layer fusion techniques so that the entire BNN inference can be packed into a single GPU kernel, to avoid high-cost frequent launching. Experiments demonstrate that our design can achieve over 1000x speedup for raw inference latency and 10x for inference throughput over the state-of-the-art full-precision simulated BNN inference for AlexNet on ImageNet.

Li, Ang↗

LP-BNN: Ultra-low-Latency BNN Inference with Layer Parallelism

High inference latency seriously limits the deployment of DNNs in real-time domains such as autonomous driving, robotic control, and many others. To address this emerging challenge, researchers have proposed approximate DNNs with reduced precision, e.g., Binarized Neural Networks (BNNs). While BNNs can be built to have little loss in accuracy, latency reduction still has much room for improvement. In this paper, we propose a single-FPGA-based BNN accelerator that achieves microsecond-level ultra-low-latency inference of ImageNet, LP- BNN. We obtain this performance via several design optimizations. First, we optimize the network structure by removing Batch Normalization (BN) functions which leads to significant latency in BNNs without any loss on accuracy. Second, we propose a parameterized architecture which is based on layer parallelism and supports nearly perfect load balancing for various types of BNNs. Third, we fuse all the convolution layers and the first fully connected layer. We process them in parallel through fine- grained inter-layer pipelining. With our proposed accelerator, the inference of binarized AlexNet, VGGNet, and ResNet are completed within 21.5µs, 355µs, and 67.8µs respectively, with no loss in accuracy as compared with other BNN implementations.

Geng, Tong↗

O3BNN: An Out-Of-Order Architecture for High-Performance Binarized Neural Network Inference with Fine-Grained Pruning

In this work, we demonstrate that the highly-condensed BNN model can be shrunk significantly further by dynamically pruning irregular redundant edges. Based on two new observations on BNN-specific properties, our out-of-order (OoO) architecture – O3BNN, can curtail remaining edge evaluation in cases where the binary output of a neuron can be determined early. Similar to Instruction-Level-Parallelism (ILP), these fine-grained, irregular, runtime pruning opportunities are traditionally presumed to be di cult to exploit. We evaluate our design on an FPGA platform using three well-known networks, including VggNet-16, AlexNet for ImageNet, and a VGG-like network for Cifar-10. Results show that our out-of-order approach can prune 27%, 16%, and 42% of the operations for the three networks respectively, without any accuracy loss, leading to at least 1.7×, 1.5×, and 2.1× speedups over state-of-the-art implementations on FPGA/GPU/CPU BNN implementations. Our approach is inference runtime pruning, so no retrain or fine-tuning is needed. We demonstrate our design on an FPGA platform. However, this is only for showcasing the method; the approach does not rely on any FPGA-specific features and can thus be adopted by other devices as well.

Geng, Tong↗

Engineering SrSnO 3 Phases and Electron Mobility at Room Temperature Using Epitaxial Strain

High-speed electronics require epitaxial films with exceptionally high carrier mobility at room temperature (RT). Alkaline-earth stannates with high RT mobility show outstanding prospects for oxide electronics operating at ambient temperatures. However, despite significant progress over the last few years, mobility in stannate films has been limited by dislocations because of the inability to grow fully coherent films. Here, we demonstrate the growth of coherent, strain-engineered phases of epitaxial SrSnO 3 (SSO) films using a radical-based molecular beam epitaxy approach. Compressive strain stabilized the high-symmetry tetragonal phase of SSO at RT, which, in bulk, exists only at temperatures between 1062 and 1295 K. We achieved a mobility enhancement of over 300% in doped films compared with the low-temperature orthorhombic polymorph. Using comprehensive temperature-dependent synchrotron-based X-ray measurements, electronic transport, and first principles calculations, crystal and electronic structures of SSO films were investigated as a function of strain. Furthermore, we argue that strain-engineered films of stannate will enable high mobility oxide electronics operating at RT with the added advantage of being optically transparent.

36 MATERIALS SCIENCE↗

Molecular beam epitaxy growth of SnO{sub 2} using a tin chemical precursor

The authors report on the development of a molecular beam epitaxy approach for atomic layer controlled growth of phase-pure, single-crystalline epitaxial SnO{sub 2} films with scalable growth rates using a highly volatile precursor (tetraethyltin) for tin and rf-oxygen plasma for oxygen. Smooth, epitaxial SnO{sub 2} (101) films on r-sapphire (101{sup ¯}2) substrates were grown as a function of tin precursor flux and substrate temperatures between 300 and 900 °C. Three distinct growth regimes were identified where SnO{sub 2} films grew in a reaction-, flux-, and desorption-limited mode, respectively, with increasing substrate temperature. In particular, with increasing tin flux, the growth rates were found to increase and then saturate indicating any excess tin precursor desorbs above a critical beam equivalent pressure of tin precursor. Important implications of growth kinetic behaviors on the self-regulating stoichiometric growth of perovskite stannates are discussed.

36 MATERIALS SCIENCE↗

Nanocomposite film of TiO{sub 2} nanotube and polyoxometalate towards photocatalytic degradation of nitrobenzene

Highlights: • The film of POMs and TiO{sub 2} nanotubes was prepared by electrodeposition. • The photocatalytic activity of the composite film for nitrobenzene was investigated. • The composite film showed higher photocatalytic activity than pure TiO{sub 2} nanotubes. • The introduction of POMs into TiO{sub 2} could retard electron–hole recombination. - Abstract: The composite film based on polyoxometalates (POMs)-modified TiO{sub 2} nanotubes was prepared by electrodeposition method for the photocatalytic degradation of nitrobenzene. The composite film was characterized by field-emission scanning electron microscopy, X-ray diffraction, energy dispersive X-ray spectroscopy and X-ray photoelectron spectroscopy, which indicated that the POMs were well introduced into the TiO{sub 2} nanotubes. Furthermore, the photocatalytic properties of the TiO{sub 2} nanotubes and POMs-modified TiO{sub 2} nanotubes were evaluated by the decomposition of nitrobenzene. POMs-modified TiO{sub 2} nanotubes showed much higher photocatalytic activity than pure TiO{sub 2} nanotubes. These results provide a promising route to effectively photocatalytic degradation of nitrobenzene by POMs-modified TiO{sub 2} nanotubes.

36 MATERIALS SCIENCE↗

Critical thickness and strain relaxation in molecular beam epitaxy-grown SrTiO{sub 3} films

We report on the study of the critical thickness and the strain relaxation in epitaxial SrTiO{sub 3} film grown on (La{sub 0.3}Sr{sub 0.7})(Al{sub 0.65}Ta{sub 0.35})O{sub 3} (001) (LSAT) substrate using the hybrid molecular beam epitaxy approach. No change in the film's lattice parameter (both the in-plane and the out-of-plane) was observed up to a film thickness of 180 nm, which is in sharp contrast to the theoretical critical thickness of ∼12 nm calculated using the equilibrium theory of strain relaxation. For film thicknesses greater than 180 nm, the out-of-plane lattice parameter was found to decrease hyperbolically in an excellent agreement with the relaxation via forming misfit dislocations. Possible mechanisms are discussed by which the elastic strain energy can be accommodated prior to forming misfit dislocations leading to such anomalously large critical thickness.

36 MATERIALS SCIENCE↗