Madan Musuvathi

Sarah Fakhoury, Saikat Chakraborty, Madan Musuvathi et al.

Large language models (LLMs), such as OpenAI's Codex, have demonstrated their potential to generate code from natural language descriptions across a wide range of programming tasks. Several benchmarks have recently emerged to evaluate the ability of LLMs to generate functionally correct code from natural language intent with respect to a set of hidden test cases. This has enabled the research community to identify significant and reproducible advancements in LLM capabilities. However, there is currently a lack of benchmark datasets for assessing the ability of LLMs to generate functionally correct code edits based on natural language descriptions of intended changes. This paper aims to address this gap by motivating the problem NL2Fix of translating natural language descriptions of code changes (namely bug fixes described in Issue reports in repositories) into correct code fixes. To this end, we introduce Defects4J-NL2Fix, a dataset of 283 Java programs from the popular Defects4J dataset augmented with high-level descriptions of bug fixes, and empirically evaluate the performance of several state-of-the-art LLMs for the this task. Results show that these LLMS together are capable of generating plausible fixes for 64.6% of the bugs, and the best LLM-based technique can achieve up to 21.20% top-1 and 35.68% top-5 accuracy on this benchmark.

Aashaka Shah, Abhinav Jangda, Binyang Li et al.

Modern cutting-edge AI applications are being developed over fast-evolving, heterogeneous, nascent hardware devices. This requires frequent reworking of the AI software stack to adopt bottom-up changes from new hardware, which takes time for general-purpose software libraries. Consequently, real applications often develop custom software stacks optimized for their specific workloads and hardware. Custom stacks help in quick development and optimization, but incur a lot of redundant efforts across applications in writing non-portable code. This paper discusses an alternative communication library interface for AI applications that offers both portability and performance by reducing redundant efforts while maintaining flexibility for customization. We present MSCCL++, a novel abstraction of GPU communication based on separation of concerns: (1) a primitive interface provides a minimal hardware abstraction as a common ground for software and hardware developers to write custom communication, and (2) higher-level portable interfaces and specialized implementations enable optimization for different workloads and hardware environments. This approach makes the primitive interface reusable across applications while enabling highly flexible optimization. Compared to state-of-the-art baselines (NCCL, RCCL, and MSCCL), MSCCL++ achieves speedups of up to 5.4$\times$ for collective communication and up to 15% for real-world AI inference workloads. MSCCL++ is in production of multiple AI services provided by Microsoft Azure, and is also adopted by RCCL, the GPU collective communication library maintained by AMD. MSCCL++ is open-source and available at https://github.com/microsoft/mscclpp.

Saeed Maleki, Madan Musuvathi, Todd Mytkowicz et al.

Stochastic gradient descent (SGD) is an inherently sequential training algorithm--computing the gradient at batch $i$ depends on the model parameters learned from batch $i-1$. Prior approaches that break this dependence do not honor them (e.g., sum the gradients for each batch, which is not what sequential SGD would do) and thus potentially suffer from poor convergence. This paper introduces a novel method to combine gradients called Adasum (for adaptive sum) that converges faster than prior work. Adasum is easy to implement, almost as efficient as simply summing gradients, and is integrated into the open-source toolkit Horovod. This paper first provides a formal justification for Adasum and then empirically demonstrates Adasum is more accurate than prior gradient accumulation methods. It then introduces a series of case-studies to show Adasum works with multiple frameworks, (TensorFlow and PyTorch), scales multiple optimizers (Momentum-SGD, Adam, and LAMB) to larger batch-sizes while still giving good downstream accuracy. Finally, it proves that Adasum converges. To summarize, Adasum scales Momentum-SGD on the MLPerf Resnet50 benchmark to 64K examples before communication (no MLPerf v0.5 entry converged with more than 16K), the Adam optimizer to 64K examples before communication on BERT-LARGE (prior work showed Adam stopped scaling at 16K), and the LAMB optimizer to 128K before communication on BERT-LARGE (prior work used 64K), all while maintaining downstream accuracy metrics. Finally, if a user does not need to scale, we show LAMB with Adasum on BERT-LARGE converges in 30% fewer steps than the baseline.

Yuhan Liu, Yuyang Huang, Jiayi Yao et al.

Compound AI systems, such as agentic systems, are an emerging trend in large-scale enterprise settings, with multiple LLMs specialized for different users, tasks, and/or roles working together. In these scenarios, different models often process inputs that share the same context prefix. Although much work was done in the past to enable the reuse of prefix KV caches across inputs for a single model, how to enable one model to reuse the prefix KV caches of a different model remains an open question. We introduce DroidSpeak, the first distributed LLM inference system that enables KV cache reuse across distributed nodes running inference of different LLMs, so long as the LLMs have the same architecture. We present the first study that aims at understanding the impact of sharing KV caches across different LLMs, and if/when such sharing affects quality. Inspired by the findings, we present DroidSpeak, which selectively recomputes a few layers of the KV cache produced by another LLM and reuses the remaining layers, with negligible quality loss. Moreover, carefully pipelining the layer-wise re-computation and the loading of reused KV cache further improves the inference performance. Experiments on diverse datasets and model pairs demonstrate that DroidSpeak achieves up to 4x throughput improvement and about 3.1x faster prefill (time to first token), with negligible loss of quality in F1 scores, Rouge-L or code similarity score, compared to the baseline which does not allow any sharing across models.

Jubi Taneja, Avery Laird, Cong Yan et al.

Vectorization is a powerful optimization technique that significantly boosts the performance of high performance computing applications operating on large data arrays. Despite decades of research on auto-vectorization, compilers frequently miss opportunities to vectorize code. On the other hand, writing vectorized code manually using compiler intrinsics is still a complex, error-prone task that demands deep knowledge of specific architecture and compilers. In this paper, we evaluate the potential of large-language models (LLMs) to generate vectorized (Single Instruction Multiple Data) code from scalar programs that process individual array elements. We propose a novel finite-state machine multi-agents based approach that harnesses LLMs and test-based feedback to generate vectorized code. Our findings indicate that LLMs are capable of producing high performance vectorized code with run-time speedup ranging from 1.1x to 9.4x as compared to the state-of-the-art compilers such as Intel Compiler, GCC, and Clang. To verify the correctness of vectorized code, we use Alive2, a leading bounded translation validation tool for LLVM IR. We describe a few domain-specific techniques to improve the scalability of Alive2 on our benchmark dataset. Overall, our approach is able to verify 38.2% of vectorizations as correct on the TSVC benchmark dataset.

Aashaka Shah, Vijay Chidambaram, Meghan Cowan et al.

Machine learning models are increasingly being trained across multiple GPUs and servers. In this setting, data is transferred between GPUs using communication collectives such as AlltoAll and AllReduce, which can become a significant bottleneck in training large models. Thus, it is important to use efficient algorithms for collective communication. We develop TACCL, a tool that enables algorithm designers to guide a synthesizer into automatically generating algorithms for a given hardware configuration and communication collective. TACCL uses a novel communication sketch abstraction to get crucial information from the designer to significantly reduce the search space and guide the synthesizer towards better algorithms. TACCL also uses a novel encoding of the problem that allows it to scale beyond single-node topologies. We use TACCL to synthesize algorithms for three collectives and two hardware topologies: DGX-2 and NDv2. We demonstrate that the algorithms synthesized by TACCL outperform the Nvidia Collective Communication Library (NCCL) by up to 6.7x. We also show that TACCL can speed up end-to-end training of Transformer-XL and BERT models by 11%--2.3x for different batch sizes.

Sharadha Srinivasan, Madan Musuvathi

In recent times, there have been increasing accusations on artificial intelligence systems and algorithms of computer vision of possessing implicit biases. Even though these conversations are more prevalent now and systems are improving by performing extensive testing and broadening their horizon, biases still do exist. One such class of systems where bias is said to exist is facial recognition systems, where bias has been observed on the basis of gender, ethnicity, skin tone and other facial attributes. This is even more disturbing, given the fact that these systems are used in practically every sector of the industries today. From as critical as criminal identification to as simple as getting your attendance registered, these systems have gained a huge market, especially in recent years. That in itself is a good enough reason for developers of these systems to ensure that the bias is kept to a bare minimum or ideally non-existent, to avoid major issues like favoring a particular gender, race, or class of people or rather making a class of people susceptible to false accusations due to inability of these systems to correctly recognize those people.

Gurbinder Gill, Roshan Dathathri, Saeed Maleki et al.

Many applications today, such as NLP, network analysis, and code analysis, rely on semantically embedding objects into low-dimensional fixed-length vectors. Such embeddings naturally provide a way to perform useful downstream tasks, such as identifying relations among objects or predicting objects for a given context, etc. Unfortunately, the training necessary for accurate embeddings is usually computationally intensive and requires processing large amounts of data. Furthermore, distributing this training is challenging. Most embedding training uses stochastic gradient descent (SGD), an "inherently" sequential algorithm. Prior approaches to parallelizing SGD do not honor these dependencies and thus potentially suffer poor convergence. This paper presents a distributed training framework for a class of applications that use Skip-gram-like models to generate embeddings. We call this class Any2Vec and it includes Word2Vec, DeepWalk, and Node2Vec among others. We first formulate Any2Vec training algorithm as a graph application and leverage the state-of-the-art distributed graph analytics framework, D-Galois. We adapt D-Galois to support dynamic graph generation and repartitioning, and incorporate novel communication optimizations. Finally, we introduce a novel way to combine gradients during distributed training to prevent accuracy loss. We show that our framework, called GraphAny2Vec, matches on a cluster of 32 hosts the accuracy of the state-of-the-art shared-memory implementations of Word2Vec and Vertex2Vec on 1 host, and gives a geo-mean speedup of 12x and 5x respectively. Furthermore, GraphAny2Vec is on average 2x faster than the state-of-the-art distributed Word2Vec implementation, DMTK, on 32 hosts. We also show the superiority of our Gradient Combiner independent of GraphAny2Vec by incorporating it in DMTK, which raises its accuracy by > 30%.