Daily Papers

W4A4 quantization promises full utilization of INT4 Tensor Cores, yet group dequantization overhead on CUDA Cores has driven existing systems to mixed-precision fallbacks. We present the first systematic study of how intra-SM compute balance governs this bottleneck. Through controlled benchmarks across four GPUs from Ampere and Ada architectures, we identify the Tensor Cores to CUDA Cores throughput ratio (ρ) as the primary hardware indicator: the W4A4-g128 kernel yields 2.0--2.5times speedup on RTX~3090 (ρ=16) yet degrades to 0.43--0.47times on A100 (ρ=64) in compute-bond scenarios, establishing W4A4 viability as platform-dependent rather than universally infeasible. Guided by this finding, we build APEX4, which co-designs pure INT4 GEMM kernels with ρ-aware granularity adaptation to mitigate the CUDA Cores dequantization bottleneck. APEX4 achieves perplexity within 0.63 of FP16 on LLaMA-2-70B and outperforms W4Ax Atom-g128 by 4.0\%--4.4\% in zero-shot accuracy. Deployed as a drop-in replacement in unmodified vLLM, it delivers up to 1.66times end-to-end speedup on L40S (ρ=8), and 1.78times on RTX~3090 (ρ=16), 2.09times on A40 (ρ=16), while recovering A100 (ρ=64) to 1.20--1.40times via the mixed-granularity mode. Our code is available at https://github.com/APEX4-W4A4/APEX4-W4A4.

Modern transformer-based Large Language Models (LLMs) are constructed with a series of decoder blocks. Each block comprises three key components: (1) QKV generation, (2) multi-head attention, and (3) feed-forward networks. In batched processing, QKV generation and feed-forward networks involve compute-intensive matrix-matrix multiplications (GEMM), while multi-head attention requires bandwidth-heavy matrix-vector multiplications (GEMV). Machine learning accelerators like TPUs or NPUs are proficient in handling GEMM but are less efficient for GEMV computations. Conversely, Processing-in-Memory (PIM) technology is tailored for efficient GEMV computation, while it lacks the computational power to handle GEMM effectively. Inspired by this insight, we propose NeuPIMs, a heterogeneous acceleration system that jointly exploits a conventional GEMM-focused NPU and GEMV-optimized PIM devices. The main challenge in efficiently integrating NPU and PIM lies in enabling concurrent operations on both platforms, each addressing a specific kernel type. First, existing PIMs typically operate in a "blocked" mode, allowing only either NPU or PIM to be active at any given time. Second, the inherent dependencies between GEMM and GEMV in LLMs restrict their parallel processing. To tackle these challenges, NeuPIMs is equipped with dual row buffers in each bank, facilitating the simultaneous management of memory read/write operations and PIM commands. Further, NeuPIMs employs a runtime sub-batch interleaving technique to maximize concurrent execution, leveraging batch parallelism to allow two independent sub-batches to be pipelined within a single NeuPIMs device. Our evaluation demonstrates that compared to GPU-only, NPU-only, and a na\"ive NPU+PIM integrated acceleration approaches, NeuPIMs achieves 3times, 2.4times and 1.6times throughput improvement, respectively.

General Matrix Multiplication (GEMM) is a critical kernel in high-performance computing and deep learning. While modern architectures like ARM's Scalable Matrix Extension (SME) introduce dedicated hardware for matrix operations, existing linear algebra libraries fail to fully exploit its potential, particularly for large matrices. This paper presents MpGEMM, an open-source library that leverages key architectural features of SME to optimize GEMM across multiple precisions. Through a systematic characterization of SME, we derive optimization guidelines that inform our design. MpGEMM employs cache-aware partitioning, efficient data packing with on-the-fly transposition, and specialized micro-kernels that utilize multi-vector loads and all available tile registers. Evaluated on an Apple M4 Pro with real-world workloads from DeepSeek and LLaMA, MpGEMM achieves an average speedup of 1.23x over the vendor-optimized Apple Accelerate library and significantly outperforms other open-source alternatives.

Deep learning frameworks commonly implement convolution operators with GEMM-based algorithms. In these algorithms, convolution is implemented on top of matrix-matrix multiplication (GEMM) functions, provided by highly optimized BLAS libraries. Convolutions with 1x1 kernels can be directly represented as a GEMM call, but convolutions with larger kernels require a special memory layout transformation - im2col or im2row - to fit into GEMM interface. The Indirect Convolution algorithm provides the efficiency of the GEMM primitive without the overhead of im2col transformation. In contrast to GEMM-based algorithms, the Indirect Convolution does not reshuffle the data to fit into the GEMM primitive but introduces an indirection buffer - a buffer of pointers to the start of each row of image pixels. This broadens the application of our modified GEMM function to convolutions with arbitrary kernel size, padding, stride, and dilation. The Indirect Convolution algorithm reduces memory overhead proportionally to the number of input channels and outperforms the GEMM-based algorithm by up to 62% on convolution parameters which involve im2col transformations in GEMM-based algorithms. This, however, comes at cost of minor performance reduction on 1x1 stride-1 convolutions.

Large foundation models have become central to modern machine learning, with performance scaling predictably with model size and data. However, training and deploying such models incur substantial computational and memory costs, motivating the development of low-precision training techniques. Recent work has demonstrated that 4-bit floating-point (FP4) formats--such as MXFP4 and NVFP4--can be successfully applied to linear GEMM operations in large language models (LLMs), achieving up to 4x improvements in compute throughput and memory efficiency compared to higher-precision baselines. In this work, we investigate the recently proposed HiFloat4 FP4 format for Huawei Ascend NPUs and systematically compare it with MXFP4 in large-scale training settings. All experiments are conducted on Ascend NPU clusters, with linear and expert GEMM operations performed entirely in FP4 precision. We evaluate both dense architectures (e.g., Pangu and LLaMA-style models) and mixture-of-experts (MoE) models, where both standard linear layers and expert-specific GEMMs operate in FP4. Furthermore, we explore stabilization techniques tailored to FP4 training that significantly reduce numerical degradation, maintaining relative error within 1% of full-precision baselines while preserving the efficiency benefits of 4-bit computation. Our results provide a comprehensive empirical study of FP4 training on NPUs and highlight the practical trade-offs between FP4 formats in large-scale dense and MoE models.

Scaling laws for Large Language Models (LLMs) establish that model quality improves with computational scale, yet edge deployment imposes strict constraints on compute, memory, and power. Since General Matrix Multiplication (GEMM) accounts for up to 90% of inference time, efficient GEMM acceleration is critical for edge AI. The Adaptive Intelligent Engines available in the AMD Versal adaptive SoCs are well suited for this task, but existing state-of-the-art (SOTA) frameworks maximize performance through spatial scaling, distributing workloads across hundreds of cores -- an approach that fails on resource-limited edge SoCs due to physical implementation failures, bandwidth saturation, and excessive resource consumption. We propose Tempus, a Resource-Invariant Temporal GEMM framework for the AMD Versal AI Edge SoC. Rather than expanding hardware resources with matrix size, Tempus employs a fixed compute block of 16 AIE-ML cores, achieving scalability through iterative graph execution and algorithmic data tiling and replication in the Programmable Logic. High-speed cascade streaming ensures low-latency partial sum reduction at Initiation Interval (II) of 1, while a deadlock-free DATAFLOW protocol maximizes transfer-compute overlap and PLIO reuse. Evaluated on GEMM workloads, Tempus achieves 607 GOPS at 10.677 W total on-chip power. By characterizing system-level efficiency through the Platform-Aware Utility (PAU) metric, we prove that Tempus achieves a 211.2x higher prominence factor than the leading spatial SOTA (ARIES). Furthermore, the framework maintains a 0.00% utilization of URAM/DSP, yielding 22.0x core frugality, 7.1x power frugality, and a 6.3x reduction in I/O demand, establishing a sustainable, scalable foundation for edge LLM inference.

General matrix-matrix multiplication (GEMM) is a cornerstone of AI computations, making tensor processing engines (TPEs) increasingly critical in GPUs and domain-specific architectures. Existing architectures primarily optimize dataflow or operand reuse strategies. However, considering the interaction between matrix multiplication and multiply-accumulators (MACs) offers greater optimization potential. This work introduces a novel hardware perspective on matrix multiplication, focusing on the bit-weight dimension of MACs. We propose a finer-grained TPE notation using matrix triple loops as an example, introducing new methods for designing and optimizing PE microarchitectures. Based on this notation and its transformations, we propose four optimization techniques that improve timing, area, and power consumption. Implementing our design in RTL using the SMIC-28nm process, we evaluate its effectiveness across four classic TPE architectures: systolic array, 3D-Cube, multiplier-adder tree, and 2D-Matrix. Our techniques achieve area efficiency improvements of 1.27x, 1.28x, 1.56x, and 1.44x, and energy efficiency gains of 1.04x, 1.56x, 1.49x, and 1.20x, respectively. Applied to a bit-slice architecture, our approach achieves a 12.10x improvement in energy efficiency and 2.85x in area efficiency compared to Laconic. Our Verilog HDL code, along with timing, area, and power reports, is available at https://github.com/wqzustc/High-Performance-Tensor-Processing-Engines

Low precision (LP) datatypes such as MXFP4 can accelerate matrix multiplications (GEMMs) and reduce training costs. However, directly using MXFP4 instead of BF16 during training significantly degrades model quality. In this work, we present the first near-lossless training recipe that uses MXFP4 GEMMs, which are 2times faster than FP8 on supported hardware. Our key insight is to compute unbiased gradient estimates with stochastic rounding (SR), resulting in more accurate model updates. However, directly applying SR to MXFP4 can result in high variance from block-level outliers, harming convergence. To overcome this, we use the random Hadamard tranform to theoretically bound the variance of SR. We train GPT models up to 6.7B parameters and find that our method induces minimal degradation over mixed-precision BF16 training. Our recipe computes >1/2 the training FLOPs in MXFP4, enabling an estimated speedup of >1.3times over FP8 and >1.7times over BF16 during backpropagation.

As the Large Language Model (LLM) becomes increasingly important in various domains. However, the following challenges still remain unsolved in accelerating LLM inference: (1) Synchronized partial softmax update. The softmax operation requires a synchronized update operation among each partial softmax result, leading to ~20% overheads for the attention computation in LLMs. (2) Under-utilized computation of flat GEMM. The shape of matrices performing GEMM in LLM inference is flat, leading to under-utilized computation and >50% performance loss after padding zeros in previous designs. (3) Performance loss due to static dataflow. Kernel performance in LLM depends on varied input data features, hardware configurations, etc. A single and static dataflow may lead to a 50.25% performance loss for GEMMs of different shapes in LLM inference. We present FlashDecoding++, a fast LLM inference engine supporting mainstream LLMs and hardware back-ends. To tackle the above challenges, FlashDecoding++ creatively proposes: (1) Asynchronized softmax with unified max value. FlashDecoding++ introduces a unified max value technique for different partial softmax computations to avoid synchronization. (2) Flat GEMM optimization with double buffering. FlashDecoding++ points out that flat GEMMs with different shapes face varied bottlenecks. Then, techniques like double buffering are introduced. (3) Heuristic dataflow with hardware resource adaptation. FlashDecoding++ heuristically optimizes dataflow using different hardware resource considering input dynamics. Due to the versatility of optimizations in FlashDecoding++, FlashDecoding++ can achieve up to 4.86x and 2.18x speedup on both NVIDIA and AMD GPUs compared to Hugging Face implementations. FlashDecoding++ also achieves an average speedup of 1.37x compared to state-of-the-art LLM inference engines on mainstream LLMs.

Mixed-precision quantization is a promising approach for compressing large language models under tight memory budgets. However, existing mixed-precision methods typically suffer from one of two limitations: they either rely on expensive discrete optimization to determine precision allocation, or introduce hardware inefficiencies due to irregular memory layouts. We propose SFMP, a search-free and hardware-friendly mixed-precision quantization framework for large language models. The framework is built upon four novel ideas: Fractional bit-width, which extends integer bit-width for weight matrix to fractional value and transforms discrete precision allocation as a continuous problem; 2)Block-wise mixed-precision, enabling fine-grained precision within weight matrices while remaining hardware-friendly; 3)Row-column weight reordering, which aggregates salient weights via row and column reordering, incurring only a small activation reordering overhead during inference; 4)Unified GEMM kernel, which supports mixed-precision GEMM at arbitrary average bit-width. Extensive experiments demonstrate that SFMP outperforms state-of-the-art layer-wise mixed-precision methods under the same memory constraints, while significantly reducing quantization cost and improving inference efficiency. Code is available at https://github.com/Nkniexin/SFMP

Low-bit floating-point formats and semi-structured sparsity are increasingly supported by modern accelerators, yet combining them for LLM activation compression remains challenging: activations contain input-dependent outliers that dominate block scales in FP4 quantization, and directly applying N:M sparsity masks discards moderate values, coupling sparsification loss with quantization error. We introduce SharQ, a training-free inference method that bridges activation sparsity and FP4 quantization through an online sparse--dense decomposition. For each activation tensor, SharQ generates an input-adaptive N:M mask to extract an outlier-dominated sparse backbone, quantizes it to FP4, and defines a dense residual relative to the quantized sparse backbone rather than the unquantized sparse values. A sparse FP4 GEMM processes the backbone while a dense FP4 GEMM compensates for both mask-induced activation loss and sparse-path quantization error. The two paths share a single FP4 weight payload with path-specific scale views, and a fused preparation kernel absorbs mask generation, residual construction, and layer normalization into one operator. SharQ requires no calibration data, retraining, or model-specific tuning. Evaluated on Llama-3.1-8B, Qwen2.5-7B, Qwen3-30B-A3B, and Qwen3-VL-8B, SharQ recovers 43--63% of the NVFP4-to-FP16 accuracy gap across language and vision-language tasks, and generalizes across NVFP4, HiF4, and MXFP4 formats. On an RTX 5090, SharQ delivers 2.2--2.4times latency reduction over FP16 and 1.2--1.4times throughput improvement over FP8 in language model serving, and up to 1.58times speedup on Wan2.2-T2V-A14B video generation when combined with SageAttention. Our code is available at https://github.com/actypedef/SharQ.

As Large Language Models (LLMs) scale, weight-only quantization (W4A16: 4-bit weights, 16-bit activations) becomes critical for reducing memory footprint with minimal accuracy loss. However, its efficient deployment on Huawei's Ascend 910 Neural Processing Unit (NPU) is challenging due to limited native mixed-precision support and the accelerator's decoupled compute architecture. To enable quantization on such architecture, we present the first practical W4A16 matrix multiplication kernel tailored for the Ascend 910 NPU. Our design leverages vector cores for on-the-fly INT4-to-FP16 dequantization, cube cores for high-throughput GEMM, and Split-K parallelization to mitigate memory latency. Performance evaluations across diverse matrix shapes and batch sizes show our method outperforms data-parallel approaches when K >> N, a typical scenario in LLM decoding. Specially, our method can achieve a speedup ranging from 1.01x to 1.74x. In addition, our profile reveals the primary bottleneck is not dequantization compution itself, but extra global memory transfer for the weight, making W4A16 only reaching a maximum speedup of 1.48x over native FP16xFP16 matrix multiplication in PyTorch. In the long run, our method lays a solid foundation and provides insightful views for the efficient deployment of quantized large language models on various domain-specific accelerators.

Large matrix multiplication is a cornerstone of modern machine learning workloads, yet traditional approaches suffer from cubic computational complexity (e.g., O(n^3) for a matrix of size ntimes n). We present Low-Rank GEMM, a novel approach that leverages low-rank matrix approximations to achieve sub-quadratic complexity while maintaining hardware-accelerated performance through FP8 precision and intelligent kernel selection. On a NVIDIA RTX 4090, our implementation achieves up to 378 TFLOPS on matrices up to N=20480, providing 75\% memory savings and 7.8times speedup over PyTorch FP32 for large matrices. The system automatically adapts to hardware capabilities, selecting optimal decomposition methods (SVD, randomized SVD) and precision levels based on matrix characteristics and available accelerators. Comprehensive benchmarking on NVIDIA RTX 4090 demonstrates that Low-Rank GEMM becomes the fastest approach for matrices Ngeq10240, surpassing traditional cuBLAS implementations through memory bandwidth optimization rather than computational shortcuts.

The emergence of fine-grained numerical formats like NVFP4 presents new opportunities for efficient Large Language Model (LLM) inference. However, it is difficult to adapt existing Post-Training Quantization (PTQ) strategies to these formats: rotation-based methods compromise fine-grained block isolation; smoothing techniques struggle with significant 4-bit quantization errors; and mixed-precision approaches often conflict with hardware constraints on unified-precision computation. To address these challenges, we propose ARCQuant, a framework that boosts NVFP4 performance via Augmented Residual Channels. Distinct from methods that compromise block isolation or hardware uniformity, ARCQuant maintains a strictly unified NVFP4 format by augmenting the activation matrix with quantized residual channels. This design integrates the error compensation process directly into the matrix reduction dimension, enabling the use of standard, highly optimized GEMM kernels with minimal overhead. Theoretical analysis confirms that the worst-case error bound of our dual-stage NVFP4 quantization is comparable to that of standard 8-bit formats such as MXFP8. Extensive experiments on LLaMA and Qwen models demonstrate that ARCQuant achieves state-of-the-art accuracy, comparable to full-precision baselines in perplexity and downstream tasks. Furthermore, deployment on RTX 5090 and RTX PRO 6000 GPUs confirms practical benefits, achieving up to 3x speedup over FP16. Our code is available at https://github.com/actypedef/ARCQuant .

Convolution is the most time-consuming operation in deep neural network operations, so its performance is critical to the overall performance of the neural network. The commonly used methods for convolution on GPU include the general matrix multiplication (GEMM)-based convolution and the direct convolution. GEMM-based convolution relies on the im2col algorithm, which results in a large memory footprint and reduced performance. Direct convolution does not have the large memory footprint problem, but the performance is not on par with GEMM-based approach because of the discontinuous memory access. This paper proposes a window-order-based convolution paradigm on GPU, called im2win, which not only reduces memory footprint but also offers continuous memory accesses, resulting in improved performance. Furthermore, we apply a range of optimization techniques on the convolution CUDA kernel, including shared memory, tiling, micro-kernel, double buffer, and prefetching. We compare our implementation with the direct convolution, and PyTorch's GEMM-based convolution with cuBLAS and six cuDNN-based convolution implementations, with twelve state-of-the-art DNN benchmarks. The experimental results show that our implementation 1) uses less memory footprint by 23.1% and achieves 3.5times TFLOPS compared with cuBLAS, 2) uses less memory footprint by 32.8% and achieves up to 1.8times TFLOPS compared with the best performant convolutions in cuDNN, and 3) achieves up to 155times TFLOPS compared with the direct convolution. We further perform an ablation study on the applied optimization techniques and find that the micro-kernel has the greatest positive impact on performance.

Quantization is a proven effective method for compressing large language models. Although popular techniques like W8A8 and W4A16 effectively maintain model performance, they often fail to concurrently speed up the prefill and decoding stages of inference. W4A8 is a promising strategy to accelerate both of them while usually leads to a significant performance degradation. To address these issues, we present QQQ, a Quality Quattuor-bit Quantization method with 4-bit weights and 8-bit activations. QQQ employs adaptive smoothing and Hessian-based compensation, significantly enhancing the performance of quantized models without extensive training. Furthermore, we meticulously engineer W4A8 GEMM kernels to increase inference speed. Our specialized per-channel W4A8 GEMM and per-group W4A8 GEMM achieve impressive speed increases of 3.67times and 3.29 times over FP16 GEMM. Our extensive experiments show that QQQ achieves performance on par with existing state-of-the-art LLM quantization methods while significantly accelerating inference, achieving speed boosts up to 2.24 times, 2.10times, and 1.25times compared to FP16, W8A8, and W4A16, respectively.

Emerging AI accelerators increasingly adopt wafer-scale manufacturing technologies, integrating hundreds of thousands of AI cores in a mesh architecture with large distributed on-chip memory (tens of GB in total) and ultra-high on-chip memory bandwidth (tens of PB/s). However, current LLM inference systems, optimized for shared memory architectures like GPUs, fail to exploit these accelerators fully. We introduce WaferLLM, the first wafer-scale LLM inference system. WaferLLM is guided by a novel PLMR model (pronounced as "Plummer") that captures the unique hardware characteristics of wafer-scale architectures. Leveraging this model, WaferLLM pioneers wafer-scale LLM parallelism, optimizing the utilization of hundreds of thousands of on-chip cores. It also introduces MeshGEMM and MeshGEMV, the first GEMM and GEMV implementations designed to scale effectively on wafer-scale accelerators. Evaluations show that WaferLLM achieves up to 200times higher accelerator utilization than state-of-the-art methods. Leveraging a wafer-scale accelerator (Cerebras WSE2), WaferLLM delivers GEMV operations 606times faster and 16times more energy-efficient than on an NVIDIA A100 GPU. For full LLM inference, WaferLLM achieves 10-20times speedups over A100 GPU clusters running SGLang and vLLM. These advantages are expected to grow as wafer-scale AI models, software, and hardware continue to mature. WaferLLM is open-sourced at https://github.com/MeshInfra/WaferLLM.

Quantizing model weights is critical for reducing the communication and inference costs of large models. However, quantizing models -- especially to low precisions like int4 or int2 -- requires a trade-off in model quality; int2, in particular, is known to severely degrade model quality. Consequently, practitioners are often forced to maintain multiple models with different quantization levels or serve a single model that best satisfies the quality-latency trade-off. On the other hand, integer data types, such as int8, inherently possess a nested (Matryoshka) structure where smaller bit-width integers, like int4 or int2, are nested within the most significant bits. This paper proposes Matryoshka Quantization (MatQuant), a novel multi-scale quantization technique that addresses the challenge of needing multiple quantized models. It allows training and maintaining just one model, which can then be served at different precision levels. Furthermore, due to the co-training and co-distillation regularization provided by MatQuant, the int2 precision models extracted by MatQuant can be up to 10% more accurate than standard int2 quantization (using techniques like QAT or OmniQuant). This represents significant progress in model quantization, demonstrated by the fact that, with the same recipe, an int2 FFN-quantized Gemma-2 9B model is more accurate than an int8 FFN-quantized Gemma-2 2B model.

Transformer training systems are built around dense linear algebra, yet a nontrivial fraction of end-to-end time is spent on surrounding memory-bound operators. Normalization, activations, residual updates, reductions, and related computations repeatedly move large intermediate tensors through global memory while performing little arithmetic, making data movement an increasingly important bottleneck in otherwise highly optimized training stacks. We introduce CODA, a GPU kernel abstraction that expresses these computations as GEMM-plus-epilogue programs. CODA is based on the observation that many Transformer operators exposed as separate framework kernels can be algebraically reparameterized to execute while a GEMM output tile remains on chip, before it is written to memory. The abstraction fixes the GEMM mainloop and exposes a small set of composable epilogue primitives for scaling, reductions, pairwise transformations, and accumulation. This constrained interface preserves the performance structure of expert-written GEMMs while remaining expressive enough to cover nearly all non-attention computation in the forward and backward pass of a standard Transformer block. Across representative Transformer workloads, both human- and LLM-authored CODA kernels achieve high performance, suggesting that GEMM-plus-epilogue programming offers a practical path toward combining framework-level productivity with hardware-level efficiency.

The application of on-device language models (ODLMs) on resource-constrained edge devices is a multi-dimensional problem that strikes a fine balance between computational effectiveness, memory, power usage, and linguistic capacity across heterogeneous tasks. This holistic study conducts a thorough investigation of the trade-offs between domain-specific optimization and cross-domain robustness, culminating in the proposal of the Generalized Edge Model (GEM), a new architecture that aims to balance specialization and generalization in a harmonious manner. With a rigorous experimental approach testing 47 well-chosen benchmarks in eight domains--healthcare, law, finance, STEM, commonsense, conversational AI, multilingual, and domain-adaptive tasks--we show that conventional optimization techniques decrease target task perplexity by 18-25% but result in a precipitous decline in general-task performance with F1 scores decreasing by 12-29%, as reported by Liu et al. GEM employs a Sparse Cross-Attention Router (SCAR) to dynamically allocate computation to a variable number of computing resources with a cross-domain F1 accuracy of 0.89 on less than 100ms latency across Raspberry Pi 4, Pixel 6, iPhone 13, and bespoke custom neural processing units (NPUs). Compared to GPT-4 Lite, GEM enhances the general-task level by 7% with respect and parity in domain-specific performance. We propose three new measurement tools--Domain Specialization Index (DSI), Generalization Gap (GG), and Cross-Domain Transfer Ratio (CDTR)--which show strong correlation between model compression intensity and brittleness.