AITOPAI热报
PapersSelected

PyTorch-to-Spyre compiler pipeline explained by IBM team

官方账号PyTorch@PyTorch
Post
View the post
Why we picked it

The piece reveals the intermediate compiler layers most accelerator benchmarks hide: KTIR and the Dataflow Scheduler sit between Inductor and the hardware, shaping how tiling and memory placement decisions actually execute. Engineers compiling PyTorch for non-GPU accelerators, or evaluating Spyre for inference stacks, get a concrete map of the software path their models take. It also shows how IBM leans on PyTorch's open source compiler stack rather than a proprietary frontend.

IBM's Spyre team published a blog explaining how PyTorch workloads are compiled into instructions executable on IBM's Spyre AI accelerator. The post covers open source compiler technology, including KTIR and the Dataflow Scheduler. It details how PyTorch Inductor's decisions on work distribution, memory placement and tiling are translated into dataflow pipelines that coordinate computation and data movement on Spyre.

Original · PyTorch

How does a PyTorch workload become something an AI accelerator can actually execute?

In our latest blog, the @IBM Spyre team explores the open source compiler technology connecting PyTorch to IBM's Spyre accelerator, including how KTIR and the Dataflow Scheduler translate PyTorch Inductor’s decisions around work distribution, memory placement, and tiling into dataflow pipelines that coordinate computation and data movement on Spyre.

Read more: https://t.co/mHcvSDSwGW

#PyTorch#IBM Spyre#PyTorch Inductor#builders

Keep reading

Recent stories with the same tag or section
ProductsBuildersMulti-sourceSelected04:07
PyTorchCon Demo Theater: 13 Sessions on Training, Inference and Agents

The lineup reveals where the PyTorch ecosystem is concentrating effort: inference speed for video models, MoE expert parallelism on NVL72, and fault-tolerant training that survives link and node failures live. Engineers choosing accelerator backends can use these demos to compare AWS Trainium, Intel AI PCs, TPUs and Rebellions hardware on real workloads. The Huawei CRCR CI relay is also a practical pattern for teams maintaining forks against upstream PyTorch.

#PyTorch#PyTorchCon#Wan 2.2#vLLM
Event
官方账号PyTorch@PyTorch11 sourcesSource
How-toBuildersSelected03:43
PyTorch Examines Linear Algebra Kernels Behind Faster Optimizers

The speed of an optimizer depends as much on kernel-level linear algebra as on step count, a factor most training guides ignore. Engineers tuning large-scale training can apply the QR, Cholesky and eigendecomposition implementation patterns to cut data movement on GPUs. The talk ties directly to GPU MODE's competition, giving concrete reference implementations to study.

#PyTorch#optimizers#GPU#linear algebra
官方账号PyTorch@PyTorchSource
ProductsBuildersSelected02:54
vLLM-TorchTPU brings native PyTorch serving to Google TPUs

The practical effect is that vLLM's optimizations — prefix caching and speculative decoding — now apply to TPU hardware, giving teams a second accelerator option beyond NVIDIA GPUs for self-hosted inference. Organisations already locked into the PyTorch and vLLM stack can evaluate TPU economics without rewriting their serving code. It also signals Google's continued push to open TPU inference beyond JAX.

#vLLM#TPU#PyTorch#Google
官方账号vLLM@vllm_projectSource
ProductsBuildersSelected02:01
FlyDSL kernel DSL now supports PyTorch on AMD MI350 GPUs

AMD MI350 support matters because GPU kernel tooling has historically centred on NVIDIA's stack, and this gives teams serving open models a Python-native path to tune kernels without CUDA-only tools. Benchmark and deployment engineers running vLLM on AMD accelerators can adopt the DSL directly from PyTorch's ecosystem and measure the 3.01x serving gain against their own baselines.

#FlyDSL#PyTorch#AMD#vLLM
官方账号PyTorch@PyTorchSource