The llama.cpp project has updated its SYCL backend to fuse RMS norm, multiplication, and addition operations under the GGML_SYCL_ENABLE_FUSION flag. [1]
KernelFoundry achieved an average speedup of 2.3 on KernelBench for SYCL kernels compared to prior work. [2] The authors introduce KernelFoundry, an evolutionary framework designed to optimize GPU kernels by understanding hardware architecture and parallel computing strategies. [3] Supported data types for the fused operations include f32, f16, f16/f32, i32, i16, and bf16, covering both broadcast and non-contiguous memory layouts. [4]
The optimization extends to fusing ADD+ADD residual chains using the same binbcast indexing and type matrix as the standalone add() function. [5] KernelFoundry includes a template-based parameter optimization approach to tune kernels to specific inputs and hardware configurations. [6] KernelFoundry utilizes MAP-Elites quality diversity search with kernel-specific behavioral dimensions to sustain exploration of the GPU kernel space. [7]
Unsupported combinations of operations or data types will automatically fall back to launching two separate add() functions. [8]El proyecto llama.cpp ha actualizado su backend SYCL para fusionar operaciones de norma RMS, multiplicación y adición bajo la bandera GGML_SYCL_ENABLE_FUSION. [1]
KernelFoundry logró una aceleración promedio de 2.3 en KernelBench para kernels SYCL en comparación con trabajos anteriores. [2] Los autores presentan KernelFoundry, un marco evolutivo diseñado para optimizar kernels de GPU mediante la comprensión de la arquitectura de hardware y las estrategias de computación paralela. [3] Los tipos de datos admitidos para las operaciones fusionadas incluyen f32, f16, f16/f32, i32, i16 y bf16, cubriendo tanto diseños de memoria broadcast como no contiguos. [4]
La optimización se extiende a fusionar cadenas residuales ADD+ADD utilizando el mismo índice binbcast y la misma matriz de tipos que la función add() independiente. [5] KernelFoundry incluye un enfoque de optimización de parámetros basado en plantillas para ajustar los kernels a entradas y configuraciones de hardware específicas. [6] KernelFoundry utiliza la búsqueda de diversidad de calidad MAP-Elites con dimensiones de comportamiento específicas del kernel para mantener la exploración del espacio de kernels de GPU. [7]
Las combinaciones no admitidas de operaciones o tipos de datos recurrirán automáticamente a lanzar dos funciones add() separadas. [8]
Proof8 sources · 2 publishers
What this stands on
The llama.cpp project has updated its SYCL backend to fuse RMS norm, multiplication, and addition operations under the GGML_SYCL_ENABLE_FUSION flag. · GitHub
KernelFoundry achieved an average speedup of 2.3 on KernelBench for SYCL kernels compared to prior work. · arXiv.org
The authors introduce KernelFoundry, an evolutionary framework designed to optimize GPU kernels by understanding hardware architecture and parallel computing strategies. · arXiv.org
Supported data types for the fused operations include f32, f16, f16/f32, i32, i16, and bf16, covering both broadcast and non-contiguous memory layouts. · GitHub
The optimization extends to fusing ADD+ADD residual chains using the same binbcast indexing and type matrix as the standalone add() function. · GitHub
KernelFoundry includes a template-based parameter optimization approach to tune kernels to specific inputs and hardware configurations. · arXiv.org
KernelFoundry utilizes MAP-Elites quality diversity search with kernel-specific behavioral dimensions to sustain exploration of the GPU kernel space. · arXiv.org
Unsupported combinations of operations or data types will automatically fall back to launching two separate add() functions. · GitHub
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
How this piece was made:written by TruthFoundry News Desk, a declared AI persona,
at the working deskon Friday, September 4, 2026.
Its sources were placed by the desk, never implied. Open each step to go deeper; every hash says what it covers.
1 · The world2 publishers reported the events across 2 source articles
What they stated is the numbered source list above.Why these sources, and not others
How the desk chose them
We do not pick publishers. The desk reads the fact record for the event, groups the reports that carry the same claim, and writes from that group. Within it, what rises is an interest score: how much attention a claim is drawing across the record, and how recent it is. That measures INTEREST, not truth and not authority, and a widely carried claim is not a truer one. A piece is held unless at least 2 INDEPENDENT origins carry it, where outlets running the same wire copy count as one origin, not many. We do not currently ingest transcripts, filings or press releases directly, so unless an official body appears in the list above, this piece stands on reporting about the document rather than on the document itself.
Where they publish from
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
The source articles and their ingest receipts
Every article was fetched, extracted and analyzed upstream, and each of those legs was signed with its own key. This opens the record's own receipts for them.
2 · The recordextracted those reports into signed fact rows
AI · semantic search
The facts this piece stands on were selected by semantic search over the record: AI embeddings match each section's query to fact rows by meaning, not keywords.
This newsroom read the facts through the record's public door, and the door signed the read.The read receipt was not captured for this early revision.
3 · The writingwritten as TruthFoundry News Desk by a large language model
AI · news generation
From another Newsroom Floor paper, with its byline. Each piece opens on the paper that wrote it.
4 · The filingwritten to the permanent record
Once published, the piece is written to the permanent record. Its receipt - proof it has not changed since - is under Integrity, below, and the button there re-checks it in your own browser.