#sycl
3 posts3 participants1 post today
My brain is absolutely fried.
Today is the last day of coursework submissions for this semester. What a hectic month.
DNN with PyTorch, Brain model parallelisation with MPI, SYCL and OpenMP offloading of percolation models,hand optimizing serial codes for performance.
Two submissions due today. Submitted one and finalising my report for the second one.
Definitely having a pint after this
Started SYCL this semester in my MSc, and I have a coursework on it.
I have never been more frustrated in my life.
I am not saying SYCL is bad. I might just be too dumb to master it in a sem in order to port an existing CPU code to use MPI & SYCL together.
CUDA was much easier for me for the same task.
hgpu.org · The Shamrock code: I- Smoothed Particle Hydrodynamics on GPUsWe present Shamrock, a performance portable framework developed in C++17 with the SYCL programming standard, tailored for numerical astrophysics on Exascale architectures. The core of Shamrock is a…
hgpu.org · ML-Triton, A Multi-Level Compilation and Language Extension to Triton GPU ProgrammingIn the era of LLMs, dense operations such as GEMM and MHA are critical components. These operations are well-suited for parallel execution using a tilebased approach. While traditional GPU programm…
Concurrent Scheduling of High-Level Parallel Programs on Multi-GPU Systems
hgpu.org · Concurrent Scheduling of High-Level Parallel Programs on Multi-GPU SystemsParallel programming models can encourage performance portability by moving the responsibility for work assignment and data distribution from the programmer to a runtime system. However, analyzing …
Continued thread
Even now, Thrust as a dependency is one of the main reason why we have a #CUDA backend, a #HIP / #ROCm backend and a pure #CPU backend in #GPUSPH, but not a #SYCL or #OneAPI backend (which would allow us to extend hardware support to #Intel GPUs). <https://doi.org/10.1002/cpe.8313>
This is also one of the reason why we implemented our own #BLAS routines when we introduced the semi-implicit integrator. A side-effect of this choice is that it allowed us to develop the improved #BiCGSTAB that I've had the opportunity to mention before <https://doi.org/10.1016/j.jcp.2022.111413>. Sometimes I do wonder if it would be appropriate to “excorporate” it into its own library for general use, since it's something that would benefit others. OTOH, this one was developed specifically for GPUSPH and it's tightly integrated with the rest of it (including its support for multi-GPU), and refactoring to turn it into a library like cuBLAS is
a. too much effort
b. probably not worth it.
Again, following @eniko's original thread, it's really not that hard to roll your own, and probably less time consuming than trying to wrangle your way through an API that may or may not fit your needs.
6/
I'm getting the material ready for my upcoming #GPGPU course that starts on March. Even though I most probably won't get to it,I also checked my trivial #SYCL programs. Apparently the 2025.0 version of the #Intel #OneAPI #DPCPP runtime doesn't like any #OpenCL platform except Intel's own (I have two other platforms that support #SPIRV, so why aren't they showing up? From the documentation I can find online this should be sufficient, but apparently it's not …)
CPU-GPU co-execution through the exploitation of hybrid technologies via SYCL
#SYCL #OpenCL #CUDA #LLVM #PerformancePortability #LoadBalancing #HybridComputing
hgpu.org · CPU-GPU co-execution through the exploitation of hybrid technologies via SYCLThe performance and energy efficiency offered by heterogeneous systems are highly useful for modern C++ applications, but the technological variety demands adequate portability and programmability.…
Exploring data flow design and vectorization with oneAPI for streaming applications on CPU+GPU
hgpu.org · Exploring data flow design and vectorization with oneAPI for streaming applications on CPU+GPUIn recent times, oneAPI has emerged as a competitive framework to optimize streaming applications on heterogeneous CPU+GPU architectures, since it provides portability and performance thanks to the…
HiPEAC 2025 kicks off next week and we're excited to be featured in two great sessions on Safety Critical and developing highly parallel applications. If you're attending HiPEAC in Barcelona, be sure to come and see us!
More information in this sessions: https://www.khronos.org/events/hipeac-2025
#SYCL #opencl #vulkan
The Khronos Group · HiPEAC 2025Deploying and developing royalty-free open standards for 3D graphics, Virtual and Augmented Reality, Parallel Computing, Neural Networks, and Vision Processing
We're used to leaning on children's books in Computer Science - with Gulliver's big-endian vs little-endian. Back at Supercomputing hashtag#SC24, I spoke at the hashtag#Intel booth all about open standards, performance portability, and the journey up the Yellow Brick Road to see the Wizard of Oz. Check out the video of the talk on YouTube:
https://youtu.be/xO8FGAOScpo?si=_BnVilvTBa0Ns6dX
#performanceportability #OpenMP #SYCL
Analyzing the Performance Portability of SYCL across CPUs, GPUs, and Hybrid Systems with Protein Database Search
#SYCL #oneAPI #Bioinformatics #Databases #HPC #PerformancePortability #Package
hgpu.org · Analyzing the Performance Portability of SYCL across CPUs, GPUs, and Hybrid Systems with Protein Database SearchThe high-performance computing (HPC) landscape is undergoing rapid transformation, with an increasing emphasis on energy-efficient and heterogeneous computing environments. This comprehensive study…
Performance portability via C++ PSTL, SYCL, OpenMP, and HIP: the Gaia AVU-GSR case study
#HIP #SYCL #OpenMP #CUDA #PerformancePortability #HPC #Astrophysics #Package
hgpu.org · Performance portability via C++ PSTL, SYCL, OpenMP, and HIP: the Gaia AVU-GSR case studyApplications that analyze data from modern scientific experiments will soon require a computing capacity of ExaFLOPs. The current trend to achieve such performance is to employ GPU-accelerated supe…
This Tuesday at Supercomputing, ANARI will have a BOF in Room B209 starting at 12:15pm.
Or join the SYCL BOF just down the hall in room B212. We hope to see you there.
More info: https://www.khronos.org/events/supercomputing-2024
#sc24 #SYCL #ANARI
The Khronos Group · Supercomputing 2024Deploying and developing royalty-free open standards for 3D graphics, Virtual and Augmented Reality, Parallel Computing, Neural Networks, and Vision Processing
Read all about it! News update with latest on #SYCL including a view of the future roadmap. https://www.khronos.org/blog/khronos-releases-sycl-rev-9-specification-and-roadmap-updates-for-sc24-conference
The Khronos Group · Khronos Releases SYCL Rev 9 Specification and Roadmap Updates for SC24 ConferenceThe SYCL™ Working Group at Khronos® has been hard at work advancing the ecosystem for this open, royalty-free, cross-platform abstraction layer that enables code for heterogeneous and offload processors to be written using modern ISO C++. Through active public development, SYCL continues to grow as developers contribute and provide feedback, shaping the future of this portable programming model. Recently, the SYCL Working Group announced the r