{"id":19349,"date":"2021-09-23T15:33:12","date_gmt":"2021-09-23T13:33:12","guid":{"rendered":"https:\/\/ritme.com\/le-traitement-heterogene-exige-la-parallelisation-des-donnees-sycl-et-dpc-sont-un-bon-debut\/"},"modified":"2021-09-24T13:00:51","modified_gmt":"2021-09-24T11:00:51","slug":"heterogeneous-processing-requires-data-parallelization-sycl-and-dpc-are-a-good-start","status":"publish","type":"post","link":"https:\/\/ritme.com\/en\/heterogeneous-processing-requires-data-parallelization-sycl-and-dpc-are-a-good-start\/","title":{"rendered":"Heterogeneous Processing Requires Data Parallelization: SYCL and DPC++ Are a Good Start"},"content":{"rendered":"\n
<\/div>
\n
<\/div>
\n
<\/div>
<\/div><\/div>
\n

James Reinders<\/span><\/strong><\/p>\n<\/div><\/div><\/div><\/div><\/div>\n\n\n\n

\"\"<\/a><\/figure>\n\n\n\n
<\/div>
<\/div><\/div>
<\/div><\/div><\/div><\/div><\/div>\n<\/div><\/div>\n\n\n\n
<\/div>
\n

Engineer at Intel, has more than three decades experience in parallel computing, and is an author\/co-author\/editor of ten technical books related to parallel programming. James works at Intel promoting parallel programming in a heterogeneous (XPU) world.<\/p>\n\n\n\n

Here is his article published on The New Stack<\/a><\/em>.<\/p>\n<\/div><\/div>\n<\/div><\/section>\n\n\n\n

I like to say that \u201cIt\u2019s all about XPUs.\u201d<\/p>\n\n\n\n

We are in a wonderful time when hardware innovation is leading to an explosion in CPUs, GPUs, FPGAs, DSPs, ASICs and more\u2014which I simply abbreviate as XPUs<\/strong>. XPUs<\/strong> is a shorthand for any type of \u201cprocessing unit\u201d \u2014any hardware that can help my application compute.<\/p>\n\n\n\n

As developers, the onslaught of XPUs<\/strong> means that we are increasingly challenged to code for a larger collection of diverse processing units<\/strong>. We are tasked with factoring in extra time, and money, to rewrite and test code to boost application performance for new architectures.<\/p>\n\n\n\n

More than ever, to preserve our sanity and the maintainability of our code, it is paramount that the code we write is applicable to as many XPUs as possible. Moving to cross-architecture models for application development has shown that this can save organizations significant time and money<\/strong>, and this becomes an even more pressing concern with the rise in popularity of heterogeneous computing.<\/p>\n\n\n\n

Underway today is a rethinking because our world is rapidly becoming a world of XPUs<\/strong> that will eventually transform all of computing.<\/strong><\/p>\n\n\n\n

XPUs: Reinventing Software for Accelerated Compute<\/strong><\/h2>\n\n\n\n