Three years from now, two-thirds of all new desktop systems will be mutli-GPU capable and of those, 30 percent will be rocking multiple graphics chips. Or at least that's the not-too-distant future Jon Peddie Research Group (JPR) laid out last week in a report on the history, technology, and future of multi-GPU computing. But are we really on the verge of widespread multi-GPU computing?
Not so fast, says Arstechnica. The JPR report points to the desire for high performance computing as the driving force for multi-GPU setups, noting high performance workloads are highly parallel and unsuited for CPU applications. But according to Arstechnica, JPR hasn't thought through the manufacturing angle.
"GPUs are composed of many parallel processing units, so any multi-GPU system involves simply ganging together still more of such small, simple processor cores," Arstechnica writes. "Because the cores are small and the workload is parallel, there is no limit on core count analogous to the limit on the number of processors that can profitably be used in a single x86 CPU. The limits on single-die GPU horsepower are manufacturing limits."
But it's not just about manufacturing. As Ars points out, only two percent of all desktop PCs sold last year came with multiple GPUs, and in Q4 of last year, only 15.2 million out of 38.5 million PCs sold came with even a single discrete graphics card. It's hard to imagine such a dramatic shift towards multiple GPUs in just three short years from now.
There's more to Ars' argument, which you can read here .
Image Credit: techreport.com