Peter McGuinness

Google Pixel Visual Core

In a world where much of the burden of smartphone differentiation rests on the quality of the camera, Google’s otherwise unremarkable Pixel 2 has been praised as one of the best cameras out there. This distinction comes in part from the inclusion of the Google-designed ‘Pixel Visual Core,’ a companion chip intended to “make it so your Snapchat pictures don’t … Read more

IoT redesigns Qualcomm in its own image

We’ve all seen that thing where people start to look like their dog. Apparently, it’s a real phenomenon that has produced lots of academic studies to try to explain it. No-one really knows for sure but the consensus opinion is that if an owner can’t find a dog that looks like them, they will change themselves to look like their … Read more

Machine learning: the new nirvana for FPGAs?

That’s the view of a significant number of players in the ML game, anyway. Even discounting the fact that this opinion is largely heard from companies with FPGAs to sell, it has gained enough traction that it has to be taken seriously;  Microsoft has based its entire Azure business strategy on FPGA-accelerated servers and Americo-Chinese company DeePhi, recently acquired by … Read more

Baidu Kunlun 昆仑

A few days ago, Baidu joined the ranks of vertically integrated companies to produce Kunlun,  their own machine learning accelerator. Announcing two versions of the Kunlun processor, while giving few details outside of their 260 TeraOPS raw performance and 100w power dissipation, Baidu signaled its intention to own all the machine learning components of its datacenters as well as expanding … Read more

Mali-V76 VPU from Arm.

Arm’s recent Premium Launch gave us an insight into upcoming products aimed at the higher end of the mobile and consumer device markets. In there with the new Cortex-A76 CPU and the Mali-G76 GPU (subscribers see related story) was the Mali-V76 VPU, a new video codec IP core that delivers 8K UHD encode and decode.  Integration of Arm’s IP engines … Read more

Imagination joins the AI race with new products

This week, following on from its ‘soft launch’ of the PowerVR Series2NX inferencing accelerator architecture late last year, Imagination Technologies has delivered on its promise of real products during 2018 with two IP cores announced. Just as significantly, IMG has confirmed that they have two customers, including one in China.  The AX2185 and AX2145 target high- and low-end applications, respectively, … Read more

Khronos builds out the ecosystem ready for NNEF launch

Khronos this week announced the launch of a project with Au-zone of Calgary, Canada to produce two conversion tools that will allow developers to import NNEF files into TensorFlow and Caffe2 as well as export NNEF files from those training frameworks. In line with Khronos’ recent policy, the tools will be made available in the second half of 2018 as … Read more

Arm and Nvidia spot a squirrel

Breathless reporting makes news of Nvidia's and Arm's collaboration on Xavier sound like earth shattering news.  Nvidia and Arm recently announced that they would be working together to make Nvidia’s Deep learning Accelerator (NVDLA) accessible to programmers using Arm’s Trillium platform. Project Trillium was announced by Arm in February this year as a way for developers to get direct access from … Read more

The evolving future of processors

We’ve been racing to keep pace with Moore’s Law for decades. GPUs have become our daily workhorse for visualization and aspects of compute. We’re close to adding FPGAs and quantum computing to that mix. This is the era of processors. Hardware is important again. The story of the last twenty years has been that everything is software; as long as … Read more

Arm and Nvidia spot a squirrel

Breathless reporting makes news of Nvidia's and Arm's collaboration on Xavier sound like earth shattering news.  Nvidia and Arm recently announced that they would be working together to make Nvidia’s Deep learning Accelerator (NVDLA) accessible to programmers using Arm’s Trillium platform. Project Trillium was announced by Arm in February this year as a way for developers to get direct access from … Read more

GTC: Nvidia’s evolution takes it into new territory with new customers and competitors

It’s ten years since Nvidia launched the GPU Technology Conference (albeit under a different name) to promote the use of their technology. This year, they are claiming an attendance of more than 8000 delegates along with one of the largest press and analyst delegations you will see outside of behemoths like CES and MWC.  What started as a bold statement … Read more