Telecom Lead America: Vivante Corporation announced that
Vivante GC Cores has passed the Khronos Group OpenCL 1.1 Embedded Profile (EP)
conformance test suite on Freescale’s i.MX 6 platform.
The GC Cores use the latest programmable ScalarMorphic
architecture to accelerate parallel data workloads on thousands of concurrent
threads to achieve Gigaflops (GFLOPS) of computational performance.
Applications taking advantage of Vivante cores to
significantly speed-up processing includes image processing, computer vision,
analytics, augmented reality and gesture-motion tracking.
OpenCL uses the Vivante parallel execution SIMD engines
to enhance computation density by performing massively parallel data processing
on multiple data streams, across several compute engines. Each compute engine
has its own ALU, including pipelined floating point (FP) and integer (INT)
units for math operations, plus a special function unit (SFU) for transcendental
operations.
“We are pleased that Vivante passed OpenCL 1.1 EP
using GPU acceleration on Freescale’s i.MX 6 triple-play graphics
platform,” said Rajeev Kumar, i.MX product line manager for Freescale
Semiconductor.
The OpenCL (Open Computing Language) API takes advantage
of all resources to efficiently process the growing number of incoming data
from multiple sources including cameras, sensors, IPUs and other complex data
streams that require data or task parallelism. The standard targets a wide range
of devices including smartphones, tablets, PCs and CE devices, to embedded
applications like automotive ADAS (Advanced Driver Assistance System).
“New innovations inside our GPUs that accelerate
OpenCL include enhanced multi-level memory access support, synchronization,
math functions, ULP precision and atomics. Our advantages in mobile GPU
architectures give us the opportunity to bring high performance compute to low
power form factors, without sacrificing computational performance or features,”
said Wei-Jin Dai, president and CEO of Vivante.