ARM's new GPU design aims to improve performance tenfold

Leeky

Posts: 3,357   +116

ARM has released a press statement today announcing that its upcoming Mali-T658 GPU chip is set to offer up to ten times the performance of the current Mali-400 MP, which is found in phones such as the Samsung Galaxy S II and is the most powerful GPU present on a mobile device today. The new chip builds on the Mali-T604, which itself is a next generation chip currently in development, but will offer a fourfold performance increase over that one as well.

ARM's Mali-T658 GPU will pave the way for faster games, desktop-like performance and the possibility of more complex software becoming a reality on smartphones and tablets. The firm expects the performance of its new chip to equal the video performance of a Playstation 3 games console.

"Next generation consumer devices based on the Mali-T658 GPU will address the growing user expectation for slick user interfaces and desktop-class graphics," said Pete Hutton, General Manager of ARM's Media Processing Division. "Intuitive user interfaces will mean that consumers can access the full functionality of their connected devices, for richer user experiences. This includes HD gaming and new compute-intensive applications, such as augmented reality."

arm gpu mali-t658

The new design should be available for widespread use in about two year's time and manufacturers will be able to link up to eight cores together with the new chip -- in comparison to the four cores maximum with the current Mali-T400 chip.

As with previous Midgard architecture GPUs, the new Mali-T658 chip will support a wide range of graphics and compute APIs, including Microsoft DirectX 11, Khronos OpenGL ES, OpenVG, Khronos OpenCL, Google Renderscript and Microsoft's DirectCompute.

"The idea is that when the full performance is required, you power up all of those cores. But having a multi-core architecture means that when the performance needs are lower, we can power some of those cores down," said Jem Davies, of ARM's Media Processing Division.

The chip designer has also stated it will work with all upcoming CPU designs in perfect harmony, as well as alongside the new ARMv8 64-bit architecture announced to the world last week.

Just last month, ARM published outstanding third quarter earnings, which saw profits double for the U.K. Cambridge-based company despite a slowdown in the industry.

Permalink to story.

 
I was fairly certain the GPU performance crown belonged to Apple's A5 SoC... or the Tegra 3 (there are some benchmarks on the net).
 
The new design should be available for widespread use in about two year's time
Two years makes a big difference. This is the same timeframe as Logan (Tegra 5), which is planned for release in 2013 and is alegedly 50 times faster than Tegra 2 or 10 times faster than Tegra 3.

When this hits the market, I don't think it will be very competitive (assuming both company's predictions are acurate).
 
@Apple's A5 SoC

(From Wiki - http://en.wikipedia.org/wiki/Apple_A5)

The A5 contains a rendition of a chip based upon the dual-core ARM Cortex-A9 MPCore CPU with NEON SIMD accelerator and a dual core PowerVR SGX543MP2 GPU.
 
spydercanopus said:
Imagine if Nvidia or ATI made that claim about their next GPU.

AMD told the press their bulldozer is crazy fast

Except they forgot to mention you need nuclear cooling in-order to power it fast enough.

But i dont think the radiation and heat is worth it.
 
Wait... someone please link me to the comparison on the current mobile GPUs... was Mali-400 MP really considered more performant (silently) than Tegra 2, or was that statement about performance correct in the line of ARM designs?

And remember PS3 GPU is of raw Nvidia 7000+ series gen. it is reasonable to think that moble GPU will reach this or surpass it by 2013-14 for the screen resolutions supported. Under the same strain probably will not push much more than 720 high fadelity content... and will be a battery muncher regardless of the GPU manufacturer...
 
He said GPU...

Anyway, I'm also kind of confused as to why AMD or Nvidia make every new graphics generation 1.5 to 2 times faster than the previous one, while ARM and PowerVR claim anything from 10 to 50 times more speed (in every generation for the past few years)?

I think it's some sort of marketing gimmick. Perhaps their calculations are based on some huge hypothetical SoC with 24 cores and over 10W in TDP :D
 
Back