[Exophase]

Originally Posted by Arun

As for Vivante, Freescale claims some *extremely* impressive benchmark numbers for the GC2000 in i.MX61, but I'd like to see numbers on a platform with VSync on before passing judgement... Still much more impressive than anyone expected I think.

I went looking around for this and if anyone is interested see page 28 here:

http://www.freescale.com/files/32bit.../IMX6SRSFS.pdf

I'd love to know more about the architecture of Vivante's higher end GPUs. All I could really glean is that they're unified and that the highest end ones are "multicore", although from the block diagram I don't think this is in the full-on sense it is for SGX543MP.

BTW, the GC2000 in i.MX6 dual/quad runs at 533MHz:

http://www.engineeringtv.com/video/F...-ARM-Cortex-A9

[mboeller]

Originally Posted by Exophase

I went looking around for this and if anyone is interested see page 28 here:

http://www.freescale.com/files/32bit.../IMX6SRSFS.pdf

What page 28? The pdf has only 2 pages.

[Arun]

Originally Posted by mboeller

What page 28? The pdf has only 2 pages.

He meant this presentation: http://www.freescale.com.cn/cstory/f...1/pdf/1166.pdf
And yes, that's mostly where I got that data from Didn't want to link it because of what happened when I linked to the Qualcomm roadmap one and I didn't feel the GPU data was *that* valuable since it's clearly without VSync unlike the iPad, and I doubt the die size estimate for GC2000 (half iPad's 32nm) is even fair in any way (e.g. 40nm vs 45nm and excluding GC350 for OpenVG). The overall presentation definitely has truckloads of data on i.MX6x though which is nice. Sounds like an very strong product except for the fairly low (1.2GHz) clock speed on the A9s. One likely explanation is that Page 38 states "core voltage: 1.1v" which implies it's probably pure 40LP rather than 40LPG like NVIDIA.

The tables on Page 20 and Page 21 are very interesting when it comes to tablet TDP and 2xA15 vs 4xA19. They claim A15 has 25% less DMIPS/W than A9 - I don't think that's enough to justify 4xA9 given the subpar quad-core scaling, and the data on Page 22 seem even less reliable. Oh and is that A15 DMIPS number even right? 12K@2x2GHz implies only 3 DMIPS/MHz rather than the previously rumoured >3.5DMIPS/MHz. That seems awfully low. Is that right or is Freescale just making up all these numbers as they go?

[Exophase]

I've been kind of eagerly checking for i.MX6 availability. Unlike most other SoC manufacturers Freescale tends to make their i.MX series available in low quantity orders through distributors or directly, so it's a viable choice for low volume or even hobbyist hardware.

One thing I really like about i.MX6, outside of the quite competitive GPU and quad-core option, is just how much it has integrated. For instance, it has a lot of embedded regulators, meaning you only need to supply it two DC/DC regulated voltages, foregoing the need for complex PMICs. And it can put it I/Os on 3.3V or 1.8V, allowing easier peripheral interface with level shifters (especially for say, LCDs). It's also one of the few ARM SoCs this generation with DDR3 support.

Hope it's going to be available soon..

[slyd]

@darkblu:
yes certainly possible.


@Exophase:
i.MX6 is also available in industrial/automotive temp. grades - that's not the case for most other smartphone SoCs.

The memory Interface is 64 Bit DDR3-1066 / dual 32 Bit LPDDR2... Tegra 2+3 have 32 Bit LPDDR2 only (so speed should be like DDR2-800).


Here's a recent Video of it driving 4 Displays (2x 1080p + 2x XGA) at the same time:
http://www.youtube.com/watch?v=F74zo...eature=related

http://media.bestofmicro.com/B/O/310...inal/photo.JPG

Notice the "GiQuila"

[Ailuros]

http://www.youtube.com/watch?v=k4hN5...eature=related

Framecounter is badly visible in that video, but there's not a lot of action going on either way.

[Ailuros]

First GC1000 GLBenchmark2.1 results: http://www.glbenchmark.com/phonedeta...chmark=glpro21

MSAA still takes a significant drop. Is that product brief new or have I missed it in the past? http://www.vivantecorp.com/Product_Brief.pdf

At least according to that one it's clear now how the variants exactly scale in core amounts. Frequency of that GC1000 is unknown of course, but if it should be for the comparable region in the iMX6/GC2000 it still doesn't sound like a iPad2 killer.

Unless of course they're also having driver troubles.

[ToTTenTranz]

It seems to have twice the performance of the previous GC800.

According to the pdf that Ailuros pointed out, the GC1000 seems to be a dual-core GC800, with GC2000 being a quad GC800 and GC4000 being eight GC800 cores.

Looking at the synthetic scores and assuming a linear scaling with core amount, the GC2000 should be right next to Exynos 4210 in pixel shader speed and quite a bit better in vertex shader speed.. so the GPU to battle a SGX543MP2 should be the GC4000.

At 12.9mm^2, how does the GC4000 compare to the SGX543MP2 in size?

[Ailuros]

Originally Posted by ToTTenTranz

At 12.9mm^2, how does the GC4000 compare to the SGX543MP2 in size?

Apple's A5 doesn't help for such a comparison since it's a huge SoC die anyway. It's over 120mm2 big under Samsung 45nm and Apple obviously didn't care a bit about die area. Roughly 1/4th of the A5 are the 2 A9 CPUs@1 GHz and roughly 1/4th the block for the MP2 GPU block. In that regard you can't just claim that the MP2 there is over 30mm2@45nm, disregarding how generous Apple was with die area. Else the CPU block in A5 is also roughly over 30mm2, while Tegra2 as an entire SoC weighs 49mm2@40nm.

I don't have much hope to get any details for the Renesas SoC with the SGX543MP2, but it could probably help.

Looking at the synthetic scores and assuming a linear scaling with core amount, the GC2000 should be right next to Exynos 4210 in pixel shader speed and quite a bit better in vertex shader speed.. so the GPU to battle a SGX543MP2 should be the GC4000.

I think that's way too pessimistic. I'd guess that JohnH's estimate of the GC2000 probably ending in the =/>70% range of a SGX543MP2 shouldn't be too far from reality. That would give roughly twice the performance of the GC1000 which sounds reasonable. If true the GC4000 (that's 8 cores and since the product brief smells like single TMU/core it could mean 8 TMUs), I'd rather place that one against a SGX54xMP4, with the latter still being a notch ahead.

Their die area figures (if true) remain impressive.

[codedivine]

Vivante GPU on iMX6 achieves OpenCL 1.1 EP conformance:
http://www.vivantecorp.com/vtopcl.html

[Ailuros]

http://www.glbenchmark.com/phonedeta...tgroup=overall

Is that a iMX6 with a GC2000? At least it's listed as a GC2000 under results by GPU.

[Laurent06]

I guess it's this board: https://wiki.linaro.org/Boards/MX6QSabreLite

Triple Play Graphics system consisting of a Quad-shader 3D unit, and a separate 2-D and separate OpenVG Vertex acceleration engine for superior 3D, 2D and user interface acceleration

Not many info I'm afraid

[Ailuros]

Originally Posted by Laurent06

I guess it's this board: https://wiki.linaro.org/Boards/MX6QSabreLite



Not many info I'm afraid

The GC2000 is actually a quad core: http://www.vivantecorp.com/Product_Brief.pdf

I'm still wondering where those claimed floating point, triangle rates and fillrate get lost considering the results. However if their claimed die area should be close to reality and considering the list of API support it's still impressive hw.

[Brad Grenz]

Looks like the first quad core A9 Android stick I've seen in powered by this chip. $99 shipped: http://www.mcbub.com/item/Zealz-GK80...-Black-153763/