Samsung's ARM roadmap lays coordinates through 2013: Aquila, Venus, and Draco (oh my)
Ah, leaked company presentation slides, they have a clarity only Mr. Blurrycam would despise. EETimes got a batch of them from Samsung dated November 2009 making the rounds, but more important than revealing its equal love for both Roman and Greek mythology, we get a glimpse at its then-planned ARM chip roadmap (yeah, another one) through 2013. In a nutshell, for the Cortex A9 crowd we've got the 800MHz dual core "Orion" due for mass production in Q1 2011, a 1GHz single core "Pegasus" for Q4 2011, a 1GHz dual core "Hercules" for Q1 2012, and for sometime in 2012 / 2013, a 1.2GHz dual core "Draco" and quad core "Aquila." Fear not, Cortex A5 fanatics, you've got gifts as well, in the form of 600MHz single core "Mercury" and dual core "Venus" chips, slated for 2010 / 2011 and 2012 / 2013, respectively. We don't expect the nomenclature to extend beyond internal usage, but frankly, who cares -- it's the devices that count, and unfortunately all we can do is doodle our future gadget hopes and dreams onto scraps of paper while we wait.
























Quad core in mobile phone ftw!
!!! QUAD DAMAGE !!!
@Kloc
Cortex A9 + Dell Lightning plox!
@Kloc
I think the Motorola XT701 Android phone lists Quad Core A8 cpu on the specs
http://nexus404.com/Blog/2009/12/19/motorola-xt701-xt800-specs-features-revealed-new-china-bound-motorola-android-phones-details-available/
So... one day, Hercules, Pegasus and Orion walked into a bar and ordered some chips...
wait... what?!
@who said what And an Apple employee left his prototype iPhone at the bar.
@who said what
And did some ARM wrestling
Cool! Can't wait for the A9 to become mainstream. Out of order processing FTW.
Multi-core on mobile phone great!!! Laptop's minimum specs is quad core!! Soon...
ARM > ALL
Where is KRATOS ? :x
1.2 Ghz is sluggish for 2013.
@JS
WHOOSH!!! There went the point!
@JS Yeah... don't compare MHz in ARM devices, just don't.
For example, Cortex-A8, running at the same clock speed as ARM11 (last gen stuff like what's in the Hero) can complete approximately 40% more instructions per second, and that doesn't even start to touch all the stuff underneath, like the significantly more efficient ARMv7 instruction set (to ARM11's ARMv6), memory bandwidth improvements, power efficiency, etc.
Seriously, there's so much going on underneath it all that comparing clock speeds or even instructions per second is pretty much pointless. More pointless than comparing clock speeds on AMD and Intel processors, in fact.
Suffice to say that a multi-core Cortex-A9 will do a lot more with 1.2 GHz than a Cortex-A8 or Snapdragon will. And let's not even get into what will happen once manufacturers start to tweak Cortex-A9 or Qualcomm introduces a next-gen SoC...
Quadcore on 4th Gen iPads with 21 inch screens. WIN!!! then I get downranked. Anyone that will click that minus sign will have a week's Badluck.
@staticjethro
21" is way too big.
@n8equalsd Thats what she said.
@n8equalsd That's not what she said.
@n8equalsd
I'm just playing you know.
We already knew about the S5PV210, it’s a sister chip to the S5PC110 “Hummingbird” in the Samsung Wave and Galaxy S, and is primarily intended for tablets or MIDs. It was announced summer of last year when they also introduced the S5PC110, though I’m surprised it’s not in any products already. I suspect that the S5L8930 in the Apple A4 is a modified S5PV210, but that’s just my theory.
Oh, I should also point out to those who don't know, Cortex-A9 is going to stomp the hell out of Cortex-A8 processors like those in the Droid, Palm Pre, N900, iPhone 3GS, etc.
It's also going to destroy Snapdragon, which is similar to Cortex-A8 but is really a custom architecture designed by Qualcomm (based upon the ARMv7 instruction set they licensed from ARM). However, given the amount of time and money they've invested into Snapdragon, I'm sure Qualcomm will make efforts to keep Snapdragon looking competitive.
But Cortex-A9 is going to be on a whole different level. Dual-channel memory controller, more efficient instruction handling, out-of-order execution, power efficiency improvements and lots of other goodies, including some stuff I'm sure I haven't read about yet.
I’ve been considering picking up a Samsung Galaxy S this summer with its 45 nm Cortex-A8 goodness. However, TI has talked about releasing it's 1 GHz dual-core Cortex-A9 OMAP 4440 late 2010/early 2011, though I'm not sure if that means releasing to manufacturers or actually on the market. Probably the former.
We should be seeing a 1.3 GHz, 45 nm refresh of the Snapdragon we all know and love (dubbed QSD8650A) in the second half of this year. I've also seen claims that we'll see dual-core Snapdragons (QSD8672) this fall / holiday season which should be fun; rumor has it that one will end up in a flagship WM7 device.
Exciting stuff. If only I could afford to plunk down $600 on an off-contract smartphone so that I'm not tied down with a contract when next-gen tech starts hitting the market, because when Cortex-A9 shows up, I'll be waiting...
@Electrofreak 90% I'm getting the Galaxy S, hummingbird is a powerhouse, I hope they'll be able to fully utilize it and get some special partnerships going on to fully utilize that Super AMOLED screen. That's part of my thought, I see the chips as important, but that screen is 6-12 months ahead of the competition so I have no problem using a previous generation's great chips along with current screen technology in 18 months when it's time to start considering an upgrade.
@juanvaldez I wouldn't even really call the Hummingbird last-gen as it seems pretty obvious it's going to be about a year before we see Cortex-A9 in a smartphone (particularly given that this leak was for netbooks and they will probably get the chips before smartphones).
In fact, Hummingbird is the first released Cortex-A8 or Snapdragon SoC released on the 45 nm scale, so that makes it pretty cutting-edge in it's own right. 45 nm SoCs consume less power, generate less heat, and can be clocked at higher speeds than their currently-available 65 nm bretheren.
Let's look at it this way; 45 nm SoCs are about 20-30% more efficient than 65 nm ones. Given that the 1 GHz Hummingbird S5PC110 is an evolution of the 833 MHz 65 nm S5PC100 (which, downclocked to 600 MHz is in the iPhone 3GS), we're looking at a clock bump of 20%. Factor in the logic gate enhancements that improve CPU calculation efficiency by nearly 10%, and it's pretty competitive to Snapdragon's 5-15% lead over Cortex-A8 (my estimate based upon a number of factors, including a 5% improvement in instructions per second, limited out-of-order execution support, expanded SIMD bus, etc). What we know about the S5PC110 is still pretty limited though, so it's very possible that other improvements have been made to Hummingbird that allow it to push past Snapdragon in performance; and in fact Samsung has made claims that it does.
I'll bet you that it'll overclock pretty well too if you're willing to risk it. Whether it'll hit a stable 1.3 GHz like the 45 nm Snapdragon will remain to be seen, but Hummingbird is definitely a strong rival for Snapdragon, and has the current 65 nm chips beat in terms of performance.
The claim of 90 million triangles per second from the Galaxy S GPU indicate that Samsung has done something interesting with the graphics solution (possibly involving multiple-cores and/or large GPU cache sizes), because the Hummingbird's single-channel memory controller and LPDDR2 could not possibly be providing the 12.6 GB/s memory bandwidth I estimate it would need to have with an SGX 540 tripled in clock speed to 600 MHz to produce 84 Mt/s. Samsung is either fudging numbers, feeding us some sort of very brief non-sustainable value, or has something impressive up their sleeve. I must be missing something because Qualcomm's dual-core Snapdragons will apparently push 80 Mt/s as well.
Okay, this is where I shut up... I love talking about this crap because the ARM manufacturing industry is always clouded in secrecy and the attentive and perservering (or, in my case, obssessed) can sometimes tease out details they're not supposed to know about...
is netbook the codeword for mobile nowadays? none of those are above 1.0GHz...aka cannot run a full OS
@artshark
with my limited knowlege (compared to others on this site), i will attempt to explain why your claim is wrong and irrelevant.
These cortex a8 and a9 processors are not designed for a full OS, rather a mobile OS. You also may have noticed the multi core processors, which increase the power and split the work, thus making it more powerful. Another thing described above in previous comments, is that clock speed isnt everything. i dont actually know what changes, but the next generation proccessors can be more powerful at the same clock speed, as well as being more efficient.
By the way, netbook is mobile, but these are for smartphones and Tablets, so these are not for netbooks. Also, W7 can run on low power processors. i'm not sure exactly what the minimum is, but there are still small netbooks/smartbooks with 1.3GHz and lower.
@masta vaan While I don't have enough knowledge to pat you on the back, I'll say you probably did a fine job for a humble man.
But let's dissect the comment further and take it in context of what Samsung is trying to do:
Firstly, depending on the accuracy of language in an article I'll bring up, the top end of this may actually be more powerful than anything we are seeing in netbooks these days, if that's true I guess that's the end of the conversation here; but I'll go on...
"aka cannot run a full OS" - This, obviously, isn't being done for OSX, so do you currently consider Windows the only other full OS? "The new chips represent Samsung's desire to move away from a Wintel - Windows OS on Intel chips - platform for its netbooks and towards a future where its ultra-portables run Ubuntu and Chrome on Samsung's own processors."1 If you don't think Ubuntu and Chrome are full OSes then perhaps we should stop the conversation here, but I'll continue...
"including an up-coming quad-core chip that might finally close the performance gap between netbooks and notebooks."2
To dismiss all of these as "unable" to run a full OS seems a bit premature. Yes, it's not a core i3, but it doesn't have to be. Also, full OSes, while obviously now "more complete [full]", have been run in the past on less effecient and powerful processors. I'm not sure what you're completely getting at, and if you need a full (read: Windows) OS you might still be getting an Intel or AMD processor. But that will add cost due to the processor and the OS amongst other things.
Lastly, you also fail to consider what the cloud will, potentially, mean for Chrome OS and even Windows in the future. Should they chose, your processors clock speed will be of lesser importance than they are these days as they could serve you data; it's just the way the market will dictate what is the cheapest and most effective way to distribute power that will be the judge of that.
1&2 both quotes from http://www.bit-tech.net/news/hardware/2010/04/23/samsung-plans-own-netbook-chips/1
addendum: fine, perhaps the quadcore, but i can't see the others leading to a very good W7 experience
You think clock speed is everything in a CPU? Think again. The Cortex A8 in iPhone 3GS clocked at 600 Mhz is much more powerful than Snapdragon I tells ya!