amd has shown that the cpu power draw is pretty low. also clock for clock amd definetely has it.
Baytrail has a very powerful CPU and a reasonably powerful GPU. the power consumption numbers are impressive. Silvermont is definitely faster than Cortex A15. Jaguar is a capable core but draws a lot more power. Also Kabini has no Turbo. AMD's Kabini/Temash are badly handicapped by the lack of aggressive turbo implementation. Silvermont has a very good Turbo implementation.
In Cinebench R11.5 single thread a 2.4 Ghz Silvermont matches a 1.5 Ghz Jaguar. Thats because Jaguar has a 128 bit FPU while Silvermont has a 64 bit FPU. Even in multithread Cinebench R11.5 the Baytrail chip would be running close to 2 Ghz.
For integer performance Jaguar can compete with Silvermont but only if it matches clocks. Without turbo Kabini has no chance. But the most important fact is Silvermont is doing all this at less than half the power of Jaguar. Thats the Intel manufacturing advantage which AMD can do nothing about.
AMD is going to find it very difficult. from servers to desktops to laptops and down to the tablets AMD just has no respite from the Intel juggernaut.
The Tech Report measured 2W idle and AnandTech measured 0.8-1.2W for Z3770 running CB11.5 @ single-core. Are The Tech Report results just for the CPU or the entire device?
For integer performance Jaguar can compete with Silvermont but only if it matches clocks.
I am very interested in the battle royale between Intel and the Arm manufacturers over the next couple years. Now for me. I would like to see a dockable tablet with KB, USB, and video out running Win 8 Pro to replace our laptops. Feel like we are still one gen away. But getting closer.
with all that money and prowess they cant beat qualcomm's s800 on an aging 28nm process with their new atom products. If feels as if this process advantage is just way too overhyped!Exactly how I feel. I'm really impressed by Bay Trail's CPU performance, we're inching ever so closer to the point where super thin and super light windows 8.1 portables are a reality - and with a low cost. If you want core performance? You can get that too in super portable form factors.
Additionally, with Iris Pro we're inching closer to 1080p gaming being a "real" thing with the smallest form factors. I'm pretty excited about these prospects - it will get better another generation from now when intel is on 14nm. Intel has - hands down - the best technology in the world in terms of lithography process, and they're worlds ahead of any ARM SOC in that respect. Can't wait to see what Broadwell and Silvermont have to offer, should be good stuff.
It make take another generation or two to be -perfected- but i'm really impressed.
The Verge said:We tried the Android version of Grand Theft Auto III at 2560 x 1440 resolution with maxed out settings, and it ran smooth as butter on the new Atom chip — smoother than the Tegra 4 inside Nvidia's Shield gaming handheld. Bay Trail won't be your first choice for Windows gaming, but it's already impressive in mobile.
I think what he means is, AMD has an easy free boost to come by simply adding a turbo in the next version.
Except it's hard to run at high frequencies, especially if you have to raise the voltage significantly and your inherent dynamic capacitance is high. So maybe it's not just as easy as changing a knob/fuse?
In mobile turbo reduce the battery life , it s more of a marketing
feature for benchmarks purpose that an advatageous feature
for the user since reducing execution time by 40% imply increasing
comsumption by roughly 60% , hence the energy/task ratio is degraded.
The 3-4W was platform power during video playback. That's still not all that good.
Precisely. Turbo speeds come at a cost and AMD hasn't demonstrated any ability to keep power consumption down at turbo/frequency levels.
That's actually an excellent number with a 10" 2560x1440 display.
The 3-4W was platform power during video playback. That's still not all that good.
The iPad gets ~15 hours, this would be ~10 hours given everything else is equal.
This just isn't accurate. Everyone has their name for it "hurry up and go idle, race to halt, etc." Further, the 60% number is arbitrary and depends on the frequency/voltage curve of the device. Do you have some information about the voltage delta between 1.5GHz and 2.4GHz on Bay Trail?
True, it's not good, merely comparable to the competition (iPad 4 is around 3.16W on Anand's 720p video playback battery life test.) Which in this case is a good thing considering that video playback is one of the weak points for Haswell ULT at closer to 5.2W. Of course there are too many variables there to draw a firm conclusion that it's comparable, but it's at least a good indication.
The iPad gets ~15 hours, this would be ~10 hours given everything else is equal.
....wiiiiith 15% less pixels.
wiiiiith no information on reference battery size.
Stop hating.
Seriously, it' gets tiresome.
This is a serious step - and it's not a perfect ultimate home run but it's a damn well serious step that should cast respect.
Honestly at this point what's the point of ARM?
Inferior design with an inferior OS... Whats to stop Intel?
The complaints about Windows 8 are from desktop users, tablet isn't nearly as poorly received it's just the hardware performance hasn't been there but it's quite obviously coming quite quickly.
You're confusing "hating" with questioning. Asking questions instead of meekly swallowing is what makes us different.
I have already said MANY TIMES in this thread that this is a really nice chip. Now it's time to look at what isn't so great.
It really depends on whether it's closer to 3W or 4W I guess, but yes "decent" is the word I think.
Power does appear to have taken a (very) small step backwards compared to Clover Trail, but obviously the SoC is far superior.