Epic Citadel On Clover Trail+

rflcptr said:
Hmm, to cross the threshold of impressiveness we need a dual-core Atom?
Click to expand...

In our pants. Yes!

I am actually interested in having something like a Galaxy 3 with a decent gaming capable output to a tv for example. So use the phone as a controller on the tv. Instant mini console.
 
spaceman said:
In our pants. Yes!
Click to expand...
So it was originally... Epic Fail?
I am actually interested in having something like a Galaxy 3 with a decent gaming capable output to a tv for example. So use the phone as a controller on the tv. Instant mini console.
Click to expand...
I could have sworn there was another phone capable... :p
 
Not very useful for comparison as I can already run this on my nexus4 at 60fps..
 
We need more demanding demos for these new phones. Even my dinosaur gnex eats this one up

 
My Nexus 10 can run this at 2560x1600 at 60fps so... this isn't really impressive.

Let's see it do that!

The PowerVR chip grafted on that "upcoming" chip is slower than a great number of already available chips so this is really not too impressive.
 
I have to say, that is pretty impressive.
Looks like the Atom has come a long ways since the days of the N270.
 
The A4 in the iPhone 4 was doing this 2+ years ago. I don't see how this is impressive.
 
I think for years people have wondered if Intel could get x86 into a power envelop close to ARM what was capable of running x86 most decently. I think the next version of the Atom, Bay Trail, just might be the trick. While Clover Trail certainly isn't a speed demon and can't power much in the way of higher end desktop software, the fact that x86 devices that are as think and light as many ARM tablets that can run full x86 desktop OSes is pretty impressive, while getting the battery life of ARM devices.

Now Intel just needs to get the speed cranked up. It looks like Bay Trail on paper is promising roughly the performance of Sandy Bridge ULVs from 2011. If that's the case and the Bay Trail devices are as thin and light and a power efficient as Clover Trail devices or better and the prices are good, Intel could have a very interesting part on it's hands.

Intel has a lot ridding on Atom, it is for now their mobile strategy and they have some aggressive timelines and goals to reach.
 
Ugh this demo again. Nevermind it's lacking actual stuff that games need like characters or anything moving really... It's kind of silly "hey you can run 20% of a real game". :D
 
When I first ran it on my galaxy nexus I was amazed! I thought we were nearing PS2 quality with phones, but I see that we surpassed it.

But on topic: my phone which is a year+ old can run this almost the same, i'm not really impressed by that intel.
 
heatlesssun said:
I think for years people have wondered if Intel could get x86 into a power envelop close to ARM what was capable of running x86 most decently. I think the next version of the Atom, Bay Trail, just might be the trick. While Clover Trail certainly isn't a speed demon and can't power much in the way of higher end desktop software, the fact that x86 devices that are as think and light as many ARM tablets that can run full x86 desktop OSes is pretty impressive, while getting the battery life of ARM devices.
Click to expand...

Done having your "Whatever" Trail Atom processor wetdream yet? :rolleyes:
While it is impressive for x86 to achieve this power/performance ratio, I do agree with the others that ARM processors have been doing this for years.


Now Intel just needs to get the speed cranked up. It looks like Bay Trail on paper is promising roughly the performance of Sandy Bridge ULVs from 2011. If that's the case and the Bay Trail devices are as thin and light and a power efficient as Clover Trail devices or better and the prices are good, Intel could have a very interesting part on it's hands.
Click to expand...

Today's best Atom is only marginally faster than a 330 from 2008; yes the power envelope is better, but certainly not the performance.
Now, suddenly, the next Atom will have the performance of Sandy Bridge?

No. Just, no.
 
heatlesssun said:
Yes, just yes.
Click to expand...

Because white paper specs have always been 100% accurate.
An Atom with the performance of Sandy Bridge... where do you come up with this bullshit?! :confused:

Yeah, maybe an 800MHz Sandy Bridge with one core enabled.
 
Bay Trail chips are expected to offer twice the performance of the Clover Trail processors powering Windows 8 tablets today, while offering lower power consumption (and heat generation). That means we should see tablets with longer battery life and thinner cases — Intel says you can build a Bay Trail tablet that’s just about 8mm (0.3 inches) thick.
Click to expand...

Twice the performance of Clover Trail along with support for SATA SSDs which I believe it will support in addition to an HD 4000 offshoot GPU should very well put an Atom based tablet within close performance range of a Sandy Bridge ULV based device.
 
heatlesssun said:
Twice the performance of Clover Trail
Click to expand...

That alone makes no sense.
Clover Trail itself isn't "twice the performance" of a four year old Atom 330 dual-core from 2009.

I'm sorry, but your beliefs in these white paper specs of these upcoming Atom processors are making me laugh.
Bay Trail will be lucky to be 20% faster than Clover Trail in real world applications, and maybe 50% in general.


From PassMark:

Intel Atom Z2760 @ 1.8GHz - 679 (Q2 2012)
Intel Atom 330 @ 1.6GHz - 597 (Q1 2009)

Clover Trail had four years to gain a little over 12% in average performance, and you're telling me that by Q3 2013 that Bay Trail will be twice as fast as Clover Trail, with near Sandy Bridge performance???
wtf are you smoking... :rolleyes:
 
If doubled the overall performance of this:

Atom%20Z2760%20WEI%208%20.png


not sure why it's unreasonable to think you wouldn't be pretty close to this:

Sandy%20Bridge%20i5-2467M%20WEI%208.png
 
heatlesssun said:
If doubled the overall performance of this:

Atom%20Z2760%20WEI%208%20.png


not sure why it's unreasonable to think you wouldn't be pretty close to this:

Sandy%20Bridge%20i5-2467M%20WEI%208.png
Click to expand...

Sorry, but that Windows CPU benchmark is bunk, and wildly inaccurate at times.
As an example, I've seen instances where single-channel DDR2 gets a higher score than dual-channel DDR3.

Intel Atom Z2760 @ 1.8GHz - 679

Intel Core i5 2467M @ 1.6GHz - 2304

That means the i5 is over 70% faster on average at a slower clock rate.
Even if Bay Trail was literally twice as fast as the existing Clover Trail, it would still be 41% slower than the i5 @ 1.6GHz.

I'm sorry, but you have no clue what you are talking about, and your blind faith in white paper specs only shows your complete ignorance of this tech.


Sly said:
Is it being rendered in software mode?
Click to expand...
Just as heatlesssun, I'm sure he'll know. :rolleyes:
 
Red Falcon said:
Clover Trail had four years to gain a little over 12% in average performance, and you're telling me that by Q3 2013 that Bay Trail will be twice as fast as Clover Trail, with near Sandy Bridge performance???
wtf are you smoking... :rolleyes:
Click to expand...

You do realize that the two CPUs you linked are basically identical right? You do realize that Intel has done almost zero work on Atom since its introduction. The only major differences have been more device integration (reduced cost to Intel) and reduced idle power consumption. They have done precisely nothing on performance improvements. The newer result you link is 12% faster... because its clocked 12% faster. They got that for free when moving from 45nm to 32nm.

BayTrail will push Atom to 22nm, finally catching up with IvyBridge. That gives maybe 10-20% headroom in clockspeed (leaks show turbo up to 2.7ghz). As well BayTrail is quad core. So yeah, were already at 2x theoretical. Now add in the most important change in Atom's history... out-of-order execution finally. Feel free to lookup the performance improvements between Cortex A9 and A15 as ARM included out-of-order.

I'll wait for the real benchmarks personally, but your responses seem really uninformed.
 
mabellon said:
You do realize that the two CPUs you linked are basically identical right? You do realize that Intel has done almost zero work on Atom since its introduction. The only major differences have been more device integration (reduced cost to Intel) and reduced idle power consumption. They have done precisely nothing on performance improvements. The newer result you link is 12% faster... because its clocked 12% faster. They got that for free when moving from 45nm to 32nm.

BayTrail will push Atom to 22nm, finally catching up with IvyBridge. That gives maybe 10-20% headroom in clockspeed (leaks show turbo up to 2.7ghz). As well BayTrail is quad core. So yeah, were already at 2x theoretical. Now add in the most important change in Atom's history... out-of-order execution finally. Feel free to lookup the performance improvements between Cortex A9 and A15 as ARM included out-of-order.

I'll wait for the real benchmarks personally, but your responses seem really uninformed.
Click to expand...

I'll wait to see what the real benchmarks show as well, but uninformed or not, I think you guys are getting your hopes up.
Yes, it will move to a new architecture, and out-of-order execution will be huge, but putting it on par with Sandy Bridge... those clock speeds better be at at least 2.7GHz if it hopes to achieve this.


You do realize that Intel has done almost zero work on Atom since its introduction.
Click to expand...
Tell that to heatlesssun.
He's the one that's been going on and on about Clover Trail and its mighty performance compared to other Atom processors, which, as you stated, are all basically the same thing (which they are).

Who's uninformed? :rolleyes:
 
Red Falcon said:
Tell that to heatlesssun.
He's the one that's been going on and on about Clover Trail and its mighty performance compared to other Atom processors, which, as you stated, are all basically the same thing (which they are).

Who's .uninformed? :rolleyes:
Click to expand...

When it comes to running Windows 8, from performance to battery life to heat output, the Samsung Clover Trail tablet I've been using everyday for the last four months now is far superior compared to the HP Slate 500 Z540 Atom tablet that I also run Windows 8 on. In almost the same weight with the same capacity battery the Samsung runs all of my common desktop applications much better and gets twice the battery life. Metro doesn't work at 1024x600, the native resolution of the HP Slate 500, though it is possible to emulate 1024x768 and run Metro apps. However it's a bad experience.

All I've done is report my real world experience and I never said that the Samsung performed much better solely because of the Z2760 or how mighty Clover Trail is. I would imagine Ion based Atom devices would do much better. But nothing in those benchmarks tells a person how fast and accurate the handwriting recognition is or how well a web page scrolls or pinch zooms and that's what I was talking about, not theoretical benchmarks.
 
Sly said:
I asked because i'd credit the performance to the GPU rather than the Atom.
Click to expand...

Red Falcon said:
haha, yeah I would too.
Click to expand...

I've mentioned this plenty of times as well. Windows 8 is much more GPU dependent than even Windows 7 so that would explain much of the performance gains. Still a platform is a platform, not just a CPU. You can't just look at CPU benchmarks and actually determine how well an platform runs end to end. And for a lot of things that one would do a tablet such as Samsung 500T or HP Slate there are no benchmarks. Show me Inkmark bench that quantifies inking responsiveness and handwriting recognition accuracy.

If you look at just benchmarks you'd never appreciate the face that a device can run much better on the same sized battery and get twice the battery life.
 
heatlesssun said:
When it comes to running Windows 8, from performance to battery life to heat output, the Samsung Clover Trail tablet I've been using everyday for the last four months now is far superior compared to the HP Slate 500 Z540 Atom tablet that I also run Windows 8 on. In almost the same weight with the same capacity battery the Samsung runs all of my common desktop applications much better and gets twice the battery life. Metro doesn't work at 1024x600, the native resolution of the HP Slate 500, though it is possible to emulate 1024x768 and run Metro apps. However it's a bad experience.
Click to expand...

That's because the Z540 was paired with a awful north bridge that consumed far more wattage than the CPU; the N270 was the exact same way.
Also, that processor and resolution were released in Q2 2008, before even Win 7 emerged on the scene, so I would hope that your Clover Trail tablet from 2012 would outperform and have better battery life than a tablet from 2008.

Your point is...? :rolleyes:


All I've done is report my real world experience and I never said that the Samsung performed much better solely because of the Z2760 or how mighty Clover Trail is. I would imagine Ion based Atom devices would do much better. But nothing in those benchmarks tells a person how fast and accurate the handwriting recognition is or how well a web page scrolls or pinch zooms and that's what I was talking about, not theoretical benchmarks.
Click to expand...
Handwriting recognition doesn't eat many CPU cycles, hence why I haven't seen any serious benchmarks for it.
That would be like comparing how fast an i5 is against an Atom on how quickly it can calculate 2+2 = 4 on CALC.exe; in short, who gives a fuck.

The accuracy wouldn't be determined by the processor, it would be by the screen type, pen type, and controller.
Do you have a clue how any of this tech works? At all???

According to what you just said, a faster processor means more accurate touch sensors. Right. :rolleyes:
Also, on how fast a web browser scrolls: any modern processor will handle this well, even an ARM; there are no need for benchmarks for such a simplistic task.

There are no benchmarks for such tasks, because, who the hell cares if one processor outperforms another in web page scrolling by 0.0000000000023ns. :rolleyes:
This is why stand-alone synthetic and real-world (such as rendering and transcoding) benchmarks exist; they test and push processors and systems to their limits, and give a general number to compare the system tested against other systems.

Dude, you are so not a computer enthusiast, professional, hobbyist, or IT 'anything'.
You are a hipster, aka, one who dabbles in surface-level, consumer-based, media-consumption technology (fucking tablets), and has no real concept of how any of the back-end components/programs/OSes truly function or interact with one another, nor their purposes.

You thought Microsoft Server 2012 Server Core was a CLI-only OS and was GUI-less:
Sconfig-in-Windows-Server-2012-Server-Core.png


(psst, here's a hint, that still has a GUI...)


What truly amazes me about you is that you aren't an Apple user; that I could believe with what you say and do.
But you're a pure-Microsoft user, strictly on the Microsoft side.

I expect better from the Microsoft side. Much, much better.
 
I think we've learned three things here:

1. Clover Trail looks promising.

2. Epic Citadel is not the most hardcore graphics benchmarking app in the world.

3. If you want to build an AMD-based Linux box, talk to Red Falcon.
 
Red Falcon said:
Your point is...? :rolleyes:
Click to expand...

I'm talking about real world use and you're talking about synthetic benchmarks. That's my point.


Red Falcon said:
Handwriting recognition doesn't eat many CPU cycles, hence why I haven't seen any serious benchmarks for it.
Click to expand...

That of course depends on the CPU. Core i5/i7 CPUs obviously show much less of a spike than of course Atoms but inking is quite a bit more responsive on this Samsung 500T than the HP Slate. I'm not saying that that's all due to the Z2760 but again something simply runs better in the real world.
 
Red Machine D said:
3. If you want to build an AMD-based Linux box, talk to Red Falcon.
Click to expand...
Damn straight.


heatlesssun said:
I'm talking about real world use and you're talking about synthetic benchmarks. That's my point.
Click to expand...
I gave examples of real-world benchmarks such as rendering and transcoding, you know, things that actually use more than 1% of the CPU.
According to you, real-world benchmarks are how fast a web page scrolls and how accurate a touch screen is; almost 26,000 posts on here, and this is all the better you can muster?

I'm talking about people who know what they are talking about and you're talking about hipsters. That's my point.
I may not know everything, but at least I'm not basing my knowledge upon make-believe benchmarks that would have been relevant in the mid-90's, back when scrolling a webpage probably would have been a big thing for a processor to handle.

Oh that's right, I forgot you operate on tablets with no storage and pathetic processors.
Kind of like a kick-back to 90's computing, no wonder you want benchmarks for how fast a webpage scrolls, it's all the more your CPU can handle. :rolleyes:

...and this is coming from the guy with an AMD Linux box. :eek:

Pro tip: don't judge a person by their sig, heatlesssun has a sweet gaming system listed in his sig and yet all he cares about are benchmarks that might eat 1% of a tablet processor.
So if one has an Intel system running Windows in their sig, that somehow automatically makes them credible?

heatlesssun could have a supercomputer in his sig, and it would not make his posts any more credible.
 
heatlesssun said:
That of course depends on the CPU. Core i5/i7 CPUs obviously show much less of a spike than of course Atoms but inking is quite a bit more responsive on this Samsung 500T than the HP Slate.
Click to expand...
Should I even mention that you are comparing two tablets from 2012 to one from 2008?

I'm not saying that that's all due to the Z2760 but again something simply runs better in the real world.
Click to expand...
"Something simply runs better in the real world", you mean like the controllers, drivers, and touch functionality that have all vastly improved in that time-frame?
I'll state again, those two Atom processors have a 12% performance difference on average, which is what I stated, and yes, you were saying that it was all due to the Z2760; go back and read your own fucking posts Mr. Wishywashy.

Again, your point is... just to apparently waste our time with your nonsense.
 
You must log in or register to reply here.
Back
Top