Does Kabini suffer the same 64-bit penalty as Brazos/Zacate did?

[Cerb]

Try opening several Newegg.com product pages, with Flash Player installed in Waterfox, and without NoScript or FlashBlock installed.

plugin-container.exe (wrapper .exe for Flash Player plugin) will balloon out to 2.7GB of memory, real quick.

Never seen it that high, but I can't handle not having Flashblock for more than a matter of minutes, once I hit any flash-using site with a new browser profile (IE, a new OS install). My near-1GB experiences are with Flashblock. I'm guessing that with more RAM, it doesn't GC much, not being denied new allocations.

 

[Abwx]

?? If we assume 30% IPC improvement, 1.4*1.3=1.82. So even the N2830 base clock will be higher performing than the E1-2500.

Assuming the N2830 doesnt somewhat throttle since it s a laptop, anyway the difference is small wich means that for the same tasks one cant be at 40-45% CPU usage and the other at say 10%, this would imply 4x the CPU perf in difference , wich is obviously not the case.

 

[ph2000]

64bit flash is pretty much cpu and memory hogger
for TS is this happen also when no flash playing ?

 

[ninaholic37]

I found this interesting:

What makes Waterfox fast?

Waterfox is compiled with Intel's C++ Compiler with the following optimisations: Intel's Math Library, SSE3, AVX for supported Intel processors, Profile-Guided Optimisation and the /O3 switch. Customisations to the base code were also made.

Intel's C++ Compiler huh?

Maybe try Cyberfox 64-bit and compare, it has an AMD version:

https://8pecxstudios.com/cyberfox-amd-version

 

[ShintaiDK]

I found this interesting:

Intel's C++ Compiler huh?

Maybe try Cyberfox 64-bit and compare, it has an AMD version:

https://8pecxstudios.com/cyberfox-amd-version

Its funny enough also the fastest compiler for AMD CPUs.

 

[ShintaiDK]

Just wondering. I'm running Win 8.1 64-bit on this E1-2500 Kabini 1.4Ghz dual-core, and the Waterfox 32.0.3 64-bit browser. It lags a lot more than my Bay Trail N2830 laptop, which is running Win 8.1 w/Bing 64-bit, running the same browser, with the same amount of RAM and HDD size / spec (though I don't think that they are the same exact HDD model).

I'm running the newest 14.9 WHQL Cat drivers for the IGP too.

There's just this... lag... with the AMD Kabini rig. It's kind of annoying. I mean, right-clicking and waiting for the menu to pop up. It always takes like a second.

I think that I remember reading that basically, Brazos supports 64-bit operation, but isn't optimized for it, in terms of cache size / width, bus width, opcode latency in 64-bit mode, something like that.

Is Kabini similarly handicapped?

I think you simply overestimate the capabilities of the ultra lowend CPUs you buy. As soon as its something that cant be offloaded it goes terrible wrong.

CPU wise the N2830 is something like 30% faster. And thats still a really slow CPU.

 

[ninaholic37]

There's just this... lag... with the AMD Kabini rig. It's kind of annoying. I mean, right-clicking and waiting for the menu to pop up. It always takes like a second.

Google gives a ton of interesting results when I search this, like:

http://www.online-tech-tips.com/windows-8/slow-right-click-context-menu/
http://www.eightforums.com/performance-maintenance/19373-right-click-menu-lag-my-pc.html#post187343

My dad's old AMD desktop does this too. He has a lot of junk so I'm guessing one of his extensions is slowing down the context menu (as the second link mentions).

 

[VirtualLarry]

I think you simply overestimate the capabilities of the ultra lowend CPUs you buy. As soon as its something that cant be offloaded it goes terrible wrong.

CPU wise the N2830 is something like 30% faster. And thats still a really slow CPU.

I agree that the N2830 seems faster, to me.

One thing that I should mention, is that this PC has "Sound Blaster Cinema", which is a sound processing / "enhancement" feature, and presumably, it takes a chunk of CPU time whenever audio is playing. In fact, you can sometimes hear the sound change, when your CPU load gets high. That may be why listening to internet radio takes 50% CPU time on this PC. I attribute most of that, though, to Flash Player bloat.

 

[pw257008]

Google gives a ton of interesting results when I search this, like:

http://www.online-tech-tips.com/windows-8/slow-right-click-context-menu/
http://www.eightforums.com/performance-maintenance/19373-right-click-menu-lag-my-pc.html#post187343

My dad's old AMD desktop does this too. He has a lot of junk so I'm guessing one of his extensions is slowing down the context menu (as the second link mentions).

very cool stuff

 

[MeldarthX]

VT are you running the gesture software in the background with everything else?

 

[VirtualLarry]

VT are you running the gesture software in the background with everything else?

It's not enabled. I selected it in the systray, and selected "exit", along with two other programs. That helped with the right-click lag, a bit.

 

[Abwx]

Some funny basic maths..

Considering that your E1 2500 plateform can use let say 20W at full usage measured at the main an elementary calculation point to an Athlon 5350 using perhaps 5W more wich should translate to 27W at the main, that is 35% more power comsumption at full load but for 200% more performance CPU wise, not counting that it will likely consume the same at equal task, that s an exemple that show that there s abolutely no point in shooting systematicaly for the lower clocked items as the rest of the plateform power usage is extremely dominant in this case.

 

[dorion]

I've been using a Kabini 5350 rig as a HTPC, running on Win 8.1 64bit. No lag whatsoever in Firefox, even while copying files in the background via a gigabit connection or making Windows updates.

I've been using the Sempron 3850 as a HTPC, I even play Civ 5 on it(in Strategic View), troubles with slow context menus sound more like a driver or OS install issue.

 

[Enigmoid]

Some funny basic maths..

Considering that your E1 2500 plateform can use let say 20W at full usage measured at the main an elementary calculation point to an Athlon 5350 using perhaps 5W more wich should translate to 27W at the main, that is 35% more power comsumption at full load but for 200% more performance CPU wise, not counting that it will likely consume the same at equal task, that s an exemple that show that there s abolutely no point in shooting systematicaly for the lower clocked items as the rest of the plateform power usage is extremely dominant in this case.

Yes. This is completely true. If you look at total system power the comparison gets even worse once you count the monitor.

 

[Abwx]

Yes. This is completely true. If you look at total system power the comparison gets even worse once you count the monitor.

Indeed, a 22" monitor will drain about 30-40W depending of the manufacturer,
efficency of thoses low power system is easily countered by saids monitors whose excess comsumption is often 10-12W due to an archaic power factor correction circuit wich is also used in most ATX PSUs and DC blocks, thoses losses are somewhat negligible with a 200W or more system but in low power builds they are the real factors, ahead of the CPU itself, for exemple the AM1 system i have was using 58W at iddle and 42W once changing the PSUs and using a frugal 22" monitor.

 

[escrow4]

Why do people bother with these trash CPUs? I don't get it. Life is too short to be twiddling your thumbs waiting for slow junk hardware to keep up. If a laptop or desktop doesn't have an i3 or i3 equivalent minimum, move on. These systems have no puff and never will.

 

[coercitiv]

That may be why listening to internet radio takes 50% CPU time on this PC. I attribute most of that, though, to Flash Player bloat.

I just did some subjective testing on my HTPC, I lowered my Athlon 5350 from 2Ghz to 1.2Ghz and tried browsing while 2 other tabs had Youtube videos running in the background. Firefox was struggling, badly, even if reported CPU load was around 50% equally spread on all 4 cores.

Getting the CPU back to nominal speed improved the situation back to a level where browsing performance was satisfactory, even with 1080p flash video playing in the background.

My advice for people considering either Kabini or Bay Trail for a cheap, low power system, is always go for 4 physical cores. In the case of Kabini, aim for Athlon only.

 

[AtenRa]

I just did some subjective testing on my HTPC, I lowered my Athlon 5350 from 2Ghz to 1.2Ghz and tried browsing while 2 other tabs had Youtube videos running in the background. Firefox was struggling, badly, even if reported CPU load was around 50% equally spread on all 4 cores.

Getting the CPU back to nominal speed improved the situation back to a level where browsing performance was satisfactory, even with 1080p flash video playing in the background.

My advice for people considering either Kabini or Bay Trail for a cheap, low power system, is always go for 4 physical cores. In the case of Kabini, aim for Athlon only.

Yeap, those low power dual cores CPUs are not ment for todays OS and applications, 1.6GHz quads and up are fine both Kabini and ATOM.

 

[AtenRa]

Just wondering. I'm running Win 8.1 64-bit on this E1-2500 Kabini 1.4Ghz dual-core, and the Waterfox 32.0.3 64-bit browser. It lags a lot more than my Bay Trail N2830 laptop, which is running Win 8.1 w/Bing 64-bit, running the same browser, with the same amount of RAM and HDD size / spec (though I don't think that they are the same exact HDD model).

I'm running the newest 14.9 WHQL Cat drivers for the IGP too.

There's just this... lag... with the AMD Kabini rig. It's kind of annoying. I mean, right-clicking and waiting for the menu to pop up. It always takes like a second.

I think that I remember reading that basically, Brazos supports 64-bit operation, but isn't optimized for it, in terms of cache size / width, bus width, opcode latency in 64-bit mode, something like that.

Is Kabini similarly handicapped?

Go in to BIOS and disable every energy saving setting, including Cool n Quiet, C6 etc. This is a low power dual core APU, it doesnt need all this crap.