Active Topics

 


Reply
Thread Tools
Posts: 3,319 | Thanked: 5,610 times | Joined on Aug 2008 @ Finland
#21
Originally Posted by nashith View Post
@mrojas, I have experienced a lot of failures based on flash degradation on various smartphones utilizing flash for storage of vital system information (not removable flash). But since Maemo is Linux, I was wondering if such a scenario happens, can we partition a µSD with a swap and configure Maemo to use that partition (will be slower, but could work right?)
You can make that happen, but it's quite a bit of trouble for no real gain. You're swapping on a 32GB eMMC device, so, as beaten to death in other threads, there is no realistic chance of wearing that out in the device's lifetime.
 

The Following 3 Users Say Thank You to attila77 For This Useful Post:
Posts: 147 | Thanked: 42 times | Joined on Sep 2009
#22
Originally Posted by texaslabrat View Post
clock for clock they should be very comparable in raw processing as they are both based on the ARM cortex a8 I believe. At least, that's what I gather from this:
http://www.engadget.com/2009/10/14/c...ehind-android/
He kinda lost me at "But the Cortex A8 is built using 65nm transistors, which makes it roughly twice the size (and thus twice the cost to manufacture) of an ARM11 core (which has 90 nm transistors)".
 
Posts: 3,319 | Thanked: 5,610 times | Joined on Aug 2008 @ Finland
#23
Just replace 'size' with 'area' mentally, the point is it grows as a square. 45nm would have been one quarter of the 'size' of 90 nm.
 
Posts: 329 | Thanked: 142 times | Joined on Oct 2009
#24
Originally Posted by Nexus7 View Post
He kinda lost me at "But the Cortex A8 is built using 65nm transistors, which makes it roughly twice the size (and thus twice the cost to manufacture) of an ARM11 core (which has 90 nm transistors)".
I'm pretty sure it means size as in "number of transistors". Smaller transistors on the same-sized chip = more of them. And yes, 65*2 is not 90, because, as it has been pointed out, it's not something linear.
 
Posts: 4 | Thanked: 2 times | Joined on Oct 2009
#25
With regards to the CPU speed, the OMAP3 in the Pandora has been known to happily overclock to 900mhz (cpufreq support for the OMAP3 is in kernel mainline) so with any luck decent overclocks may be possible on the N900 too. One thing to bear in mind though is that the CPU on the N900 has no heatsink, the case has no air vents and you really don't want to fry your £500 phone!!
 
javispedro's Avatar
Posts: 2,355 | Thanked: 5,249 times | Joined on Jan 2009 @ Barcelona
#26
Originally Posted by mickeyjaw View Post
With regards to the CPU speed, the OMAP3 in the Pandora has been known to happily overclock to 900mhz (cpufreq support for the OMAP3 is in kernel mainline) so with any luck decent overclocks may be possible on the N900 too. One thing to bear in mind though is that the CPU on the N900 has no heatsink, the case has no air vents and you really don't want to fry your £500 phone!!
Heh. It's been said it fries at 600Mhz already. The builtin CPUFreq forces it to spend most of the time at < 500Mhz frequencies.
 

The Following User Says Thank You to javispedro For This Useful Post:
Posts: 119 | Thanked: 110 times | Joined on Sep 2009 @ Prague
#27
don't forget about the powerVR graphics, and integrated DSP +ISP... I don't know how much they're currently used - but I would guess the situation could be similar to that on PCs - that the additional hardware is even more powerful than the main CPU...

Must be madness to cool down such hardware (it's even cooler to think about that this hardware is probably more powerful than regular pc's were just 10 years ago )
 
GeneralAntilles's Avatar
Posts: 5,478 | Thanked: 5,222 times | Joined on Jan 2006 @ St. Petersburg, FL
#28
Originally Posted by mickeyjaw View Post
With regards to the CPU speed, the OMAP3 in the Pandora has been known to happily overclock to 900mhz (cpufreq support for the OMAP3 is in kernel mainline) so with any luck decent overclocks may be possible on the N900 too. One thing to bear in mind though is that the CPU on the N900 has no heatsink, the case has no air vents and you really don't want to fry your £500 phone!!
Heat isn't really the issue. Overvoltage is.
__________________
Ryan Abel
 
GunnerzMate's Avatar
Posts: 70 | Thanked: 11 times | Joined on Oct 2009
#29
Originally Posted by GeneralAntilles View Post
Heat isn't really the issue. Overvoltage is.
Тhats just what i was thinking.Is there any possible way to manipulate the voltages?
 
Posts: 271 | Thanked: 220 times | Joined on Sep 2009
#30
Originally Posted by Nexus7 View Post
He kinda lost me at "But the Cortex A8 is built using 65nm transistors, which makes it roughly twice the size (and thus twice the cost to manufacture) of an ARM11 core (which has 90 nm transistors)".
Speaking in very rough numbers, the cortex A8 has about 4 times the number of transistors as ARM11. However, when cortex A8 is manufactured on a 65nm process, the actual size of the chip is roughly twice the size of the ARM11 core when the ARM11 is manufactured on a 90nm process and thus twice as expensive as that ARM11. If they were both manufactured on a 65nm process to maintain an apples-to-apples comparison...the Cortex A8 would be about 4 times as large, and thus about 4 times as expensive to make as that hypothetical 65nm ARM11.

Without going into *too* much detail... large volume IC manufacturing costs are very strongly linearly related to the size of the chips (within the same technology "generation/family" eg SOI, strained silicon, copper interconnects, etc) due to the built-in costs of the wafers they are made from as well as the staggering up-front capital costs for the equipment.

Hope that clears it up a bit.
 
Reply


 
Forum Jump


All times are GMT. The time now is 16:54.