something interesting I saw is that in the update picking a processor, it shows rk3188 as the rockchip processor you were going to reverse engineer. on the other hand, your rhombus-tech link shows that your looking at rk3288?
Not to be annoying constantly, but I am curious are you looking at one or both?
I appolgize if there is something I may have missed.
On Sun, Apr 16, 2017 at 10:56 PM, zap zapper@openmailbox.org wrote:
something interesting I saw is that in the update picking a processor, it shows rk3188 as the rockchip processor you were going to reverse engineer. on the other hand, your rhombus-tech link shows that your looking at rk3288?
Not to be annoying constantly, but I am curious are you looking at one or both?
just the 3288. the 3188 i considered but it's older, cortex a9 based so is *really* power-hungry. i got a dev board (ok an IPTV box), opened it up, immediately noted the heat-sink on the 3188 and went "absolutely not".
the 3288 on the other hand was the first commercial A17 (i believe) which is a sort-of uprated version of the Cortex A7 - something like that, anyway.
also it can do HDMI 2.0, 4k video playback (if you push the dual-channel memory up to a stonking 1600mhz that is...) and it kicks the stuffing out of the high-end variants of the intel atom.
more than that, its popularity in chromebooks has meant that the 3288 has a *lot* of modding and OS messing-about behind it. that's a good thing in that it's well-understood, but it does make it a f*****g nuisance to try and find decent instructions. i've given up on using google search and now go directly to #linux-rockchip on freenode. people there know what they're doing.
l.
On 04/16/2017 06:53 PM, Luke Kenneth Casson Leighton wrote:
On Sun, Apr 16, 2017 at 10:56 PM, zap zapper@openmailbox.org wrote:
something interesting I saw is that in the update picking a processor, it shows rk3188 as the rockchip processor you were going to reverse engineer. on the other hand, your rhombus-tech link shows that your looking at rk3288?
Not to be annoying constantly, but I am curious are you looking at one or both?
just the 3288. the 3188 i considered but it's older, cortex a9 based so is *really* power-hungry. i got a dev board (ok an IPTV box), opened it up, immediately noted the heat-sink on the 3188 and went "absolutely not".
the 3288 on the other hand was the first commercial A17 (i believe) which is a sort-of uprated version of the Cortex A7 - something like that, anyway.
also it can do HDMI 2.0, 4k video playback (if you push the dual-channel memory up to a stonking 1600mhz that is...) and it kicks the stuffing out of the high-end variants of the intel atom.
more than that, its popularity in chromebooks has meant that the 3288 has a *lot* of modding and OS messing-about behind it. that's a good thing in that it's well-understood, but it does make it a f*****g nuisance to try and find decent instructions. i've given up on using google search and now go directly to #linux-rockchip on freenode. people there know what they're doing.
Glad to hear it! The rk3288 does sound faster judging by the specs and considering what you said, just now I mean, it is the better choice.
I wonder how fast it will be with the eoma standard? Probably at least 4x faster than the fastest processor on that "x200 libreboot device I am using."
Yeah I know its not as good as what your planning, but I can wait for ya. This is good news.
Also, it is a quad core processor and uses very little wattage hah.
I am curious by the way, how much does the intel crap use for wattage traditionally...
Probably an insane amount I imagine. :)
(I am glad because that will give more power to arm in the future.) ;)
Thanks for the info hehe... I look forward to debian 9 on that.
l.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
Intel is on about the same ballpark as the big arm tablet chips but they can't shrink it any further. They got this low simply from node shrinks, but at this point making a new core design only for the tablet market would require very high sale volumes. And they failed to infiltrate the tablet market so...They even canceled their atom line, only the premium Core M for products like the 12' apple macbook, which don't seem to gather much traction either.
On Tue, Apr 18, 2017 at 11:41 PM, Stefan Monnier monnier@iro.umontreal.ca wrote:
4x faster than the fastest processor on that "x200 libreboot device I am
I highly doubt it would be nearly that fast (at least for "general computing").
Stefan
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
4x faster than the fastest processor on that "x200 libreboot device I am
I highly doubt it would be nearly that fast (at least for "general computing").
Intel is on about the same ballpark as the big arm tablet chips but they can't shrink it any further. They got this low simply from node shrinks,
I think we're talking about different things: the X200 is a laptop from a several years ago, so it's unaffected by Intel's current performance: the kinds of processors it can take has been frozen for many years already (and they're much more power hungry than the RK3288, AFAIK).
Stefan "owner of a X201s which runs about 8 times faster than a Cubietruck when compiling Elisp code. I would expect the RK3288 to be maybe 4 times faster than the A20 on this benchmark (twice the number of cores and twice higher frequency), so to get to «4x faster» it would need another factor 8 on this particular test, which seems highly unlikely"
On 04/18/2017 05:09 PM, Stefan Monnier wrote:
4x faster than the fastest processor on that "x200 libreboot device I am
I highly doubt it would be nearly that fast (at least for "general computing").
Intel is on about the same ballpark as the big arm tablet chips but they can't shrink it any further. They got this low simply from node shrinks,
I think we're talking about different things: the X200 is a laptop from a several years ago, so it's unaffected by Intel's current performance: the kinds of processors it can take has been frozen for many years already (and they're much more power hungry than the RK3288, AFAIK).
Stefan "owner of a X201s which runs about 8 times faster than a Cubietruck when compiling Elisp code. I would expect the RK3288 to be maybe 4 times faster than the A20 on this benchmark (twice the number of cores and twice higher frequency), so to get to «4x faster» it would need another factor 8 on this particular test, which seems highly unlikely"
Hmm, your probably right, I since checked the speed of my processor, and it is 2.4 per processor and its a dual core.
But, the arm processor is 1.8 per processor and its a quad core. So if I add it together, probably it would be additional 50% faster.
and considering what you said, extremely lightweight by comparison, I am not sure how many more time lightweight, but I assume it would be like ten times at least if not a lot more...
Also it would be good for long term use, where as the libreboot is probably not gonna last that long. Everyone who replied to me probably knows this and more probably already. But good luck reverse engineering that. :) it will be well worth it if you succeed.
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
Hmm, your probably right, I since checked the speed of my processor, and it is 2.4 per processor and its a dual core. But, the arm processor is 1.8 per processor and its a quad core. So if I
You can't usefully compare the frequency of processors that are internally so completely different.
Stefan
On 04/21/2017 02:18 PM, Stefan Monnier wrote:
Hmm, your probably right, I since checked the speed of my processor, and it is 2.4 per processor and its a dual core. But, the arm processor is 1.8 per processor and its a quad core. So if I
You can't usefully compare the frequency of processors that are internally so completely different.
Stefan
okay, I'll bite what do you think the difference would be in speed?
since you know more ;)
arm-netbook mailing list arm-netbook@lists.phcomp.co.uk http://lists.phcomp.co.uk/mailman/listinfo/arm-netbook Send large attachments to arm-netbook@files.phcomp.co.uk
arm-netbook@lists.phcomp.co.uk
-
Bill Kontos -
Luke Kenneth Casson Leighton -
Stefan Monnier -
zap