--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Sat, Apr 22, 2017 at 11:53 AM, sokolgeo@posteo.net wrote:
Hello. It seems like NEO900 project (http://neo900.org/) is having some problems with SoC warp soldering memory chips to their PCB:
specific details: http://lists.goldelico.com/pipermail/community/2017-February/001583.html http://lists.goldelico.com/pipermail/community/2017-February/001594.html
Luke - is there similar soldering risk in EOMA manufacturing?
fuck no :) whenever i see "PoP" as an option for a processor i run away very very fast :)
it's also why i haven't done a smart (or dumb) phone, because *without* that smaller size (due to "stacking") you simply can't make a small enough PCB for it to be socially acceptable, these days.
that's why i was so excited when i found the Ingenic M150... and so disappointed when they told me they'd stopped making it. why? because the M150 has 128mb of DDR2 RAM *on board* the same 11x11mm package, as a Multi-Chip-Module.
instead i've focussed on EOMA68 as the first modular design, which is *just* big enough to fit processor and RAM and PMICs onto one side of the PCB., and in some cases big enough to fit NAND or eMMC on as well.
so no - the primary reason why i avoid PoP RAM is because of the MOQ quantities usually being around 1,000.
*sigh* another expensive learning experience for the open / libre hardware community, there. i do wish sometimes that they'd read some of the things that i write. i've *already learned* that you don't try retro-fitting PCBs into pre-existing casework.
never mind - sometimes it's better for people to do things they won't listen on and then fail.
l.
that's why i was so excited when i found the Ingenic M150... and so disappointed when they told me they'd stopped making it. why? because the M150 has 128mb of DDR2 RAM *on board* the same 11x11mm package, as a Multi-Chip-Module.
Hmm... this reminded me of Next Thing Co.'s GR8 processor. It has a 14x14mm package with 256mb ram built in. I'm not sure they meet the libre requirements but it is worth a look (if you haven't seen it already): https://nextthing.co/pages/chippro
--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Mon, Apr 24, 2017 at 3:47 AM, Louis Pearson desttinghimgame@gmail.com wrote:
that's why i was so excited when i found the Ingenic M150... and so disappointed when they told me they'd stopped making it. why? because the M150 has 128mb of DDR2 RAM *on board* the same 11x11mm package, as a Multi-Chip-Module.
Hmm... this reminded me of Next Thing Co.'s GR8 processor. It has a 14x14mm package with 256mb ram built in. I'm not sure they meet the libre requirements
http://linux-sunxi.org/R8 - the A13's been reverse-engineered for a looong time.
but it is worth a look (if you haven't seen it already): https://nextthing.co/pages/chippro
i was aware of it but hadn't investigated it after hearing ridiculous things about pricing. the GR8 seems to be reasonable: $6. that's still quite high but it removes a lot of the hassle.
just found the ORCAD design files so... yyyeah.... that looks doable. thank you louis.
l.
Eomaphone hype
On 24 April 2017 10:59:25 GMT+03:00, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Mon, Apr 24, 2017 at 3:47 AM, Louis Pearson desttinghimgame@gmail.com wrote:
that's why i was so excited when i found the Ingenic M150... and so disappointed when they told me they'd stopped making it. why? because the M150 has 128mb of DDR2 RAM *on board* the same 11x11mm package, as a Multi-Chip-Module.
Hmm... this reminded me of Next Thing Co.'s GR8 processor. It has a
14x14mm
package with 256mb ram built in. I'm not sure they meet the libre requirements
http://linux-sunxi.org/R8 - the A13's been reverse-engineered for a looong time.
but it is worth a look (if you haven't seen it already): https://nextthing.co/pages/chippro
i was aware of it but hadn't investigated it after hearing ridiculous things about pricing. the GR8 seems to be reasonable: $6. that's still quite high but it removes a lot of the hassle.
just found the ORCAD design files so... yyyeah.... that looks doable. thank you louis.
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
On April 23, 2017 1:20:47 AM EDT, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
it's also why i haven't done a smart (or dumb) phone, because *without* that smaller size (due to "stacking") you simply can't make a small enough PCB for it to be socially acceptable, these days.
I would totally buy and use a phone large enough to fit an eoma68 CPU card. But there may not be wide enough appeal to get the costs down to a reasonable number.
There was a lot of hype around the ability to dock your phone into a desktop so you only have to carry one computer. It should be possible to ride off of that hype by demonstrating popping your CPU card out and putting it in a desktop/laptop housing and vice versa.
It just makes so much more sense to separate the core CPU card from the "peripherals" like you did with EOMA.
On Tue, Apr 25, 2017 at 12:55 AM, Adam Van Ymeren adam@vany.ca wrote:
On April 23, 2017 1:20:47 AM EDT, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
it's also why i haven't done a smart (or dumb) phone, because *without* that smaller size (due to "stacking") you simply can't make a small enough PCB for it to be socially acceptable, these days.
I would totally buy and use a phone large enough to fit an eoma68 CPU card.
i thought of a potential way to deal with that, and it's to allow the Card to stick out the side.
But there may not be wide enough appeal to get the costs down to a reasonable number.
well that's why i want to do a smaller card (EOMA50). honestly if spending the time on this i would prefer to do something that stands a chance of a larger appeal (and would reach an acceptable crowd-funding threshold)
l.
I'd say get EOMA68 cards to be fast first. Maybe even 64 bit octacore or something fancy like that. Then get the tablet out that has an EOMA68 slot. That will definitely be popular if the specs are good display/battery wise. If theres a laptop and theres a tablet, then there'll be proven customers for a phone (the people who funded the other two)
On 25 April 2017 05:18:31 GMT+03:00, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
On Tue, Apr 25, 2017 at 12:55 AM, Adam Van Ymeren adam@vany.ca wrote:
On April 23, 2017 1:20:47 AM EDT, Luke Kenneth Casson Leighton
lkcl@lkcl.net wrote:
it's also why i haven't done a smart (or dumb) phone, because *without* that smaller size (due to "stacking") you simply can't make a small enough PCB for it to be socially acceptable, these days.
I would totally buy and use a phone large enough to fit an eoma68 CPU
card.
i thought of a potential way to deal with that, and it's to allow the Card to stick out the side.
But there may not be wide enough appeal to get the costs down to a
reasonable number.
well that's why i want to do a smaller card (EOMA50). honestly if spending the time on this i would prefer to do something that stands a chance of a larger appeal (and would reach an acceptable crowd-funding threshold)
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
The octa core SoCs available right now are rather terrible when it comes to heat and power management( like the kirn 960) and we are rather limited by the fact that everything has to run on libre software
On Fri, Apr 28, 2017 at 10:31 AM, Allan Mwenda allanitomwesh@gmail.com wrote:
I'd say get EOMA68 cards to be fast first. Maybe even 64 bit octacore or something fancy like that. Then get the tablet out that has an EOMA68 slot. That will definitely be popular if the specs are good display/battery wise. If theres a laptop and theres a tablet, then there'll be proven customers for a phone (the people who funded the other two)
On 25 April 2017 05:18:31 GMT+03:00, Luke Kenneth Casson Leighton < lkcl@lkcl.net> wrote:
On Tue, Apr 25, 2017 at 12:55 AM, Adam Van Ymeren adam@vany.ca wrote:
On April 23, 2017 1:20:47 AM EDT, Luke Kenneth Casson Leighton lkcl@lkcl.net wrote:
it's also why i haven't done a smart (or dumb) phone, because *without* that smaller size (due to "stacking") you simply can't make a small enough PCB for it to be socially acceptable, these days.
I would totally buy and use a phone large enough to fit an eoma68 CPU card.
i thought of a potential way to deal with that, and it's to allow the Card to stick out the side.
But there may not be wide enough appeal to get the costs down to a reasonable number.
well that's why i want to do a smaller card (EOMA50). honestly if spending the time on this i would prefer to do something that stands a chance of a larger appeal (and would reach an acceptable crowd-funding threshold)
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
-- Sent from my Android device with K-9 Mail. Please excuse my brevity.
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
On Fri, Apr 28, 2017 at 10:40 AM, Bill Kontos vkontogpls@gmail.com wrote:
The octa core SoCs available right now are rather terrible when it comes to heat and power management
yyeah, that's no surprise. what's really nice about RISC-V btw is that it's something mad like a 40% performance-watt improvement over ARM in the same geometry.
( like the kirn 960) and we are rather limited by the fact that everything has to run on libre software
well, there are plenty of companies that _don't_ (so it makes no sense to copy them or compete with them).... and the burden is enormous on absolutely everybody. yet nobody's even *tried* an entirely libre approach *despite clearly knowing the advantages*!
http://www.h-online.com/open/news/item/Intel-and-Valve-collaborate-to-develo...
l.
Where does the 40% improvement come from ? I remember watching a video from Berkeley student where the improvement was "only" 11%.
The kirin 960 was a complete fail on their part. They downgraded the manufacturing node to one that has a lower frequency sweetspot but lower cost as well... still 14nm for whatever it's worth these days. On top of that, and especially on the gpu side they have very aggressive clock steps, even missing the efficiency sweet spot and resulting in peak draws of up to 10 watts. All for benchmarks on reviewing websites, most reviews just run a benchmark once and show the result but do not test overheating. Totally not suitable for eoma, no matter how much of a sucker for high performance I am.
On Fri, Apr 28, 2017 at 2:12 PM, Luke Kenneth Casson Leighton <lkcl@lkcl.net
wrote:
On Fri, Apr 28, 2017 at 10:40 AM, Bill Kontos vkontogpls@gmail.com wrote:
The octa core SoCs available right now are rather terrible when it comes
to
heat and power management
yyeah, that's no surprise. what's really nice about RISC-V btw is that it's something mad like a 40% performance-watt improvement over ARM in the same geometry.
( like the kirn 960) and we are rather limited by the fact that everything has to run on libre software
well, there are plenty of companies that _don't_ (so it makes no sense to copy them or compete with them).... and the burden is enormous on absolutely everybody. yet nobody's even *tried* an entirely libre approach *despite clearly knowing the advantages*!
http://www.h-online.com/open/news/item/Intel-and-Valve- collaborate-to-develop-open-source-graphics-drivers-1649632.html
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
--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Fri, Apr 28, 2017 at 12:31 PM, Bill Kontos vkontogpls@gmail.com wrote:
Where does the 40% improvement come from ? I remember watching a video from Berkeley student where the improvement was "only" 11%.
i read a paper somewhere where they'd put a RISCV64 design into silicon @ 40nm, ran it at 1.5ghz... but i have, if i am honest, a "fuzzy memory"
The kirin 960 was a complete fail on their part. They downgraded the manufacturing node to one that has a lower frequency sweetspot but lower cost as well... still 14nm for whatever it's worth these days. On top of that, and especially on the gpu side they have very aggressive clock steps, even missing the efficiency sweet spot and resulting in peak draws of up to 10 watts. All for benchmarks on reviewing websites, most reviews just run a benchmark once and show the result but do not test overheating. Totally not suitable for eoma, no matter how much of a sucker for high performance I am.
there's always the EOMA200 standard - one for the future (several more years into the future).
l.
arm-netbook@lists.phcomp.co.uk
-
Adam Van Ymeren -
Allan Mwenda -
Bill Kontos -
Louis Pearson -
Luke Kenneth Casson Leighton