[poll; updated] eMMC vs. uSD; modular eMMC

What would you prefer?

  • non replaceable eMMC - willing to pay a premium for it

    Votes: 22 24.7%
  • non replaceable eMMC - assuming roughly equal costs for both solutions

    Votes: 21 23.6%
  • uSD - assuming roughly equal costs for both solutions

    Votes: 23 25.8%
  • uSD - willing to pay a premium for it

    Votes: 23 25.8%

  • Total voters
    89

IF the failure rate of eMMC is very low, it might not be worth worrying about. Which is why I would like to know the failure rate.
It's still worth worrying about if you care about device longevity.


And even not considering that, microSD gives you the advantage to increase your OS storage size, easily move your OS into a new Pyra, (if your Pyra dies for some other reason, for example) or a way to easily switch your base OS on your device. (have multiple micro SDs for debian, arch, android, etc.) Not to mention, instead of having to spend time flashing the device before shipping, EvilDragon can have a premade image that he copies onto his microSDs that he just inserts into the device as soon as he gets it and ships it out.

-God Ginrai
I had a 1999 IBM laptop I was running ubuntu with that I had mutiple harddrives to test out linux distros at the time. All set up for that laptop. I could just pull one out and plug it in. From a production standpoint I think this approach has the most longivity. And I like the idea of being able to get options or have upgrade compatibility. But I'm just an average person with limited funds, and the idea of being able to constantly or periodically upgrade a device.. really sounds nice. 
 
The 128MB/s number I gave isn't a theoretical anything, that was measured. HardKernel claims 160MB/s. The measured speed can be artificially lower depending on access pattern and filesystem.
 
What about random access and latency ?
From the same website, random access was showing about 74MB/s. No idea about what latency would be or if it even matters as it would surely be smaller than human perception.
As far as replacing the microSD, to really be as user friendly as possible it'd probably be best to keep it under the battery just so it guarantees the power is off before they go mucking with their operating system drive but I'd be interested in some counterpoints.
 
As far as replacing the microSD, to really be as user friendly as possible it'd probably be best to keep it under the battery just so it guarantees the power is off before they go mucking with their operating system drive but I'd be interested in some counterpoints.
Agreed, that definitely does make the most sense. I'm fine with having it anywhere that's reachable, (side of the device, under the battery) and preferably somewhere where I don't have to use a screwdriver to get to it. But putting it under the battery will be the best way to prevent complaints from idiots about their card getting corrupted, or their system doing strange things, or so forth.

-God Ginrai
 
As far as replacing the microSD, to really be as user friendly as possible it'd probably be best to keep it under the battery just so it guarantees the power is off before they go mucking with their operating system drive but I'd be interested in some counterpoints.
Agreed, that definitely does make the most sense. I'm fine with having it anywhere that's reachable, (side of the device, under the battery) and preferably somewhere where I don't have to use a screwdriver to get to it. But putting it under the battery will be the best way to prevent complaints from idiots about their card getting corrupted, or their system doing strange things, or so forth.

-God Ginrai
I like this idea too. It falls inline with most cell phones with have swapable sims and honestly putting them in the same area would make it easier for people to know whats going on.. But label them or something. 
 
Over time , the failure rate for all your components in the Pyra will be 100%.  As much as you would like it to , it won't live forever. That's why some reasonable assessments need to be made.  
Id like to disagree, but thats of academic interest. The reasonable assessment would be that first and foremost battery and NAND will fail. These are part that get worse over time, but at the same time improve in performance when bought new in the future, so you have everything to gain by making them replaceable. And not locking yourself to one solution, possibly bringing the whole device down with it, means you can diversify on what kind of specs people can afford, so no need to set a fixed rate compromise.

Btw: Random IOPS on a uSD card doesnt even begin to touch the bus-speed treshold on the omap5, and thats what matters for OS use. You are lucky to get 5MB/s if that on serial extended writes... On most cards random reads and writes is really really grim, and it doesnt necessarily follow the Class rating. http://static.usenix.org/events/fast/tech/full_papers/Kim.pdf
 
Last edited by a moderator:
Everything does fail over time - I've just finished replacing the capacitors in a Sega Game Gear. Also importantly, HW also ages.  Will you still be using your Pyra over your Pyra 4 or equivalent in 20 years time?   

Anyway, that aside.

I'm leaning toward e-MMC because of the following reasoning.

It's faster.  Only some, maybe even half are going to want to boot another OS of a card.  Making every Pyra boot from uSD hampers everyone with a slower boot up - even those that didn't want it.    On the other hand an e-MMC arrangement provides everyone with a faster boot and still allows those that want to use another OS on a card to use one of the SD slots.   Call it the lesser of the two evils.
 
Last edited by a moderator:
With internal uSD there's the concern that the card could shake and lose contact at some important time. It'd be hard to get it as secured as the external SD slots.
 
Dragonbox could potentially not survive populating a 250 run batch with bad soldered flash. Just wonky flash is actually potentially worse, because it could go unnoticed out to customers, there is a lot of wishy washy stuff in the nand market.

Replaceable emmc is also a warranty against that happening full scale.
 
Last edited by a moderator:
Everything does fail over time - I've just finished replacing the capacitors in a Sega Game Gear. Also importantly, HW also ages.  Will you still be using your Pyra over your Pyra 4 or equivalent in 20 years time?
 I still play my Sega Genesis. And I would assume since you replaced the capacitors in your Sega Game Gear that you still pull that out and play it.

With internal uSD there's the concern that the card could shake and lose contact at some important time. It'd be hard to get it as secured as the external SD slots.
I think that's all comes down to proper design and part choice. The internal microSD slot never gave me a problem in my phones that used it.

-God Ginrai
 
^ Nope, bought and fixed to sell it eventually :)

I assume a lot of those that buy a Pyra will sell (or not use much) their Pandora if the Pyra is able to run the same apps and games. It's more of a question of , do you still use your old 386 PC over your new PC?

Anyway, that was an aside . I gave may reasonings for e-MMC .  Each to their own.
 
Last edited by a moderator:
I certainly want to have the ability to use my c64 now, and i like the idea of ricing it up with newly fabricated parts in 2014. Who is to tell me it makes sense, maybe not from some point of view, but the option does by the fact that there is a market for it. A lot of engineers made terribad choices for the game gear, lets leave the flame there ;) , and focus on doing it right with pyra
 
Last edited by a moderator:
Well, as I said,.  It's a choice between faster boot e-MMC and using an SD card for those that want to boot other OS's.  or slower boot uSD for everyone - just so some can use uSD instead of one of the SD card slots.   That's why I voted e-MMC
 
Well, as I said,.  It's a choice between faster boot e-MMC and using an SD card for those that want to boot other OS's.  or slower boot uSD for everyone - just so some can use uSD instead of one of the SD card slots.   That's why I voted e-MMC
I've sat through much longer boots than we are likely to experience w/ the Pyra, even if it came with a poor microSD for the OS. Hell, you ever turned on an android phone?

-God Ginrai
 
I really think the poll should feature a removable eMMC ala hardkernel option, for which there isnt a price premium, since you dont have to populate it and it can be added in by the user later on.

Paying less up front, for having options down the road, and a device less likely to wear out, is what i would vote for. Depends what one would sac for the space needed if it came to it.
 
Last edited by a moderator:
I'm not a patient person, boot time (& probably overall OS performance) matters to me  :p

Anyway, I've said my reasons. 
 
Last edited by a moderator:
E-sata is where you get that ;)   Just never shut down and you have no bootwait, its called uptime, linux is great :)
Yea, booting is a thing of the past. :p I rarely ever have to boot my Pandora due to suspend-to-ram. Same w/ my 3DS. It's been about 2 weeks since I last booted my Pandora. Why do we need to sacrifice the versatility that microSD brings just so we can make something that happens rarely (a boot) faster?

-God Ginrai
 
Because microSD makes less sense than regular SD for OS installs. This is due to low random read and write and generally worse performance. There isnt enough ram to run the whole OS in tempfs if you need more than 2GB total, but that is fast.

Correction, you can populate at least 8GB on 5432, thats as much as the controller can address. Ive seen 4 in the wild, so chips exist to do this. The dream is real.
 
Last edited by a moderator:
Back
Top