Flash write cycles

Discussion around the SwissMicros DM42 calculator
User avatar
akaTB
Posts: 794
Joined: Tue May 02, 2017 1:56 pm
Location: Milan, Italy

Re: Flash write cycles

Post by akaTB »

Walter wrote:
Mon Dec 18, 2017 6:28 am
rprosperi wrote:
Mon Dec 18, 2017 12:04 am
I wish they could do something about delivery times to USA... :(
Look at it as "justice of international trade" - usually, we in Europe have to wait for delivery from Chin... pardon ... America first! And feel lucky that the seller is willing to ship abroad ... SCNR
Amen to that! ;)
Greetings,
    Massimo
ajcaton
-+×÷ left is right and right is wrong :twisted: Casted in gold
toml_12953
Posts: 795
Joined: Wed May 03, 2017 7:46 pm
Location: Malone, NY USA

Re: Flash write cycles

Post by toml_12953 »

Walter wrote:
Mon Dec 18, 2017 6:28 am
rprosperi wrote:
Mon Dec 18, 2017 12:04 am
I wish they could do something about delivery times to USA... :(
And feel lucky that the seller is willing to ship abroad ... SCNR
I don't need a broad...I need a calculator! :o
Tom L

Some people call me inept but I'm as ept as anybody!
DM10L SN: 059/100
DM41X SN: 00023 (Beta)
DM41X SN: 00506 (Shipping)
DM42 SN: 00025 (Beta)
DM42 SN: 00221 (Shipping)
WP43 SN: 00025 (Prototype)
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: Flash write cycles

Post by Walter »

toml_12953 wrote:
Mon Dec 18, 2017 1:44 pm
Walter wrote:
Mon Dec 18, 2017 6:28 am
rprosperi wrote:
Mon Dec 18, 2017 12:04 am
I wish they could do something about delivery times to USA... :(
And feel lucky that the seller is willing to ship abroad ... SCNR
I don't need a broad...I need a calculator! :o
You need both: Computing with a gifted broad may be significantly more fun than without - YMMV depending on age and fitness.
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
mcc
Posts: 277
Joined: Fri Jun 23, 2017 5:10 am

Re: Flash write cycles

Post by mcc »

Hi,

suppose you are currently writeing a programm into your DM42.
At what moment the electronic does decide to write to the flash?

Since a flash is (as far as I know) a block device (512byte sized?)
it would be every time if 512 byte of "new data" are available.
Or does it write at each line of code - that is each block would
be written more than once.
Or -- with Thomas Okkens app on Android in mind -- it is written at
that moment one press EXIT (that is switching of the DM42).

By the way ... '42' is the answer to all question, the life, the universe
and everything. Why? "42" is decimal ASCII value of "*", which in turn
is a placeholder for ... everything in regular expressions.

How wonderful coincidence that HP choose this number long before
a hitchhiker was travelling through the galaxy... ;)

Cheers
Meino
DM 42 - SN: 00373, Firmware release v.:3.22. / DMCP 3.24. as compiled by SwissMicros
User avatar
revwillie
Posts: 133
Joined: Mon Apr 24, 2017 7:48 pm

Re: Flash write cycles

Post by revwillie »

mcc wrote:
Tue Dec 19, 2017 5:13 pm

How wonderful coincidence that HP choose this number long before
a hitchhiker was travelling through the galaxy... ;)
agree with everything you said. However I think I read HHGTTG before 1987, or maybe my memory is worn out from too many beer cycles.
DM42 SN: 00076
DM15, HP 28c, HP 12c, HP 10b, HP 20b, HP 30b, HP 17bII, HP 42s, HP 35s, WP 34s, NP-25, TI-83+, teenix multicalc
User avatar
Walter
Posts: 3070
Joined: Tue May 02, 2017 11:13 am
Location: On a mission close to DRS, Germany

Re: Flash write cycles

Post by Walter »

revwillie wrote:
Tue Dec 19, 2017 5:50 pm
mcc wrote:
Tue Dec 19, 2017 5:13 pm

How wonderful coincidence that HP choose this number long before
a hitchhiker was travelling through the galaxy... ;)
agree with everything you said. However I think I read HHGTTG before 1987, or maybe my memory is worn out from too many beer cycles.
I confirm it has been published before 1987. IIRC it must have been before 1984. Anybody knowing more precisely?

EDIT: Wikipedia is my friend: HHGTTG was published in 1979 - German translation in 1981.
WP43 SN00000, 34S, and 31S for obvious reasons; HP-35, 45, ..., 35S, 15CE, DM16L S/N# 00093, DM42β SN:00041
User avatar
Guenter
Posts: 168
Joined: Wed May 24, 2017 6:26 pm
Location: Germany

Re: Flash write cycles

Post by Guenter »

mcc wrote:
Tue Dec 19, 2017 5:13 pm
Hi,

suppose you are currently writeing a programm into your DM42.
At what moment the electronic does decide to write to the flash?
As soon as you decide to write someting to the flash memory, either via "SETUP/File/Save Program" or via "SETUP/Calc. State/Save State". The calculator itself uses SRAM on the SoC http://hhuc.us/2016/files/Speakers/Gunt ... 476RG.pdf

Günter
Günter
DM42 SN:00004 and SN:00184 -- DM41X SN:00013 and SN:00955
mcc
Posts: 277
Joined: Fri Jun 23, 2017 5:10 am

Re: Flash write cycles

Post by mcc »

Hi Günter,

Thank you for the information -- and for the link of the MCU...that is the food I need! :)

Again I am very positevly surprised of the DM42 and its design.
From your info I conclude (ohoh...I am no native speaker and this /sounds/
very arrogant and posh...without really knowing whether it is or not --
but for sure it is not intended or meant that way...) that I will be able
to write the program until it is finished (read I am no longer able to find
the included bugs... ;) ) and than as a final step I can write the whole
thing into flash.
That is nice! Really puts down the write cyckes...

(The other positive surprise was to read, that it will be possible to plot
to the whole screen. And I am SURE that there will be more...I ordered my DM42
today...)
DM 42 - SN: 00373, Firmware release v.:3.22. / DMCP 3.24. as compiled by SwissMicros
User avatar
Guenter
Posts: 168
Joined: Wed May 24, 2017 6:26 pm
Location: Germany

Re: Flash write cycles

Post by Guenter »

Dave Britten wrote:
Sun Dec 17, 2017 7:56 pm
I noticed that the manual mentions the flash chip being rated for a minimum of 100,000 write/erase cycles, and that the FAT is always stored in a static location. What actions on the calculator cause flash writes? Does this only happen when manually saving calculator state or transferring programs via USB?
Modern devices with flash memory like USB-sticks, SDcards, SSD have load balancing features. That is, although it seems that data, like FAT for instance, are always written to the same place, internal electronics prevent from doing so. I don't know if the architecture of the SoC and the additional flash memory in the DM42 are behaving in a similar way. But even if the FAT is constantly written to the same place, I think we shouldn't bother. Let's assume the DM42 is your major device in your life for the next 30 years, I'd be 100 then, you'd be able to write ~9 times a day someting to the flash memory. If you reach these numbers each day within the next 30 years, you're a real freaky nerd :) and deserve a cost free replacement of the flash memory by SwissMicros :D

Günter
Günter
DM42 SN:00004 and SN:00184 -- DM41X SN:00013 and SN:00955
Dave Britten
Posts: 137
Joined: Wed Jun 14, 2017 9:27 pm

Re: Flash write cycles

Post by Dave Britten »

Guenter wrote:
Tue Dec 19, 2017 9:34 pm
Modern devices with flash memory like USB-sticks, SDcards, SSD have load balancing features. That is, although it seems that data, like FAT for instance, are always written to the same place, internal electronics prevent from doing so. I don't know if the architecture of the SoC and the additional flash memory in the DM42 are behaving in a similar way. But even if the FAT is constantly written to the same place, I think we shouldn't bother. Let's assume the DM42 is your major device in your life for the next 30 years, I'd be 100 then, you'd be able to write ~9 times a day someting to the flash memory. If you reach these numbers each day within the next 30 years, you're a real freaky nerd :) and deserve a cost free replacement of the flash memory by SwissMicros :D

Günter
Right, if there's load balancing, then there's no way anybody is going to wear the thing out without deliberately trying. I have my doubts that this SoC is doing any load balancing, though. My assumption is that it's just strapped directly to the flash chip with no address abstraction. But even in that case, it sounds like durability won't be an issue.
Post Reply