HP Forums

Full Version: Samsung ARM9 (920T) End of Life?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

I recently picked up an HP 49G+ calculator to play with -- learning to program an ARM9 CPU seems interesting to me. I think the HP 49G+ is a nifty machine (even if build quality is, as expected, less than my HP-12C, HP-17BII, HP-42S, and HP-48GX...).

I looked up the ARM chip that I read HP uses (Samsung S3C2410X) and noticed on Samsung's website for the part, under section "Production Status", the initials "EOL" which I believe stands for End-Of-Life.

Since I recently read that HP moved off of the Saturn CPU partly because NEC stopped making the chip (probably too low volume for NEC to justify the legacy fab process), it interests me greatly to see this product note for the HP-49G+ CPU.

http://www.samsung.com/Products/Semiconductor/SystemLSI/MobileSolutions/MobileASSP/MobileComputing/S3C2410X/S3C2410X.htm

The conspiracy theorist in me wonders if this might drive HP to rev it's high end calculator again thus leading to Amazon's fire sale of the current model. ;-)

Anybody have any info on the Samsung part? My apologies in advance if this is old material on this forum.

-Todd

I looked at the web site and it appears that the S3C2410X has been replaced by the S3C2410A. The main differences are bug fixes to the hardware. See http://www.samsung.com/Products/Semiconductor/SystemLSI/MobileSolutions/MobileASSP/MobileComputing/S3C2410/s3c2410x_to_s3c2410a_revisionpoints_041206.pdf.

Thanks for the info John. Nice to know that they've moved from an "X" rev to an "A" rev. I have a little familiarity with high tech manufacturing from an "IT" perspective -- I just nod knowingly and chuckle when engineers and program managers claim they never release an X rev to customers ... just for testing you know. ;-)

Wow, I started my career with Computer Concepts of Gaddesden Place, Hert in July 1988 programming (in assembler) the Acorn Archimedes which was based on a 4MHz (?) ARM3-core. Man the details kinda escape me now, but at the time, there was no Windows, and our PCB layout machine was a top of the line Dell 286 10 MHz machine, which was very very cool. I used to sneak in and play F19 Stealth Fighter on it when no one was looking.

Anyway, my development machine was a regular Archimedes A340 production machine with *no* hard-drive, just a floppy and no software beyond the fairly nasty looking RiscOS, BBC Basic interpreter and inbuilt (into the Basic) ARM assembler. In practice, the only basic we used was a for...next loop to do the 2-pass assembly. Oh, and there wasn't any network either. Just sneakernet.

With this unpromising start, I actually managed to program a large chunk of the UI for the Impression DTP package, and probably the worst printer driver ever. We also worked on a sadly unreleased OS called Impulse which was pretty neat.

The performance of the original ARM chips was actually pretty good. It's a huge testament that a 26,000 transistor chip running at half the rate of the contemporary 286 actually ran similarly to the as yet unheard of 486.

Great that these chips are still doing something useful. I think there may be one in my Clie, but I have my hands full trying to learn my CX to do much Clie stuff.

Thanks John for relating your interesting experience with early ARM! I'm glad my failure to read far enough into Samsung's product web page illicited something good.

I started programming with C, Pascal, and FORTRAN. Nowadays, I mostly work with Java, Oracle PL/SQL, and some UNIX shell scripting.

I've always wanted to learn a little bit of Assembly and do some HelloWorld type stuff since I believe having lower level experience provides valuable insights into what is going on with the higher level abstractions that we mostly deal with these days.

...however, I have no desire to learn x86 Assembler since CW is that it is strange, so perhaps ARM assembler which I have read is clean and "friendly".

-Todd

Yes, Todd, I think I share your sentiments on x86 assembly, though some of the MMX stuff might be fun.

I started as a kid with 6502, then went to 68000 and from that point ARM assembly was pretty natural.

I remember being surprised about the fuss people made about RISC not being able to do additions to memory, but I always tried to stay away from the long 68000 instructions, so I guess I was a load/add/store kinda programmer.

Good luck with your experiences. I believe there's a Archie emulator about somewhere. That probably includes BASIC and the easy to use ARM assembler. That might be worth a try.

I understand the reasons for RISC and load/store architectures. Still, I miss the PDP-11 (http://en.wikipedia.org/wiki/PDP-11) and its versatile addressing modes. The 68000 had some similarities to the PDP-11, but it's effectively dead too.

I'll have to pick up an ARM reference manual and see what clever things it can do.