How Overclock 50g Hp calculator? - Hewlett Packard

This is a discussion on How Overclock 50g Hp calculator? - Hewlett Packard ; Hi to all, im developing a software in sysrpl for structural analisys. But the operation with matrix is too slow in certain commands. I remember a software located at hpcalc.org that overclock the 49g+ calculator, and i tried in my ...

+ Reply to Thread
Results 1 to 11 of 11

Thread: How Overclock 50g Hp calculator?

  1. How Overclock 50g Hp calculator?

    Hi to all, im developing a software in sysrpl for structural analisys.

    But the operation with matrix is too slow in certain commands.

    I remember a software located at hpcalc.org that overclock the 49g+
    calculator, and i tried in my "old" 49g+ calculator with possitive
    results.

    But when i used in my "new" 50g, this software can`t run.

    I heard about the FEM49 software that have a "Fast" routine that use a
    overclock system to perform the analisys more quickly.

    Someone of yours can tellme how to develope a routine for that??

    Is possible to make a overclock in the 50g like the 49g+??


    Thanx to all, and sorry my poor english :P

  2. Re: How Overclock 50g Hp calculator?

    Hello,

    > I remember a software located at hpcalc.org that overclock the 49g+
    > calculator, and i tried in my "old" 49g+ calculator with possitive
    > results.
    >
    > But when i used in my "new" 50g, this software can`t run.

    AFAIK there is no *difference* in this hardware section of the 49G+
    and the 50G. I am most curious about this and would be very interested
    in trying this out myself on a 49G+ and a 50G. Maybe you could make
    the code/program availabe for that, so that others can trial and error
    this.

    For overclocking take a look at this posts of 'FEM48/49 v5.3 updated
    to SR4':
    http://groups.google.com/group/comp....723eb70baa8402
    This code is based on Al Borowski´s overclocking code.

    HTH,
    Andreas
    http://www.software49g.gmxhome.de

  3. Re: How Overclock 50g Hp calculator?

    On Aug 25, 6:29*am, Andreas Möller
    wrote:
    > Hello,
    >
    > > I remember a software located at hpcalc.org that overclock the 49g+
    > > calculator, and i tried in my "old" 49g+ calculator with possitive
    > > results.

    >
    > > But when i used in my "new" 50g, this software can`t run.

    >
    > AFAIK there is no *difference* in this hardware section of the 49G+
    > and the 50G. I am most curious about this and would be very interested
    > in trying this out myself on a 49G+ and a 50G. Maybe you could make
    > the code/program availabe for that, so that others can trial and error
    > this.
    >
    > For overclocking take a look at this posts of 'FEM48/49 v5.3 updated
    > to SR4':http://groups.google.com/group/comp..../thread/2a723e...
    > This code is based on Al Borowski´s overclocking code.
    >
    > HTH,
    > Andreashttp://www.software49g.gmxhome.de


    Thanx a lot Andreas!

    I use part of your code for error checking, but for clock i use de Al
    Borowski apps with successfull results.
    I drop down the time to run my app in a 40%, i confuse in something...
    the processor take more time to run my app using a 203 mhz setting,
    that using a 119 mhz.

    for this, i use de 119mhz for performance, and i implemented your code
    for error handling.
    Thanx and sorry my english, see u in adictoshp.org, bye!

  4. Re: How Overclock 50g Hp calculator?

    Hello,

    > I use part of your code for error checking, but for clock i use de Al
    > Borowski apps with successfull results.

    The code inside my post 'FEM48/49 v5.3 updated to SR4' is the one from
    Al Borowski in *binary* format which sets the calculator to 203 MHz
    and the one at the end (in binary format as well) sets it back to 75
    MHz.

    > I drop down the time to run my app in a 40%, i confuse in something...
    > the processor take more time to run my app using a 203 mhz setting,
    > that using a 119 mhz.
    > for this, i use de 119mhz for performance, and i implemented your code
    > for error handling.

    This does not make sense. I suspect that something different is
    happening here which is impossible to speculate about without seeing
    the code. The same code (in RAM of the calc) run at 203MHz can not be
    slower than running at 119 MHz ;-)

    Best wishes,
    Andreas
    http://www.software49g.gmxhome.de

  5. Re: How Overclock 50g Hp calculator?

    On Aug 24, 10:12*pm, "asrh...@gmail.com" wrote:
    > Hi to all, im developing a software in sysrpl for structural analisys.
    >
    > But the operation with matrix is too slow in certain commands.


    Have you considered C? The following is a link to a C-based Sparse
    Linear Solver that can be called from UserRPL:

    http://sense.net/~egan/hpgcc/#Exampl...inear%20Solver

    Scroll down to the "Dense vs. Sparse" section for a comparison of the
    50g's internal matrix solver vs. the one in C.

    It's just an example and may or may not apply to you. But, the
    argument is the same, you could write the time critical parts of your
    code in C and save a lot of time. E.g. the setup and solve of a 99x99
    matrix took the internal solver 90 seconds, whereas the same problem
    in C was setup and solved in 7 seconds.

  6. Re: How Overclock 50g Hp calculator?

    On Aug 25, 10:51*am, Andreas Möller
    wrote:
    > Hello,
    >
    > > I use part of your code for error checking, but for clock i use de Al
    > > Borowski apps with successfull results.

    >
    > The code inside my post 'FEM48/49 v5.3 updated to SR4' is the one from
    > Al Borowski in *binary* format which sets the calculator to 203 MHz
    > and the one at the end (in binary format as well) sets it back to 75
    > MHz.
    >
    > > I drop down the time to run my app in a 40%, i confuse in something...
    > > the processor take more time to run my app using a 203 mhz setting,
    > > that using a 119 mhz.
    > > for this, i use de 119mhz for performance, and i implemented your code
    > > for error handling.

    >
    > This does not make sense. I suspect that something different is
    > happening here which is impossible to speculate about without seeing
    > the code. The same code (in RAM of the calc) run at 203MHz can not be
    > slower than running at 119 MHz ;-)
    >


    It does make sense if you consider that:

    * Al Borowski overclocking code was part of early experiments with the
    ARM processor. The routine may not use the optimum values for the rest
    of the hardware to work including screen, keyboard, memory, etc. (it
    was reported that keyboard cannot be read at higher speeds, for
    example).
    * The increased processor speed doesn't mean you can increase the RAM
    access speed. Memory access needs to run at the same speed (for which
    proper wait cycles must be added, which has to be configured in the
    overclocking routine).

    I've made my own experiments with overclokcing for hpgcc 3.0 and once
    you get everything properly setup to run at higher speeds, you'll see
    a speed increase that varies according to your program.
    For programs that are CPU intensive but don't use much memory, the
    speed increase is almost linear.
    Programs that are memory-intensive show little or no increase at all,
    even when running at 192 MHz vs. 75 MHz. This is simply because the
    memory is accessed at the same clock speed.
    The case mentioned by the OP, where the program runs slower, is
    probably because the memory speed is "underclocked", perhaps as a bug
    or perhaps because there are limited clock division/memory wait states
    possible configurations, and memory must use a slower speed than
    normal.

    Regards,
    Claudio

  7. Re: How Overclock 50g Hp calculator?

    On Aug 25, 8:21*am, "dataj...@gmail.com" wrote:
    > On Aug 24, 10:12*pm, "asrh...@gmail.com" wrote:
    >
    > > Hi to all, im developing a software in sysrpl for structural analisys.

    >
    > > But the operation with matrix is too slow in certain commands.

    >
    > Have you considered C? *The following is a link to a C-based Sparse
    > Linear Solver that can be called from UserRPL:
    >
    > http://sense.net/~egan/hpgcc/#Exampl...inear%20Solver
    >
    > Scroll down to the "Dense vs. Sparse" section for a comparison of the
    > 50g's internal matrix solver vs. the one in C.
    >
    > It's just an example and may or may not apply to you. *But, the
    > argument is the same, you could write the time critical parts of your
    > code in C and save a lot of time. *E.g. the setup and solve of a 99x99
    > matrix took the internal solver 90 seconds, whereas the same problem
    > in C was setup and solved in 7 seconds.


    I read all of Egan Ford's article and all I can say is WOW!

    There is a lot of amazing stuff there. I did not realize all that the
    50g could do.

    If only HPGCC could be ported to the 50g

  8. Re: How Overclock 50g Hp calculator?


    > > Have you considered C? *The following is a link to a C-based Sparse
    > > Linear Solver that can be called from UserRPL:


    i thinking about programming in C, but i see that the result file size
    is very big. But, you have the reason...this can be the solution where
    the sysrpl is slow (matrix operations, etc).

    Respect to the message writed by Claudio,...yeah you have the
    reason too. This could be because some commands are CPU or Memory
    based (principaly).

    Is for that, that is necessary to experiments the best way to mannage
    the system clock..are possible a tool to clock only the memory??

    if the answer is yes, this be a great, because you can adapt the clock
    of your cpu/memory depending on the aplication or command.

    Claudio, i think that the FAST routine included in your structural
    analisys software is great, you can re-considere the option to turn
    off the LCD; not for speed gain, only for battery save, remember that
    the clock increment the battery usage, and if you can disable the LCD
    for this time, you are saving energy.

    Consider this ( please sorry the english, i aprove the english 1
    this semester!! XD )

    Sometimes, i think when are the final of the sys and the user rpl,
    because i think that the calculator progress is stopped by HP..is for
    that, that born HPGCC programming language ported to hp, for solve the
    slow emulation of older saturn.

    HP must be write a new ROM, for use the new Samsung prosessor fully :O

    but for now..i think that the relation time/size/complexity is
    balanced to sysrpl.




  9. Re: How Overclock 50g Hp calculator?

    On Aug 25, 8:08*pm, "asrh...@gmail.com" wrote:
    > > > Have you considered C? *The following is a link to a C-based Sparse
    > > > Linear Solver that can be called from UserRPL:

    >
    > i thinking about programming in C, but i see that the result file size
    > is very big.


    True, but it is not an issue. All my examples are stored on the SD
    card and run directly from the card. The most UserRPL can access is
    just under 256KB, where as C can access ~ 400KB. So, the overhead to
    load the C program from SD to RAM is not a factor. Most of my C
    programs have ~320KB of RAM to use once the program starts.

    More info here:

    http://sense.net/~egan/hpgcc/#C%20Bi...g,%20Too%20Big

    The CG library created here takes very little space because the C
    binaries are on SD:

    http://sense.net/~egan/hpgcc/#Comput...hics%20Library

    CG.LIB is only 817 bytes, whereas CG.ZIP is 151KB. CG.ZIP lives on
    the SD (MB/GB of storage vs KB).

  10. Re: How Overclock 50g Hp calculator?

    On Aug 25, 10:08*pm, "asrh...@gmail.com" wrote:

    > Claudio, i think that the FAST routine included in your structural
    > analisys software is great, you can re-considere the option to turn
    > off the LCD; not for speed gain, only for battery save, remember that
    > the clock increment the battery usage, and if you can disable the LCD
    > for this time, you are saving energy.


    I think you are confusing me with somebody else. I never wrote any
    structural analysis software, despite being a structural engineer.
    Credit goes for whoever wrote the software you mentioned.


    Regards,
    Claudio

  11. Re: How Overclock 50g Hp calculator?

    On Mon, 25 Aug 2008 21:08:26 -0500:

    > re-consider the option to turn off the LCD;
    > not for speed gain, only for battery save...


    Hopefully the amount of energy for driving the LCD
    may be so minute as to be a negligible influence on battery life,
    particularly if the only time "off" would be during
    modest intervals during program running,
    and while nothing was changing on the display.

    It could actually be disconcerting to see one's display
    turn off unexpectedly, so the chance of medical consequences
    (and their huge cost) should also be factored into the equation

    -[ ]-

+ Reply to Thread