"Turbo"

[John W Kercheval]

The word "turbo" is used a lot. For example, switching the capacitor on the HP41CV like i did to double the speed is called a "Turbo 41CV"

Using the 41/71 Translator Pac to run 41 pgms on the 71 at 5X the speed is called the "41 Twin Turbo", etc.

There are rumors the 41CL is even faster ht the 71 with the translator pac.

Be advised, this is not really turboing something as a turbo charging system is a mechanical device

Any other speed ups out there? Can you pump up the volume of other units?

[Massimo Gnerucci]

(04-27-2014 04:13 AM)John W Kercheval Wrote:  Be advised, this is not really turboing something as a turbo charging system is a mechanical device

Really? What a pity! A "Wrooom wrooooom" 41 is what I need!

Oh well, I'll have to settle for my 41CL...

[jebem]

I recall the 80's when the IBM ISA PC was set as the standard for personal computing after a smart IBM move to publish the ISA architecture and BIOS in full in a very nice book (no eBooks in those times).
After that, the market was invaded with the so called PC clones coming from the Orient. I built one of these clones myself, by importing a motherboard PCB from HK.
Those PCB's had no components installed, so one would need to acquire and solder in all the required components (mostly TTL "LS" and "F" chips).
At last I used a EPROM programmer to literally clone the original IBM PC BIOS ROM, and voila, a new PC was born.
And yes, I had installed a nice "Turbo" push button to speed up the CPU and RAM!
I recall to have a few crashes when pushing the clock speed too much!
A few years later, nearly all the compatible IBM PC's had that "Turbo" option.

Those were the times before the term "Overclocking" took over the ancient "Turbo" expression.

I digress, I'm sorry if I'm upsetting our MoHPC fellows.

There is a good reason why the manufacturers uses a specific (under)clock speed on their products: Reliability

As I see it, when overclocking a system one needs to consider a few potential issues from using higher clock frequency (meaning more gate's switching operations per second):
1. Higher electrical current consumption depleting the battery's life in shorter time;
2. Higher working temperature as the current is almost converted into heat;
3. Shorter lifetime for the hardware: forcing the semiconductors to operate under high temperatures can cause degradation in gate leakage current or even catastrophic internal failures;
4. Running errors resulting from high temperatures while overclocking: digital circuits are complex in nature and one gate depends on the other, so when the transition times (0 to 1 or vice versa) are approaching the same values as the steady state times, the circuit starts to giving running errors;
5. The digital gates implementing the logic functions are very dependent on the clock speed: again, the mentioned transition times are the limiting factor to do overclocking and get away with it.

I know that the above issues are minimized when using modern LSI chips that runs MOS circuits consuming extremely low currents, allowing for higher clock speeds.
However I'm afraid there is no way to go around point (5).

That said, I'm for overclocking everything just for fun!
It is all about the human nature - to break our own limits!

Cheers

[Thomas Radtke]

(04-27-2014 10:59 AM)jebem Wrote:  Those were the times before the term "Overclocking" took over the ancient "Turbo" expression.

Actually that was a marketing term. The CPUs usually weren't overclocked, but at slow speed there was some degree of compatibility to software using busy loops for timing purpose.

[Alvaro]

(04-27-2014 10:59 AM)jebem Wrote:  I digress, I'm sorry if I'm upsetting our MoHPC fellows.

There is a good reason why the manufacturers uses a specific (under)clock speed on their products: Reliability

There were very big (Hardware) problems with Overcloking PC´s mostly with ISA bus.
Most problematic was the Keyboard and HDD.
So "overcloking" must be done very carefully and people must know exactly what they are doing.
Not only Hardware failures are to expect but also data corruption are likely to happen.

[John W Kercheval]

What I meant was, are there any other mods for calcs other than those I mentioned? For example, a pumped up 67, or anything else?

[Raymond Del Tondo]

(04-27-2014 01:24 PM)John W Kercheval Wrote:  What I meant was, are there any other mods for calcs other than those I mentioned? For example, a pumped up 67, or anything else?

Maybe this one?

;-)

[Sylvain Cote]

(04-27-2014 11:54 AM)Alvaro Wrote:  There were very big (Hardware) problems with Overcloking PC´s mostly with ISA bus.
Most problematic was the Keyboard and HDD.
So "overcloking" must be done very carefully and people must know exactly what they are doing.
Not only Hardware failures are to expect but also data corruption are likely to happen.

Yep and this is why Monte is constantly toggling the 41CL speed (>x1 to x1 and back) to address those kind of issues (mostly IO's).

(04-27-2014 01:24 PM)John W Kercheval Wrote:  What I meant was, are there any other mods for calcs other than those I mentioned? For example, a pumped up 67, or anything else?

There is a RAM and speed upgrade available (here and here) for the HP-42S.

Best, Sylvain

[edit: add a second link]

[Gerson W. Barbosa]

(04-27-2014 01:24 PM)John W Kercheval Wrote:  What I meant was, are there any other mods for calcs other than those I mentioned? For example, a pumped up 67, or anything else?

There are these excellent ones, Takayuki Hosoda, for the HP-15C and the HP-42S:

http://www.finetune.co.jp/~lyuka/interests/calc/hp15c/

http://www.finetune.co.jp/~lyuka/interests/calc/hp42s/

[jebem]

1) Overclocking the 49G+
Some HP models, for instance the 49G+ model, can have the ARM cpu "accelerated" up to 203MHz (in several steps) and back down to 12MHz, just using software tools created by Alistair Borowski (Clockspeed Adjustment Tools - HP-CAT).
I know this is old news, but here it goes the link to the tool:
http://www.hpcalc.org/details.php?id=6081

2) Overclocking the 48GII
Apparently at least one of the 48GII different production batches also works with Alistair Borowski tools, despite not all the cpu speeds will work.
Well, this is a 49 series machine hardware, despite the name, but there are two production batches: the latest production batch uses 4 cell batteries (closer to a 50G?), while the initial batches uses just 3 batteries and a not so good keyboard.

3) Overclocking the 50G
I read some reports claiming the Alistair Borowski tools will not work on the 50G at 203MHz. Well, it should, if we assume that a 49G+ hardware is essentially the same one used in the 50G. Perhaps there are differences in batch productions (for example, if the 50G memory access is running without wait cycles, while the 49G+ can be using wait cycles, who knows!).
HP (or should I say Kinpo?) use to do silent changes, while keeping the same model number.
On top of that, there are different firmware versions for these HP calculators, and it may affect the results as well.
I couldn't find any documented work detailing all the different conditions.

4) Or forget about turbo/overclocking... Try an HP calc emulator/simulator on Windows or Linux or even on your preferred Android/iPad mobiles instead... now that is real "turbo" mode!
Some examples:
http://www.hpmuseum.org/cgi-sys/cgiwrap/...i?read=867
http://www.hpcalc.org/hp48/pc/emulators/
http://sim41.webcindario.com/sim67.htm
http://www.limpidfox.com/hp67.htm
http://www.cneufeld.ca/neufeld/hp67/hp67.html
https://play.google.com/store/apps/detai....limpidfox

[John W Kercheval]

Very cool! thanks.

[John W Kercheval]

is that 67CX thing a joke? or a real machine?

[Sylvain Cote]

(04-27-2014 04:41 PM)John W Kercheval Wrote:  is that 67CX thing a joke? or a real machine?

Follow any of the web-link and you will have your answer! ;-) [or click here]

[Garth Wilson]

Quote:switching the capacitor on the HP41CV like i did to double the speed is called a "Turbo 41CV"

Using the 41/71 Translator Pac to run 41 pgms on the 71 at 5X the speed is called the "41 Twin Turbo", etc.

There are rumors the 41CL is even faster ht the 71 with the translator pac.

Besides the hardware, efficiency in software (something I know is old-fashioned) is important, and Ángel and others here have helped tremendously in that way with M-code solutions that are far faster than what we had years ago. The 71's Forth mdule was very poorly done, and I was able to dramatically speed up many of the functions by re-defining them, in Forth, not even in assembly, the most extreme example being CHR$ at 14 times as fast as the original.

[John W Kercheval]

Garth: Very interesting concept! thanks for sharing that. never thought of it.

[Dave Frederickson]

The Portable PLUS can have it's clock increased from 5.33MHz to 8MHz.

[ttw]

We used to claim that shredders had four speeds:
1. Slow
2. Fast
3. Turbo
4. Subpoena

[Katie Wasserman]

(06-20-2014 03:53 AM)ttw Wrote:  We used to claim that shredders had four speeds:
1. Slow
2. Fast
3. Turbo
4. Subpoena

LOL!

[Oulan]

You can even have a turbo HP97+ at https://sites.google.com/site/olivier2sm...jects/hp97

Turbo speed HP97 with full memory address used : 106 regs, 896 step and 5 more instructions to fill the 256 byte code ops map.

Of course with turbo cards : one card is ok for all regs or program

Code:


1.7
new HP97 model : HP97+ fully loaded HP97

896 program steps

106 registers (A to I are regs 100 to 105)

HD magnetic cards : one side is enough for either:
- whole program memory 
- whole register set
Be careful not to mix your cards sets
(a tool to convert cards set will be done soon)

5 new instructions: one more indirect register E
RCL (e) keyed as 'RCL' 'e^x', printed as 'RCLe'
STO (e) keyed as 'STO' 'e^x', printed as 'STOe'
GTO (e) keyed as 'GTO' 'e^x', printed as 'GT->e'
GSB (e) keyed as 'GSB' 'e^x', printed as 'GS->e'
use the E register as a new indirect one, same functionnalities as I register
BACK x keyed as 'LBL' 'e^x', printed as 'Bc^x'
go back 'stack X' steps (stack X doesn't need to be negative)

97.src.lst is a commented source of the HP97 rom's
can be assembled with 'ass97.py' to produce 
- original roms (define HP97+ as 0)
- patched roms (define HP97+ as 1)

new display font (more realistic I hope)

[Ángel Martin]

(04-27-2014 04:13 AM)John W Kercheval Wrote:  There are rumors the 41CL is even faster ht the 71 with the translator pac.

Can't comment on the speed comparison between those two (although I very much also think so), but even if the translator pac on the 71 is faster - it only works on a marginal sub-set of the 41 world, so it's a bit of a moot point I think.