Threaded Mode | Linear Mode

Hlib · 09-29-2019, 08:47 PM

In recent years, it has been repeatedly noted that new models of graphing calculators have
too much power consumption at a relatively slow computational speed unlike their classic predecessors from the 1990s.
https://comp.sys.hp48.narkive.com/Zq0obS...onsumption
https://www.hpmuseum.org/forum/thread-4856.html )
I purchased the fx-9860gii-2 (OS 2.09) to get to know a typical modern member of the CASIO family. It is equipped with the custom SH-4AL(SH7305) without FPU made for CASIO. The device is supplied with two cables: SB-62 and USB mini type A.

http://images.vfl.ru/ii/1569786378/a9c84...018037.jpg
http://images.vfl.ru/ii/1569786412/1e9a7...018053.jpg
The Internet noticed the weaknesses of this model: in general it should be at least similar in speed (compared to older fx-9860g with SH-3), but it is not (due to a lot more bloated and badly optimized software). Also in the specification and on the back of the case stated: 0.7 W. All this I decided to check.
As for programming, interface, debugging, editing and mathematical capabilities - it brings together the best of all previous models. Some new features have been added: processing string variables, creating folders with files of different types, spreadsheet mode etc.
The unexpected surprise for me was the power consumption of this calculator (the measurements were carried out at a voltage of 5 V in the battery compartment and averaged over a time interval).
sleep mode (OFF) - I=30 µA
idle (cursor blinking) - I=1.5 mA (P~0.008W)
running program or graphing - I=19 mA (P~0.1W)
data send to/receive from flash 1.5 Mb memory - I=20 mA (P~0.1W)
scrolling text - 4 mA (P~0.02W)
scrolling list/matrix data - 8mA (P~0.04W)
The shut off point is 3.95 V.
Switching the backlight mode changes the perceived color of pixels from black to blue and adds to all measurements 7mA (P~0.04W). Summary: Pmax=0.15W.
The keyboard is made in the style of cheap financial calculators, but it does not cause complaints. These are first impressions of fx-9860gii-2.

http://images.vfl.ru/ii/1569786300/19649...018028.jpg
http://images.vfl.ru/ii/1569786349/90148...018034.jpg

toml_12953 · 09-30-2019, 05:14 AM

Thanks for the review! I have the fx-9860GII-2 SD (includes an SD card slot) and am very happy with it.

Thott · 09-30-2019, 09:27 AM

I don’t know the calculator at all.
But the display seems to have a quite good contrast.

Dave Britten · 09-30-2019, 11:31 AM

I love the fx-9860G series. I just got a 9860GII SD last week, after finally finding a decent deal on one. I think the Slim is my favorite version, since it has the built-in help feature, is super tiny, and has just about the best rubber dome keyboard I've ever used on a calculator (the TI-85 gives it a close race).

Hlib · 10-10-2019, 09:26 PM

13 days this calculator was subjected to heavy testing. there are two minor bugs that few people will notice

toml_12953 · 10-10-2019, 10:20 PM

(10-10-2019 09:26 PM)Hlib Wrote: 13 days this calculator was subjected to heavy testing. there are two minor bugs that few people will notice

and those bugs are...?

toml_12953 · 10-10-2019, 10:25 PM

(09-29-2019 08:47 PM)Hlib Wrote: In recent years, it has been repeatedly noted that new models of graphing calculators have
too much power consumption at a relatively slow computational speed unlike their classic predecessors from the 1990s.
https://comp.sys.hp48.narkive.com/Zq0obS...onsumption
https://www.hpmuseum.org/forum/thread-4856.html )
I purchased the fx-9860gii-2 (OS 2.09) to get to know a typical modern member of the CASIO family. It is equipped with the custom SH-4AL(SH7305) without FPU made for CASIO. The device is supplied with two cables: SB-62 and USB mini type A.

I just bought an fx-CG50. It's Basic is very compatible with the 9860 and it also runs Micro Python. While the Basic is pretty fast, Python blows it away! The full color screen is easier on the eyes (IMHO) than the 9860. I still like my 9860 but the CG50 is my new toy. I keep discovering new things about it.

Hlib · 10-12-2019, 05:19 PM

(10-10-2019 10:20 PM)toml_12953 Wrote: and those bugs are...?

1) Under certain conditions, copying files 1...6 of lists from flash, the calculator sometimes ignores the command " Overwrite OK? No [F6]".
2) The program size is displayed only every 4 bytes : 4,8,12, ... 224, 228... etc
There are probably other bugs. But I`ve been using graphing calculators as interesting toys nowadays, I don`t go deep into them.

Hlib · (This post was last modified: 10-27-2019 05:52 PM by Hlib.)

Another advantage of this model. The user manual for fx-9860gii states that this calculator communicates seamlessly with models from the past: fx-7400, 7450, 9850, 9950, graph 25, 25+, 35 etc. I tested this on the graph25 and cfx-9850gb+ models. You can create programs (data) on the 9860-gii and load them into old calculators using a standard 3-pin SB-62 cable. 9860gii has in the menu or in the catalog the entire list of commands for older models. Conversely, all programs from graph25 (for example) run in fx-9860gii.

http://images.vfl.ru/ii/1572197294/09cd1...348285.jpg

Hlib · (This post was last modified: 11-17-2019 03:40 PM by Hlib.)

Almost three years have passed since the first release of the CBASIC application for the CASIO 9860G series. 11/2019 the latest v2.36 beta version has been released and some consider CBASIC app the best program ever made for CASIO calculators.
The new interpreter speeds up user programs within 15...45 times, still making the 9860gii-2 the most economical calculator in its class in terms of autonomous power consumption. In addition to speeding up programs, there is access to hidden resources of the calculator: system date and time, intervals in ticks, 250K RAM (is now available), battery voltage and backlight control. More than 200 new commands and functions significantly expand the capabilities of the built-in BASIC language, but it takes a lot of time to learn them.
8-queens benchmark test at 29MHz CPU speed: t=0.98s (integer mode), t=1.62s (double precision mode). Using the algorithm for old models, e.g. for fx-6300g, we get a faster result: t=0.83s and t=1.27s respectively.

http://images.vfl.ru/ii/1573996866/7cc1d...590909.jpg
http://images.vfl.ru/ii/1573996791/7ad87...590896.jpg

xerxes · 11-19-2019, 10:27 PM

I've tested C.Basic 1.86 in combination with ftune on the FX-9860G SD. If I remember correctly, I've also tried the FX-6300G version, but the fastest was this one:

Code:

 0->A~Z

 8->R

 R->Dim List 1

 Do

   Isz X

   R->List 1[X]

   Do

     Isz S

     X->Y

     While Y>1

       Dsz Y

       List 1[X]-List 1[Y]->T

       If T=0 Or X-Y=Abs T Then

         0->Y

         Dsz List 1[X]

         Deg

         While List 1[X]=0

           Dsz X

           Dsz List 1[X]

         WhileEnd

       IfEnd

     WhileEnd

   LpWhile Y<>1

 LpWhile X<>R

 S

The results:

Code:

- 1.06      FX-9860G SD    C.Basic 1.86 / Integer Mode

- 0.563     FX-9860G SD    C.Basic 1.86 / Integer Mode / Fast Mode x1.9 (29.5->118 MHz)

- 0.385     FX-9860G SD    C.Basic 1.86 / Integer Mode / Fast Mode x2.8 (29.5->118 MHz / -1 wait state)

I'll make some tests with version 2.37 on the FX-9860G SD this weekend, also to see the influence of the local vars.

Mjim · 11-20-2019, 02:10 AM

First of all thank you Hlib. I had not heard of C.Basic, but am impressed at it's speed (the sequence function can build a table of 10,000 values of sine between 0 & Pi/2 in 2 seconds for my calculator....that is just madness that wouldn't of been possible without the performance and extra unlocked RAM of C.Basic). Appreciate the power measurements that you did. Would be interesting to see what kind of power usage you would get in C.Basic, considering how nicely optimized it seems to be.

In a way, I prefer the nice contrast of black LCD's on the fx-9860GII/9750GII over the colour display on the fx-CG50 as I'm not particularly keen on backlit displays for calculators (also I've heard the CG50 has SDRAM which loses memory after being powered off for about an hour vs SRAM (fx-9750GII/9860GII/CG10/CG20) which retains it until battery replacement). I would love to see a Casio graphing calculator with a high resolution black LCD (perhaps sporting a pixel density similar or better than the 991EX); or if it is to be colour, a transflective LCD display, so it can be operated without the backlight.

I just discovered that "Execute Time" flag, was trying to figure out how you got that sort of execute accuracy, so I'll post my results on my fx-9750GII (SH3 - 9860GII firmware).

Stock ftune (Nrm) settings (29MHz):

C.Basic 2.36:
DBL# NQUEENS: 1.84 sec
INT% NQUEENS: 1.11 sec

Built in Basic:
NQUEENS: 33 seconds

Max ftune (F5) settings (117.96MHz):

C.Basic 2.36:
DBL# NQUEENS: 0.69 sec
INT% NQUEENS: 0.41 sec

Built in Basic:
NQUEENS: 14.5 sec

klesl · 11-20-2019, 09:17 AM

...also I've heard the CG50 has SDRAM which loses memory after being powered off for about an hour...
I don't observe this behaviour on my fx-CG50. I turn the calc on after 48 hours and the history is still visible at least in run- matrix, statistics, spreadsheet, graph, 3D graph.

klesl · (This post was last modified: 11-20-2019 09:20 AM by klesl.)

deleted double post

Dave Britten · 11-20-2019, 12:27 PM

(11-20-2019 09:17 AM)klesl Wrote: ...also I've heard the CG50 has SDRAM which loses memory after being powered off for about an hour...
I don't observe this behaviour on my fx-CG50. I turn the calc on after 48 hours and the history is still visible at least in run- matrix, statistics, spreadsheet, graph, 3D graph.

It saves the contents of memory to flash after a couple of hours then fully powers down the memory. When you turn it on again, it reloads main memory from flash. The earlier models (fx-9860, etc.) also save to flash some time after being turned off, but they don't power down the main memory. It's intended more as a backup in case the batteries die/are removed.

Mjim · 11-20-2019, 12:57 PM

(11-20-2019 09:17 AM)klesl Wrote: ...also I've heard the CG50 has SDRAM which loses memory after being powered off for about an hour...
I don't observe this behaviour on my fx-CG50. I turn the calc on after 48 hours and the history is still visible at least in run- matrix, statistics, spreadsheet, graph, 3D graph.

The original link that first pointed out this difference was here:
https://community.casiocalc.org/topic/78...entry62236

Perhaps everything in "Main Memory" that has a filename is backed up to flash, while anything that is actually running in RAM (eg a complicated equation or program) is lost?

If it is backing up all the files in "Main Memory" (including the run-matrix history) as you pointed out, then that is better than what I was imagining when I first saw the comment about memory being lost. While I may not be particularly fond of the back-lit display (although the higher resolution is appreciated), the Xcas port, Python, better processing power and nicer looking exterior were strong reasons for me to look at purchasing the fx-CG50 calculator, but the uncertainty about the memory being retained spoiled all of those good points.

Dave Britten · 11-20-2019, 01:06 PM

(11-20-2019 12:57 PM)Mjim Wrote:
(11-20-2019 09:17 AM)klesl Wrote: ...also I've heard the CG50 has SDRAM which loses memory after being powered off for about an hour...
I don't observe this behaviour on my fx-CG50. I turn the calc on after 48 hours and the history is still visible at least in run- matrix, statistics, spreadsheet, graph, 3D graph.
The original link that first pointed out this difference was here:
https://community.casiocalc.org/topic/78...entry62236

Perhaps everything in "Main Memory" that has a filename is backed up to flash, while anything that is actually running in RAM (eg a complicated equation or program) is lost?

If it is backing up all the files in "Main Memory" (including the run-matrix history) as you pointed out, then that is better than what I was imagining when I first saw the comment about memory being lost. While I may not be particularly fond of the back-lit display (although the higher resolution is appreciated), the Xcas port, Python, better processing power and nicer looking exterior were strong reasons for me to look at purchasing the fx-CG50 calculator, but the uncertainty about the memory being retained spoiled all of those good points.

From an end-user perspective, it's almost exactly like the fx-9860g, i.e. nothing is lost by leaving the calculator powered off for hours/days. You'll simply notice a delay of a second or two next time you turn it on, while main memory is being restored. It's like the hibernation feature in Windows, and the TI-84 CE appears to do something very similar (it even shuts down the realtime clock, meaning you'll have to reset the time and date if the 84 hasn't been used for a few days).

Mjim · 11-20-2019, 01:43 PM

(11-20-2019 01:06 PM)Dave Britten Wrote: From an end-user perspective, it's almost exactly like the fx-9860g, i.e. nothing is lost by leaving the calculator powered off for hours/days. You'll simply notice a delay of a second or two next time you turn it on, while main memory is being restored. It's like the hibernation feature in Windows, and the TI-84 CE appears to do something very similar (it even shuts down the realtime clock, meaning you'll have to reset the time and date if the 84 hasn't been used for a few days).

This sounds fantastic, but this comment by piu58 seems to suggest that not all of the RAM is backed up to the flash:

"If you run a program which need some time for calculation, and you don'*t wait for the result, it may be your results are lost. The calculator switches off automatically if not used for a period of time. The cg20 stores information as long as battery life."

I figured with a hibernation feature, RAM, CPU cache and the current memory pointers would all be copied from flash so that nothing at all would be lost, including the current progress through a long calculation that is in RAM.

Hlib · 11-20-2019, 10:21 PM

(11-19-2019 10:27 PM)xerxes Wrote: I've tested C.Basic 1.86 in combination with ftune on the FX-9860G SD. If I remember correctly, I've also tried the FX-6300G version, but the fastest was this one:

Code:

0->A~Z 8->R R->Dim List 1 Do Isz X R->List 1[X] Do Isz S X->Y While Y>1 Dsz Y List 1[X]-List 1[Y]->T If T=0 Or X-Y=Abs T Then 0->Y Dsz List 1[X] Deg While List 1[X]=0 Dsz X Dsz List 1[X] WhileEnd IfEnd WhileEnd LpWhile Y<>1 LpWhile X<>R S

In 9860gii-2 this code gives the slowest result: t=1.03s(in integer mode).
The fastest result is provided by the following code (algorithm from fx-6300g and the like):

Code:

0->A~Z:8->R:{1,R}->Dim_Mat_A

Lbl_0:X=R=>Goto_4

Isz_X:R->A[1,X]

Lbl_1:Isz_S:X->Y

Lbl_2:Dsz_Y:Deg:Y=0=>Goto_0

A[1,X]-A[1,X]->T:T=0=>Goto_3

X-Y<>Abs_T=>Goto_2

Lbl_3:A[1,X]-1->A[1,X]

A[1,X]=>Goto_1

Dsz_X:Goto_3

Lbl_4:"S=":S

t=0.73s (in integer mode).
Deg: command performs here No Operation; instead of Mat_A[1,X] in C.BASIC we can write A[1,X];
A[1,X]=>Goto_1 in CASIO syntax means if_A[1,X]<>0_then_Goto_1.

(11-20-2019 02:10 AM)Mjim Wrote: ....Would be interesting to see what kind of power usage you would get in C.Basic, considering how nicely optimized it seems to be....

Battery drain in C. BASIC is almost the same: in program execution mode it is 22 mA at 5.5V (0.12W). Also, this app can run without affecting the main 60K RAM memory of the calculator.

klesl · 11-21-2019, 08:13 AM

(11-20-2019 01:43 PM)Mjim Wrote:
(11-20-2019 01:06 PM)Dave Britten Wrote: From an end-user perspective, it's almost exactly like the fx-9860g, i.e. nothing is lost by leaving the calculator powered off for hours/days. You'll simply notice a delay of a second or two next time you turn it on, while main memory is being restored. It's like the hibernation feature in Windows, and the TI-84 CE appears to do something very similar (it even shuts down the realtime clock, meaning you'll have to reset the time and date if the 84 hasn't been used for a few days).

This sounds fantastic, but this comment by piu58 seems to suggest that not all of the RAM is backed up to the flash:

"If you run a program which need some time for calculation, and you don'*t wait for the result, it may be your results are lost. The calculator switches off automatically if not used for a period of time. The cg20 stores information as long as battery life."

I figured with a hibernation feature, RAM, CPU cache and the current memory pointers would all be copied from flash so that nothing at all would be lost, including the current progress through a long calculation that is in RAM.

The running of the program is different situation, Yo can expected such behaviour since the values assigned to variables are volatile.
It's the same situation similir to the Windows, when I turn the PC off (or sleep or hibernate) during some activity (e.g.video conversion), the activity or result is lost.