• Please review our updated Terms and Rules here

Have Tandy-100? Please run it for me!

V

vol.litwr

Experienced Member
Joined
Mar 22, 2016
Messages
324
I have just made a post in CP/M thread. However this thread seems more appropriate for my request.
Thank you
 
I still need help. Would anybody like to run my program on a real Tandy model 100 (or 102)? There are two files:
1) PINT.BA - Basic code, it executes ML code;
2) CAPINT.CO - ML code, it is executed by Basic code.
I need results for 100, 1000, and 3000 digits. The calculations take about 3s, 5m, and 45m respectively. I need numbers which will be printed, they are just timings in seconds.
Thank you very much.
View attachment tandy100-upd.zip
 
MikeS said:
Impressive!

Looks like another member has already run the tests for him; nice.
Click to expand...

No, it has been another test. The Tandy-100 has very slow screen output, so to accurately calculate ER requires more checking. Please run it for me. It can help us to find out the exact speed ratio for the Z80 and 8085 on this task.
 
I removed an excess file from the archive. So now, the archive contains only two files: PINT.BA and CAPINT.CO - they are the same as in the previous archive. I hope this tiny change will make runs of PINT.BA a bit easier.
 

Attachments

  • tandy100-time.zip
    1.5 KB · Views: 1
just to make this more interesting,
I now have an M100 with a Z80 installed, same clock rate.
 
I guess you missed my point.. I can do a direct comparison of z80 vs 8085 in the same hardware. Seems like a good fir for your stated goal.
 
At first, a man graciously helped me with the required tests - https://atariage.com/forums/topic/313868-i-need-data-from-the-real-tandy-100-or-102/ - I have been a bit surprized, I thought the Tandy-100 is much slower when it prints characters.
Twospruces said:
I guess you missed my point.. I can do a direct comparison of z80 vs 8085 in the same hardware. Seems like a good fir for your stated goal.
Click to expand...
IMHO we know that pi-spigot code written specially for the Z80 is about 15% faster than this code specially written for the 8085. However you can ask me to make the Z80 code for you.
IMHO it is more interisting to run the pure 8080 code on the 8085 and Z80. IMHO the 8085 will be a bit faster. However it seems that you don't have your 8085 anymore. :( So it will be difficult for me to gather results. :(
BTW how do you transfer data to your Tandy? COM-port?
 
I can drop in 4 processors into the hardware
1) stock 8085 at 2.45MHz
2) stock 8085 @ 4.9MHz
3) NSC800 (Z80)@ 2.45MHz
4) NSC800 (Z80)@ 4.9 MHz

all will run 8080 code. you tell me what you want run
 
Twospruces said:
I can drop in 4 processors into the hardware
1) stock 8085 at 2.45MHz
2) stock 8085 @ 4.9MHz
3) NSC800 (Z80)@ 2.45MHz
4) NSC800 (Z80)@ 4.9 MHz

all will run 8080 code. you tell me what you want run
Click to expand...

Your collection is really good!
It is interesting how much unfriendlier 4.9MHz clock is to batteries than 2.45 MHz? It is also interesting what is better for batteries the 8085 or Z80? Have you checked how long your systems can work on batteries?
I have just made pi-spigot implementation for the pure 8080. So we can know what is faster with the 8080 code, the z80 or 8085. I bet it will be the 8085. However the speed difference may be very little. Let's, for instance, check timings for 100 and 1000 digits.
BTW Merry Xmas and Happy New Year to all.
View attachment pi-8080.zip
 
5MHz bumps the current by 25% or so. More than reasonable in my opinion.
NSC800 and 80C85 are equivalent on current to a first order.

Ok I will try to run these.
 
Twospruces said:
5MHz bumps the current by 25% or so. More than reasonable in my opinion.
NSC800 and 80C85 are equivalent on current to a first order.
Ok I will try to run these.
Click to expand...
Thanks for your information about the current value. I have just checked datasheets for the Z80 and 8085 timings and got the following results.

The 8085 is faster on the next instructions:
Code: 
XTHL [16] - EX (SP),HL [19]        +3
DAD B [10] - ADD HL,BC [11]        +1
DAD D [10] - ADD HL,DE [11]        +1
DAD H [10] - ADD HL,HL [11]        +1
DAD SP [10] - ADD HL,SP [11]       +1
INR M [10] - INC (HL) [11]         +1
DCR M [10] - DEC (HL) [11]         +1
IN byte [10] - IN A,(byte) [11]    +1
OUT byte [10] - OUT (byte),A [11]  +1
JNZ far [7/10] - JP NZ,far [10]   +3/0
JZ far [7/10] - JP Z,far [10]     +3/0
JNC far [7/10] - JP NC,far [10]   +3/0
JC far [7/10] - JP C,far [10]     +3/0
JPO far [7/10] - JP PO,far [10]   +3/0
JPE far [7/10] - JP PE,far [10]   +3/0
JP far [7/10] - JP P,far [10]     +3/0
JM far [7/10] - JP M,far [10]     +3/0

The Z80 is faster on the next instructions:
Code: 
PUSH B [12] - PUSH BC [11]          -1
PUSH D [12] - PUSH DE [11]          -1
PUSH H [12] - PUSH HL [11]          -1
PUSH PSW [12] - PUSH AF [11]        -1
PCHL [5] - JP (HL) [4]              -1
CALL word [18] - CALL word [17]     -1
RNZ [6/12] - RET NZ [5/11]          -1
RZ [6/12] - RET Z [5/11]            -1
RNC [6/12] - RET NC [5/11]          -1
RC [6/12] - RET C [5/11]            -1
RPO [6/12] - RET PO [5/11]          -1
RPE [6/12] - RET PE [5/11]          -1
RP [6/12] - RET P [5/11]            -1
RM [6/12] - RET M [5/11]            -1
RST 0 [12] - RST 0 [11]             -1
RST 1 [12] - RST 8 [11]             -1
RST 2 [12] - RST 10 [11]            -1
RST 3 [12] - RST 18 [11]            -1
RST 4 [12] - RST 20 [11]            -1
RST 5 [12] - RST 28 [11]            -1
RST 6 [12] - RST 30 [11]            -1
RST 7 [12] - RST 38 [11]            -1

The 8085 and Z80 have different timings (which are equal on average) on the next instructions:
Code: 
CNZ word [9/18] - CALL NZ,word [10/17]  +1/-1
CZ word [9/18] - CALL Z,word [10/17]    +1/-1
CNC word [9/18] - CALL NC,word [10/17]  +1/-1
CC word [9/18] - CALL C,word [10/17]    +1/-1
CPO word [9/18] - CALL PO,word [10/17]  +1/-1
CPE word [9/18] - CALL PE,word [10/17]  +1/-1
CP word [9/18] - CALL P,word [10/17]    +1/-1
CM word [9/18] - CALL M,word [10/17]    +1/-1

My pi-spigot implementation uses DAD and conditional jump instructions more often than the PUSH instructions. So the 8085 should show faster results.
 
You must log in or register to reply here.
Back
Top