Rate your MSX - Speed Test

Supporter (10)

10-02-2013, 05:44

100 PRINT "CREATING ARRAY"
110 DIM A(150)
120 FOR J = l to 150
130 A(J) = 151–J
140 NEXT J
150 PRINT "SORTING"
160 EX = 0
170 FOR K = 0 TO 149
180 IF A(K) > A(K + 1) THEN T = A(K) : A(K) = A(K + 1) : A(K + 1) = T : EX = 1
190 NEXT K
200 IF EX < > 0 THEN GOTO 160

My Performance (execution time):
SONY Hit-Bit F1XD 6:45

From: http://www.atarimagazines.com/compute/issue56/107_1_MSX_IS_C...

His results:

Quote:

IBM PC 5:45
GoldStar MSX 6:20
Apple II Plus 6:24
Apple IIc 6:33
Commodore VIC-20 6:34
IBM PCjr 6:59
Commodore 64 7:02
Commodore 8032 7:16
TRS-80 Color Computer 8:01
Commodore 16 8:35
Commodore Plus/4 8:36
Atari 800XL 8:55
Atari 800 9:00
TI-99/4A 12:58

Aangemeld of registreer om reacties te plaatsen

Van Manuel

Ascended (18864)

10-02-2013, 08:35

I measure 6:41 on a Toshiba HX-10 (using MSX-BASIC to time it with the TIME variable). On a turboR in R800 mode it's 1:14

Van Manuel

Ascended (18864)

10-02-2013, 09:11

If you add DEFINTA-Z at the start.... I get 4:48 on a normal MSX.

Van meits

Scribe (6509)

10-02-2013, 11:13

Panasonic FSA1WX:
60Hz turbo on: 5:08
50Hz turbo on: 5:03
60Hz turbo off: 7:05
50Hz turbo off: 6:45

Used TIME

Van Manuel

Ascended (18864)

10-02-2013, 22:45

Meits: on 60Hz, did you divide by 60 instead of 50 to calculate from TIME to seconds?

Also, where did you put the TIME=0 and ?TIME statements?

Van anonymous

incognito ergo sum (116)

11-02-2013, 19:29

well... what do you think the result of TIME was devided by the frequency I was running at...
time=0 in line 90 and ? time in the first new line after the program was done...

Van Colemu

Hero (539)

11-02-2013, 21:27

Same model as stm4il

Normal speed: 2:51
7 Mhz: 1:26

Using a stopwatch, cos I'm thick

Van Manuel

Ascended (18864)

11-02-2013, 21:43

heh, interesting that it's so much slower on 60Hz. I get the same values in openMSX with that machine for 60Hz turbo off: 7:04, but I get 6:56 on 50Hz turbo off... Not quite the same difference!

Van Manuel

Ascended (18864)

11-02-2013, 21:39

Colemu: I doubt that! That's way too short. How did you get the program on your MSX?

Van Colemu

Hero (539)

11-02-2013, 21:59

You're right Manuel, a typing error:
120 FOR J=1 TO 50
instead of
120 FOR J=1 TO 150 (Red Face)

Revised times:
Normal: 6:52
7 Mhz: 3:27
Normal slightly slower than stm4il's time, but that makes sense as my F1XD runs at a slightly slower 3.56Mhz

Van Manuel

Ascended (18864)

11-02-2013, 22:18

Aha, I can explain the differences Meits has. He was running after booting with an IDE interface. This severely affects the times. (I guess due to Disk Basic 2.)

So, when doing this test, start up without any carts/expanders in the machine

I noticed some other issues:
- if you type over the program with different spaces, the program will have a different timing
- if you do not take over the typo in the listing above (l instead of 1 in line 120), it will loop from 0 to 150... so that will also affect the timing...

So, let's start over and use exactly this program:

90 TIME = 0
100 PRINT "CREATING ARRAY"
110 DIM A(150)
120 FOR J = 1 to 150
130 A(J) = 151-J
140 NEXT J
150 PRINT "SORTING"
160 EX = 0
170 FOR K = 0 TO 149
180 IF A(K) > A(K + 1) THEN T = A(K) : A(K) = A(K + 1) : A(K + 1) = T : EX = 1
190 NEXT K
200 IF EX < > 0 THEN GOTO 160
210 PRINTTIME