Request for simple speed test

Page 1/2
| 2

By Manuel

Ascended (19468)

Manuel's picture

15-05-2005, 21:58

Hi,

Can you please do the following little test for me:

1 - disconnect all hardware from your MSX (like cartridges and slot expanders)
2 - boot normally, you should get into basic
3 - type exactly this (add no spaces or other things, just *exactly* this):

TIME=0:FORI=0TO10000:NEXT:?TIME

4) after a while (maximum 16 seconds), a number appears. Please post the model of your MSX and this number in this forum topic! Smile To be sure, try it a few times, the number should be the same all the time.

5) thanks!!

Login or register to post comments

By roadfighter

Champion (461)

roadfighter's picture

15-05-2005, 22:23

Hi,

Oke fine, but nou i am a litle courious, where is it for?

Turbo-r GT in R800 (normal boot) = 181

Turbo-r GT in z80 (boot with "1" pressed) = 1033

Sony HB F-700P = 862

NMS-8245/00 = 871

nms-8235/20 = 871

NMS-8280 with mem expansion 1 MB Sim. = 864

I am no genius when it coms to programming but, why are the resuls so different ?

By Manuel

Ascended (19468)

Manuel's picture

15-05-2005, 23:21

It is for testing if we get the same kinds of results on openMSX.

The results are sometimes very different for several reasons. One of them is that some machines run on a 60Hz interrupt (60 counts per second) and some on 50. Other differences are in the stuff that the MSX has to do in the interrupt routine, which is BIOS dependant. It can depend on installed software or hardware extensions, if I'm correct.

Thanks for submitting your results! Your turboR gives exactly the same results as mine Smile

By Edwin

Paragon (1182)

Edwin's picture

16-05-2005, 02:58

To prevent you from collecting a lot of unreproducible data, I wrote a small benchmark tool. It tests cpu speed, VDP port writes and VDP VRAM writes. Numbers should not be sensitive to any present hardware or current screen frequency. With the exception of maybe an external memory mapper. Especially a turbo R has some odd behaviour with regard to those.

You can get the tool here:
http://quor-toth.balpol.tudelft.nl/spdtest.com

Numbers for my FS-A1ST are (cpu/vdp/vram):
Z80: 1733/571/719
R800: 265/1330/1702

(notice the horrid slowdown of the VDP compared to the Z80)

On openMSX I generated:
Z80: 1733/528/664
R800: 287/1222/1565

Oh yes, the benchmark gives a lot of flicker in the middle part. This is normal. :)

By Manuel

Ascended (19468)

Manuel's picture

16-05-2005, 17:34

For my FS-A1GT:
Z80: 1734/571/719
R800: 277/1330/1702

openMSX 1724:

Z80: 1733/528/664
R800: 287/1222/1565

By tcm1998

Expert (120)

tcm1998's picture

16-05-2005, 17:47

here are the results from my 8250 and that speedtest in three situations:

1) All hardware removed
2) With slotexpander, 4MB external mapper, IDE, SCC,Philips MusMod, but using DOS 1 (Boot with INS).
3) Same as situation 2, but under DOS2

VDP port write test: 528 in all cases
VRAM write test: 664 in all cases

Since the CPU test varies, I have repeated the test several times in all situations

situation 1: 1733, 1733, 1732, 1732, 1734, 1734, 1734
situation 2: 1734, 1732, 1732, 1734, 1734, 1734, 1734
situation 3: 1733, 1734, 1732, 1734, 1734, 1734, 1734

it seems to stabelize to 1734 everytime. Something to do with cache ?

By SolidEric

Champion (332)

SolidEric's picture

16-05-2005, 18:09

Hey Manuel:

On my 8250 : 870

greets Eric

By Manuel

Ascended (19468)

Manuel's picture

16-05-2005, 18:11

btw, plain 8250 on openMSX 1724: 1733/528/664

By Manuel

Ascended (19468)

Manuel's picture

16-05-2005, 19:33

SolidEric: Thanks! What about Edwin's more extensive test program?

By SolidEric

Champion (332)

SolidEric's picture

16-05-2005, 19:56

Okay, here the NMS8250 tested with edwin's program: tested 3 times

CPU: 1732,1734,1734
VDP port write: 528,528,528
VRAM write: 664,664,664

By Manuel

Ascended (19468)

Manuel's picture

16-05-2005, 20:06

By the way, for the first test, the Sony HB-F700D on openMSX gives 862 (the same as the real F700P).

Page 1/2
| 2