Topic title says it all, of course. I always shout: 1!!! But, that not true ofcourse. I could analyze lot's of sources and stuff (or however they count these things) but there's probably someone here who knows the answer. I would guess 0.62.

I tought you didn't care about speed

MIPS - Meaningless Indication of Processor speed.

Offtopic: I'm building a lib with the original opl4 patches,including an easy why to calculate the oct and f-num. Maybe you can use it for RealFun?

*cough* vapourware *cough*

Quote:

I tought you didn't care about speed

Ofcourse I care about speed. I just think a msx2 at 3.5MHz is fast enough.

Offtopic:
I think your opl4-lib wil be very useful for Realfun. But I should see it first, ofcourse .

Quote:

*cough* vapourware *cough*

uhm, is Realfun under attack here? In that case, could you give me your definition of 'vapourware'?

Hehe, just prove me wrong, please

Quote:

Hehe, just prove me wrong, please

Should I read this as: "No, I can't give you my definition of 'vapourware'?"

If you insist: vapourware, software that has been announced but will probably never be released in a finished and usable form.

Note that there are some vapourware titles partly on my name, so don't feel bad

Realfun will be finished, if not I will feel bad

Also because Realfun is needed for other vapourware... (Hehe, the lost world, or the last dragon (to avoid confusion with the Umax product), for instance. Now THAT's vapourware ).

if you say so

BTW, enter.mp3 sounds cool. Is that realfun or mbwave?

Well, 3.57 Mhz = roughly 3.57 million T-cycles per second.

an instruction is minimal 4 cycles and maximum ... eeh .. 24 or so.
average about 12 cycles per instruction, so 3.57 / 12 = 0.29 Mips.

hmm, that doesn't sound impressive.

Quote:

BTW, enter.mp3 sounds cool. Is that realfun or mbwave?

Thanks 8)
It's Realfun, ofcourse. If you think Microsoft's marketing strategy is agressive, you don't know mine. I have forced everybody I know (and has a MoonSound) to use Realfun. Yes, I am bad I should finish Realfun just for them...

Quote:

Well, 3.57 Mhz = roughly 3.57 million T-cycles per second. an instruction is minimal 4 cycles and maximum ... eeh .. 24 or so. average about 12 cycles per instruction, so 3.57 / 12 = 0.29 Mips. hmm, that doesn't sound impressive.

But I think the shorter instructions are used more than the longer ones. So, It is a bit more. We should introduce tips (thousand instructions per second) Just to sound more impressive...

Quote:

MIPS - Meaningless Indication of Processor speed.

True, usually people talk about MFLOPS when they measure speed. According to my quick tests standard MSX speed is about 0,02 MFLOPS MSX tR got 0,18 score in my test.

To give you some idea: 486DX4/100 performance is about 6 MFLOPS

~NYYRIKKI

However, what none of these measurements take into account is the high CPU-load PC OS-es, legacy limitations and drivers take with them.

Because it's possible to adress the MSX hardware directly you can make a lot more use of the speed available. I wonder if it's possible to make a portable benchmark that takes this into account.

I have to correct you a little on that one snout, my friend.

Ofcourse PC can be programmed to address hardware directly.
You're mistaken this with the Windows crap, which does (ofically) not allow direct hardware (also mem) writing.

A benchmark testin MIPS etc, has to be done in DOS for example (although I think the same result are with windows, not sure)

Many demogroups on PC do NOT use windows, but code in DIRECT-acces ASSEMBLY. (moslty under DOS).

ah, well. who gives a fuck anyway