Page 1 of 1
TIIDO'S SEGANE MEGA BENCHMARK !!!
Posted: Wed Jun 27, 2007 11:27 am
by TmEE co.(TM)
This is a REAL benchmark for MD/Genny. It tests most of the 68K instructions in almost all possible ways. All benchmark code is run 63356 times (so it takes several minutes to complete) and while doing it, HBLs are counted up. The less HBL it takes, the better the performance. Next version (If there will be any) will use z80 and YM2612 for timing, to get exact same results on PAL and NTSC machines.
http://www.hot.ee/tmeeco/DWNLOADS/SEGANE.RAR
SEGANE means CRAZY in Estonian
My MD2 results:
PAL: 1268775, 10MHz OC: 935261
NTSC: 1571456, 10MHz OC: 1147458
Lower score means better performance !!!
Posted: Wed Jun 27, 2007 7:57 pm
by TmEE co.(TM)
I updated the benchmark, now it uses Z80 and YM2612 timers so PAL and NTSC systems give same results. Enjoy
My MD2 being benchmarked

the benchmark channel

Posted: Thu Jun 28, 2007 1:54 am
by evildragon
Can you try and compile it for Sega CD? I just need a .bin file that is compatible with the MoD Sega Loader..
If i'm not mistaken, it's just a start address change. Maybe MoD can help you with that. (i still can't port code for some reason)
Posted: Thu Jun 28, 2007 2:28 pm
by Mask of Destiny
For a small program like this you usually need to do the following:
Change the start address to either $200000 or $FF0000
Remove ROM header and TMSS check
Hook V-INT (if present) using a BIOS call
Adjust DMA commands if DMA-ing from Word RAM ($200000) to deal with a small hardware bug.
Posted: Fri Jun 29, 2007 8:32 am
by TmEE co.(TM)
Luckily no ints are used, same for DMA. What happens if TMSS check isn't removed ?
Posted: Fri Jun 29, 2007 3:50 pm
by Mask of Destiny
To be honest, I don't remember. You might end up locking the VDP though and even if it has no negative effects it's not needed for Sega CD programs.
Posted: Fri Jun 29, 2007 11:46 pm
by 8bitwizard
As long as the TMSS code checks the hardware version first it should have no negative effect.
Posted: Sat Jun 30, 2007 3:26 am
by evildragon
8bitwizard wrote:As long as the TMSS code checks the hardware version first it should have no negative effect.
Sega CD is already running, thus the system has already booted.
If I remember correctly, if TMSS code is used, the result, atleast on my Genesis (which has no TMSS), was a screen filled with "@"... Sometimes I had nothing though...
But the thing is, the Sega CD already initialized the hardware, so the TMSS code would be bad...
Posted: Sat Jun 30, 2007 8:38 am
by TmEE co.(TM)
I guess it doesn't do anything, I'm checking if there is TMSS. Too bad that SEGA didn't change the hardware version of later models, all later models have 1 as version.
Now I'm going to check what happens if you write SEGA more than once...
Is there anyone who is actually running this program ?
Posted: Sat Jun 30, 2007 4:41 pm
by evildragon
since I have no way of running it from a cartridge, I can't run it, until it's compiled for a Sega CD .bin...