Cs MKII

[Whaka]

Wow, I didn't even know it was possible to clock them that high. Any side effects? I thought the onboard SCSI failed above 66MHz or so.

no side effect at all. the mod consist to put the MkII in 040 mode only by moving 2 switch.
by doing this, you divide the bus frequency by 2. so, when you run a 060 at 100 Mhz, the bus run at 50.
so, the scsi work, and no stress at all for the logic and ram.

the only side effect is your 060 wasting a lot of cycles to wait for ram access
a real exemple : doom will run more fast at 72-75 Mhz with a 1/1 bus than 100 Mhz with a 1/2 bus...

mips are cool, but... that's not all.

Sure would like that on my MKIII!

you can, the same way. but you need to switch the cpu voltage to 5V and get a 040>060 adapter.
but you'll face the same drawback. you get more mips on benchmark, but you bottleneck ram access...

in short : i'll never trade my Mk3 running 75Mhz for a Mk2 running 100 Mhz
i have both, so i experienced.

[Whaka]

for the moment, i can only post some screenshots with sysspeed and bustest with my Mk3 at 75 (74,3 Mhz wichamiga say) , set to 60ns for ram in firmware.

[ATTACH]5778[/ATTACH]
[ATTACH]5779[/ATTACH]

i don't have screenshots with Mk2. and as it run at 100 Mhz too, it won't be useful...

but finding somone else with a OC'ed Mk2 running 1/1 bus to compare is doable i think... HammerD can you help ?

[Whaka]

Is there any form of real life application we can benchmark and see for example fps difference in a game or somthing?

yes,  doom is a real exemple. i can do a little video later today.
and i think quake have a benchmark mode too.

[Whaka]

here it is  : https://www.youtube.com/watch?v=9zCHGWytf9Q

[Whaka]

argh !! i guess it's beacause i didin't think to mute music i was listening... youtube bot probably did their work...

[Whaka]

it was doom 2.
so, i did

[Whaka]

it was doom 2.

i did a quick shot with the same version as your : https://www.youtube.com/watch?v=o78ac89CHPQ&feature=youtu.be

just simply look at the first screen : i got 32 fps, where you get 17.

[Whaka]

believe me or not, but if you roll back your mk2 to 72 Mhz, you'll have same results.
i'm gonna do this on mine. 100 Mhz with a bottlenecked bus not really worth it...

[Whaka]

64 bits memory doesn't do very much in this case.
in this case, this is not specially doom. and this is not specially graphic card too.
this is simply the effect of the 060 wasting ton of clock cycles to read/write in memory when the bus is divided by 2... it's not very hard to understand.

it's a bit like if you get a car engine from 72 to 100 HP, but divided the gas pedal run by half. (i don't know if "run" is the good word in this context but english is not my native language)

look : here is my Mk2 at 66 Mhz 1:1 bus : https://www.youtube.com/watch?v=nPKMsKGj_1g

compare here at 100 Mhz : https://youtu.be/9zCHGWytf9Q?t=73

what do you see ? and please watch carefully

i have a very little bit better framerate at 66 Mhz... why ?
because at 66 Mhz, the bus is synchronous, and the 060 despite the fact there's less raw mips... waste a lot less clock cycles to access memory.

[Whaka]

well, if you won't uderstand the simple fact that a synchronous bus is better than a bottlenecked bus in general, i can't do much.

the simple facts you can see in my videos, doom doesn't load slower and get around 1,5 fps better on the very first screen at 66 Mhz synchron, speak by itself.

did you realize that we speak about 34 Mhz less here ??
how the hell i can get better results

[Whaka]

ok, as we mostly have same configurations, you can give me some exemples to test.
as you have a 100 Mhz and me a 66 Mhz Mk2 now, we can try to do comparisons in general applications, such as loading webpage or pictures, or mp3 encoding

and yes, i said in general.
what the point of overclocking if it's for losing more than you are supposely win ?
for me, overclocking is to try getting a win/win situation...

at this point, if it's only for saying "yay ! my 060 run 100 Mhz, look those angry mips in sysspeed" but loosing half of them in memory waitstate... errr