Re: Robotron CPU performance in relation to marathoning?
No problem, but the link i posted above is from KLOV where Clay provided the technical insight into the why's and how's of my observations. I consider this a closed topic for what I was curious to confirm.
Robotron gameplay is driven by the CPU performance.
And CPU performance involves the componentry on the board.
I have messed with 4 sets of ram
1-original 30 yr old set of 4116 ram, all good (meaning they always pass the boot up test)
2- original 30 yr old set of 4116 ram, flakey (meaning that varying rams will fail about every 5th boot up test, but game still plays fine)
3- new set of 4116 ram from bob roberts
4- new set of 4164 converted rams from bob roberts.
Ram set 2 makes the game amazingly easy...boring
Ram set 1 makes the game play like i observe in online videos
Ramset 3 is tough game play, but manageable. (i'm going to assume this was like game play in 1981)
If I could stay up 100 hrs, and put an air conditioner unit blowing on the CPU, i'd play some hokey old ram and whip ***** on that fake record. And I'm just an average player at this point in my game.
If I were a turkey, I'd practice a few months on flakey ram and try to get close to the 1million 5 man game..ha ha ha (don't worry redelf my best so far is about 350k on a 5 man game with ram set 1...uggggh)
Re: Robotron CPU performance in relation to marathoning?
Did any of the 4 sets of ram let you get a score that ends in 680? TG's been carefully investigating a score like that for the last few years but don't seem to have reached a conclusion yet.
Re: Robotron CPU performance in relation to marathoning?
Oh boy. I almost didn't want to know that.
This adds a new dimension to the topic of authenticity/legitimacy of records: records broken in the last 10 years may have been done so on 20-30 year old hardware, with lessened performance, whereas the early 80's records (on the same titles) are all set on crisp hardware, thus with a faster gameplay.
But my Defender angle on all this is to try to have a dedicated/original machine actually performing better/faster than it ever could when it was new, with the help of new/better hardware. And this is possible as well, as I understand it from your initial post?
Re: Robotron CPU performance in relation to marathoning?
I have little defender skills, actually none. Going to make a control panel and learn it on 19-1.
But Defender uses the same Williams hardware so it appears the concepts would carry over.
Eugene J. was involved in both games so we know his assembly language design skills would have been the same for the coding.
I'm going to assume if you researched the timing circuits for the CPU and optimized the parts involved (caps..) and put in a better/faster crystal that the cpu is going to step faster which means the game is going to move faster.
And if you have fast fresh ram then the cpu is going to be managing the temp files faster.
And putting in fresh video decoder roms using brand new chips is probably further going to enhance the performance.
looks like the game would be a lot harder to me.
You are right though about scores and I heard this same sentiment from an arcade owner. Benchmarking old arcade game performance is an issue yet to be explored, when you are dealing with 30 year old hardware and worn out components. Probably a moot subject too until people start earning big $$$ for breaking records and that's just silly talk. Comes down to a phrase told to me by Jeff Ramsey, "My goal is to just enjoy myself while playing the game." Cheers to that!
Quote:
Originally Posted by MOS
Oh boy. I almost didn't want to know that.
This adds a new dimension to the topic of authenticity/legitimacy of records: records broken in the last 10 years may have been done so on 20-30 year old hardware, with lessened performance, whereas the early 80's records (on the same titles) are all set on crisp hardware, thus with a faster gameplay.
But my Defender angle on all this is to try to have a dedicated/original machine actually performing better/faster than it ever could when it was new, with the help of new/better hardware. And this is possible as well, as I understand it from your initial post?
Re: Robotron CPU performance in relation to marathoning?
Quote:
Originally Posted by 1500points
Here is some technical insight from Clay which seems to support the observations.
The CPU has everything to do with the performance experienced.
So that would tend to lend itself to new ram having a performance effect.
Looks like caps and crystals would come into play too.
So my final view on this topic as I let it drop, is.....
Robotron as the code is designed, and run from an efficient/fast/new school CPU is too hard for the human player.
The marathoner is benefiting from the "lag" created on the old school CPU by lots of action on the screen.
I'm going to say if an old original CPU was updated with fresh components, that the 1 million 5 man score would not happen.
Fascinating stuff. :)
http://forums.arcade-museum.com/showthr ... 82&page=71
I'm not sure you really understood what Clay was trying to say. The game speed is not CPU dependent at all as it relates to operating speed. That is set by the clock crystal. You could put a 2x fast CPU in there but it'd still take 2.5 seconds (or whatever) to run the length of the board. Now if you has 500 enemies all on the screen moving and firing it would take longer to run the screen width because the processor would have to be drawing all the items and doing all that math and it would be VERY obvious that EVERYTHING had slowed down.
There may be some degradation of the clock signal slowing things a TINY bit on some machines over another but i'm quite sure none of us would be able to tell.
The only real way to prove what you think you are seeing its to prove or disprove it with experiment. EVERYTHING else is just a guess and based on your (or someone elses) perception.
Those of you worrying about older games not playing like they did when they were new can relax. Look at Pacman, an absolute pattern based game. If age made the game slower or faster harder or easier these guys would know FOR SURE. We would hear differenced in the tones of the music in mappy or dig dug, differences in the rate at which pacman eats dots or pole position rpm's up.. We dont here about this happening because it dont. If something is obviously wrong its likely real hardware failure and not some minor degradation of caps and crystals.
-Ken
-Ken
Re: Robotron CPU performance in relation to marathoning?
The game is at Tim McVey's house. I believe you were just on the phone with him the other night?
He has streaming capabilities and there are 2 CPU's.
Experiment away! ;)
It'll be an interesting technical experiment. Love to see the outcome in statistical detail.
Quote:
Originally Posted by Scoundrl
Quote:
Originally Posted by 1500points
Here is some technical insight from Clay which seems to support the observations.
The CPU has everything to do with the performance experienced.
So that would tend to lend itself to new ram having a performance effect.
Looks like caps and crystals would come into play too.
So my final view on this topic as I let it drop, is.....
Robotron as the code is designed, and run from an efficient/fast/new school CPU is too hard for the human player.
The marathoner is benefiting from the "lag" created on the old school CPU by lots of action on the screen.
I'm going to say if an old original CPU was updated with fresh components, that the 1 million 5 man score would not happen.
Fascinating stuff. :)
http://forums.arcade-museum.com/showthr ... 82&page=71
I'm not sure you really understood what Clay was trying to say. The game speed is not CPU dependent at all as it relates to operating speed. That is set by the clock crystal. You could put a 2x fast CPU in there but it'd still take 2.5 seconds (or whatever) to run the length of the board. Now if you has 500 enemies all on the screen moving and firing it would take longer to run the screen width because the processor would have to be drawing all the items and doing all that math and it would be VERY obvious that EVERYTHING had slowed down.
There may be some degradation of the clock signal slowing things a TINY bit on some machines over another but i'm quite sure none of us would be able to tell.
The only real way to prove what you think you are seeing its to prove or disprove it with experiment. EVERYTHING else is just a guess and based on your (or someone elses) perception.
Those of you worrying about older games not playing like they did when they were new can relax. Look at Pacman, an absolute pattern based game. If age made the game slower or faster harder or easier these guys would know FOR SURE. We would hear differenced in the tones of the music in mappy or dig dug, differences in the rate at which pacman eats dots or pole position rpm's up.. We dont here about this happening because it dont. If something is obviously wrong its likely real hardware failure and not some minor degradation of caps and crystals.
-Ken
-Ken
Re: Robotron CPU performance in relation to marathoning?
Defender and Stargate don't have this emulation problem.
http://www.yakyak.org/viewtopic.php?p=565150
It's not degrading chips or crystals. It's all about blitters, and the fact the hardware is doing two things at once. mame is too good at getting everything done each refresh. I've been led to believe a perfect emulation of Robotron is virtually impossible. This guy Aaron Giles did do some work on mame to try to compensate. He did make it better, try Robotron in a .8x version of mame :twisted:
http://aarongiles.com/?m=200504
Re: Robotron CPU performance in relation to marathoning?
very cool. i'm going to check out that mame, but now i've gotta figure out joystick interface for computer....
Re: Robotron CPU performance in relation to marathoning?
OK, so to bring this whole observation to an end, I am looking for consensus.
Is it considered acceptable to run the new 4164 ram in a williams CPU, which gives benefit of only running on plus 5 volts dc so it stays cooler and doesn't fail over long hours of play?
I'm saying that it shouldn't matter based on the views presented in this thread. And 4164 ram is a beautiful thing even though it isn't stock to an original williams cpu.
Right?????
Re: Robotron CPU performance in relation to marathoning?
Quote:
Originally Posted by 1500points
OK, so to bring this whole observation to an end, I am looking for consensus.
Is it considered acceptable to run the new 4164 ram in a williams CPU, which gives benefit of only running on plus 5 volts dc so it stays cooler and doesn't fail over long hours of play?
I'm saying that it shouldn't matter based on the views presented in this thread. And 4164 ram is a beautiful thing even though it isn't stock to an original williams cpu.
Right?????
I've upgraded hundreds of Williams machines to the 4164's and it's a great fix.... in my opinion a MUST for any Williams game of that era. Never had to go back and replace a 4164 yet... but 4116's were a problem even when the games were new.
I know I wouldn't have raised any issue over it while I was on the staff, being familiar with the fix and why it's needed and all.... and the odds are way high a great many of the Williams game scores already in the database have this fix done.