Amiga.org

Amiga computer related discussion => General chat about Amiga topics => Topic started by: lunithy on September 22, 2011, 12:57:19 PM

Title: Interesting 1200
Post by: lunithy on September 22, 2011, 12:57:19 PM
(http://1.bp.blogspot.com/-mgrQMYdfv1w/Tnn0h3zfB5I/AAAAAAAAGOw/5bfq-PvgRNI/s640/060%2540100_5.JPG)

http://cosmosamiga.blogspot.com/2011/09/68060100.html
Title: Re: Interesting 1200
Post by: Hattig on September 22, 2011, 01:16:43 PM
I wish I had that sort of Amiga in 1996!
Title: Re: Interesting 1200
Post by: LaserBack on September 22, 2011, 02:05:30 PM
this overclock is well done and real
sysinfo reports accurate results ..77 mips which is exactly the double which does the 060/50
btw, sysinfo is the only tool that still report correct and accurate cpu test results

2x faster than a 060/50   38.5 mips
2.5x faster than a 040/40mhz  30 mips
4x faster than a 040/25mhz  19 mips
8x faster than 030/50
Title: Re: Interesting 1200
Post by: zipper on September 22, 2011, 02:31:32 PM
Quote from: LaserBack;660584
this overclock is well done and real
sysinfo reports accurate results ..77 mips which is exactly the double which does the 060/50
btw, sysinfo is the only tool that still report correct and accurate cpu test results


which is about half of the "real" mips as only one integer unit is used. SysSpeed gives the "right" figure above 150 if you try it - I believe.
Title: Re: Interesting 1200
Post by: LaserBack on September 22, 2011, 04:03:33 PM
Quote from: zipper;660594
which is about half of the "real" mips as only one integer unit is used. SysSpeed gives the "right" figure above 150 if you try it - I believe.


sysspeed or aibb reports wrong and inaccurate results

for example sysspeed on a 040/40 mhz reports 28 mips and on a 060/50 reports 66 mips..so means is about 2.5x faster....
that's totally inaccurate...cause the 060/50 is not 2.x faster than the 040/40....only 1.3x faster
also on winuae sysspeed reports wrong and inaccurate results ...it is a really buggy tool
Title: Re: Interesting 1200
Post by: Cosmos Amiga on September 24, 2011, 12:04:17 PM
I tried with a 110 Mhz crystal and a 3.5v regulator, but no boot with or without any Simms...
Title: Re: Interesting 1200
Post by: LaserBack on September 24, 2011, 03:00:22 PM
Quote from: Cosmos;661019
I tried with a 110 Mhz crystal and a 3.5v regulator, but no boot with or without any Simms...


maybe you can create your own regulator using a LM317 transistor and using a variable resistor you can create and try different voltages examples 3.35, 3.4, 3.40,3.45
also you must connect a big capacitor there to smooth that 3.3v  or 3.x line...something like 4700 mf or more .....if you do not use a big cap that line will not be enough stable for an overclocked 060
Title: Re: Interesting 1200
Post by: AndyLandy on September 24, 2011, 08:13:16 PM
Quote from: LaserBack;660609
sysspeed or aibb reports wrong and inaccurate results

for example sysspeed on a 040/40 mhz reports 28 mips and on a 060/50 reports 66 mips..so means is about 2.5x faster....
that's totally inaccurate...cause the 060/50 is not 2.x faster than the 040/40....only 1.3x faster
also on winuae sysspeed reports wrong and inaccurate results ...it is a really buggy tool

Interesting. I'd always heard the complete opposite, that Sysinfo doesn't use the full power of the '060 so always under-scores it, but the other programs can make use of both integer units, so will give a fairer estimate.
Title: Re: Interesting 1200
Post by: fishy_fiz on September 24, 2011, 08:19:43 PM
Actually the real truth would be somewhere in between. An '060 is roughly 1.3x as fast as an '040 at same clock speed, which'd make a 50mhz '060 something akin to a 66mhz '040,.... roughly 1.65x the speed on a 40mhz '040.
Title: Re: Interesting 1200
Post by: AndyLandy on September 24, 2011, 11:18:20 PM
Quote from: fishy_fiz;661093
Actually the real truth would be somewhere in between.

Ain't that always the way? :)
Title: Re: Interesting 1200
Post by: Kawazu on September 24, 2011, 11:47:10 PM
Quote from: AndyLandy;661091
Interesting. I'd always heard the complete opposite, that Sysinfo doesn't use the full power of the '060 so always under-scores it, but the other programs can make use of both integer units, so will give a fairer estimate.

You are right.
Dont listen to LaserBack, just read this thread and see for yourself.
Always bashing on everything that aint sysinfo.

http://www.amiga.org/forums/showthread.php?t=59096




(http://www.alinea-computer.de/amigaspeed/bilder/68kvergleiche.gif)

From: http://www.amigaspeed.de.vu/
Title: Re: Interesting 1200
Post by: Vlabguy1 on September 24, 2011, 11:49:06 PM
Neat.. I look forward to the videos..

Rich
ny
Title: Re: Interesting 1200
Post by: SpeedGeek on September 25, 2011, 12:19:57 AM
WOW! Good work Cosmos. :)
 
How could anyone really believe SysInfo is an accurate benchmark program? Would you trust a program which almost always miscalculates the CPU speed? Has a small cachable MIPS benchmark which makes 040 and 060 machines appear much faster than they really are? ...and on my A3000 9 out of 10 times reports N/A for Ramsey and Gary. Maybe thats why its called SysInfo instead of SysBench?
Title: Re: Interesting 1200
Post by: LaserBack on September 25, 2011, 05:30:00 AM
Quote from: Kawazu;661112
You are right.

http://www.amiga.org/forums/showthread.php?t=59096




(http://www.alinea-computer.de/amigaspeed/bilder/68kvergleiche.gif)

From: http://www.amigaspeed.de.vu/


don't be idiot what do you think ? that the 060 is a core i7 ? loi
that's graphic results are accurate of course were not done with the buggy sysspeed
and you can see I'm right that the 060/50 is only 1.3x faster than the 040/040
sysspeed reports 28 mips on the 040/40 and 66mips on the 060/50 and that's totally out of range and giving the 040 a weak weak integer result
to make sure what I'm talking about anyone can get sysspeed modules of both processors on aminet and compare

http://aminet.net/util/moni/A1240_040.lha
http://aminet.net/util/moni/A2k_50_60_Bliz.lha
Title: Re: Interesting 1200
Post by: stefcep2 on September 25, 2011, 08:44:48 AM
MIPS=meaningless indicator of performance.

Other than MIPS what else makes people think the 68060 @50 mhz is only 1.3 times faster than a 68040@40 mhz?
Title: Re: Interesting 1200
Post by: LaserBack on September 25, 2011, 04:36:31 PM
Quote from: stefcep2;661173
MIPS=meaningless indicator of performance.

Other than MIPS what else makes people think the 68060 @50 mhz is only 1.3 times faster than a 68040@40 mhz?


mips is something invented to calculate integer performance
you can use it to get an idea  how fast the cpu will complete certain task
however there are buggy programs like sysspeed that report wrong and inaccurate results
but the 060/50mhz will complete a task 1.3x faster than a 040/40mhz and 2x faster than a 040/25
That's because the 040 and the 060 have identical cores....in fact the 040 overclocked to 50mhz performs identical as the 060/50

and for your question ...you can use real life benchmarks to measure cpu performance
something like a picture converter or a sound converter
also someone who had both cpus...the 040/40 and the 060/50 should affirm that the 060 is not much faster...almost impossible to notice a difference on workbench
you note it on benchmarks or playing games like doom or quake where the 060 will work fast
Title: Re: Interesting 1200
Post by: zipper on September 25, 2011, 05:05:26 PM
Counting from Freescale specs 060/50 does  1.66x faster than 040/40. Check your facts  about the cores.
Title: Re: Interesting 1200
Post by: Kawazu on September 25, 2011, 05:12:46 PM
Quote from: LaserBack;661162
don't be idiot what do you think ? that the 060 is a core i7 ? loi
that's graphic results are accurate of course were not done with the buggy sysspeed
and you can see I'm right that the 060/50 is only 1.3x faster than the 040/040
sysspeed reports 28 mips on the 040/40 and 66mips on the 060/50 and that's totally out of range and giving the 040 a weak weak integer result
to make sure what I'm talking about anyone can get sysspeed modules of both processors on aminet and compare

http://aminet.net/util/moni/A1240_040.lha
http://aminet.net/util/moni/A2k_50_60_Bliz.lha

So you are calling me an idiot?
After reading that last thread nothing you write on these forums will be taken seriously.

And what are you talking about that the graph is showing that 060/50 is only 1.3 times faster then 040/40?
The 040/40 gets 41MIPS while the 060/50 gets 80 mips. How the hell do you get that to 1.3 times faster? 1.3 times faster yould be 53.3mips
Title: Re: Interesting 1200
Post by: mongo on September 25, 2011, 05:32:37 PM
Quote from: LaserBack;661209
That's because the 040 and the 060 have identical cores....in fact the 040 overclocked to 50mhz performs identical as the 060/50


That's not even close to true.
Title: Re: Interesting 1200
Post by: nicholas on September 25, 2011, 06:04:15 PM
Quote from: LaserBack;661209
That's because the 040 and the 060 have identical cores....


What drugs are you on?

Must be the milspec kind!
Title: Re: Interesting 1200
Post by: Cosmos Amiga on September 25, 2011, 06:08:22 PM
040 core : http://3.bp.blogspot.com/_ft6jgTxpbj0/TSxNeGH6AnI/AAAAAAAAFqk/qpRYDHa839c/s1600/Heart_040_4.JPG

060 core : http://1.bp.blogspot.com/-y3x21FvobL4/TVgCXzLbWEI/AAAAAAAAFu8/vT4kfn0jVtY/s1600/060_10.JPG
Title: Re: Interesting 1200
Post by: bbond007 on September 25, 2011, 06:30:15 PM
Quote from: LaserBack;661209

That's because the 040 and the 060 have identical cores....in fact the 040 overclocked to 50mhz performs identical as the 060/50


if that was true (identical cores), it would have been called the 050 and not the 060.

taken from:

http://en.wikipedia.org/wiki/Motorola_68000_family

"Odd-numbered releases had always been reactions to issues raised within the previous even numbered part; hence, it was generally expected that the 68050 would have reduced the 68040's power consumption...The 68010 was a revised version of the 68000 with minor modifications to the core, and likewise the 68030 was a revised 68020 with some more powerful features, none of them significant enough to classify as a major upgrade to the core."
Title: Re: Interesting 1200
Post by: psxphill on September 25, 2011, 06:45:17 PM
Quote from: stefcep2;661173
MIPS=meaningless indicator of performance

It actually means million instructions per second. Like any benchmark, you have to understand what it's telling you.

Just because you don't understand something doesn't make it meaningless.