|
---|
| | Alex Kemmler
Posts 26 07 May 2016 13:54
| I have seen a picture in this forum on the dyno one 68000-300 mhz, with a score of 81 mips possible? Greetings
| |
| | Samuel Crow
Posts 424 07 May 2016 16:43
| Ummm... You posted this in the Apollo core forum? The Apollo core is an FPGA softcore created to be compatible with a 68020+. A Motorola/Freescale related question should not be posted on this forum unless it pertains to the Apollo core in some way. The flat 68000 ASIC is not an Apollo softcore. Do you want to compare the two in some way?
| |
| | Gunnar von Boehn (Apollo Team Member) Posts 6254 07 May 2016 17:46
| Some background information : Motorola did produce the following 68K CPUs 68000 68010 68020 68030 68040 68060 None of these CPUS has something like an ID-Register. This means Software has _NO_ easy way to find out whether it runs on an 68020 or 68040 or 68060 ... Also there is NO easy way to read out the MHZ a CPU is clocked. How do AMIGA benchmarks programs then print this out? They run some instructions and based on how the CPU react on them, the programs "guesses" which CPU it is. Then the program runs some testloop and measures the time. Based on the measured time it then calculates the Mhz. APOLLO CORE is not from Motorola. APOLLO is by far most arvanced of all 68K cores also much more advanced than original 68060. If Motorola would have build Apollo they might have named it 68080. If you see in some AMIGA benchmark screenshots on our website this: 68000 @ 1300MHz or 68040@ 400 MHz Then this means the program did guess that Apollo might be an 68000 and based on the measured speed it "calculated" that Apollo must be running at 1500 Mhz. Or it might have guessed that Apollo is an 68040 and based on the measured speed calculated that it must be running at 400 Mhz. Did this explanation help you?
| |
| | Alex Kemmler
Posts 26 07 May 2016 19:57
| Hello I meant if the apollo I guess it will be a software can reach 300 mhz to 68000? Someone can explain me? By the way powerpc is not motorola?
| |
| | Martin Soerensen
Posts 232 07 May 2016 20:48
| The PowerPC instruction set was developed jointly by Apple–IBM–Motorola and was based on IBM's POWER instruction set. Regarding your 68000 question, I'm not sure I understand what you mean. The Apollo core is not a 68000 core as previously explained. If you mean if the performance of the Apollo core could be similar to that of a theoretical 300MHz 68000 CPU, I'd say it already is in most areas, and in some areas quite a lot faster than that.
| |
| | Mercury Thirteen
Posts 9 07 May 2016 20:55
| Apollo is not a piece of hardware in itself, it is a "soft core" which gets loaded onto FPGA hardware to do its work. Since this is the case, Apollo itself theoretically has no speed limit whatsoever. The maximum speed at which it can be clocked is limited solely by the FPGA on which it is loaded. The higher capacity FPGA used, the higher speeds it can achieve. This also means that Apollo users are not locked into a certain speed range; as newer, faster FPGAs come to the market, the Apollo core can be loaded directly onto them to take advantage of their speed benefits with zero modifications for the purpose of doing so.
| |
| | Alex Kemmler
Posts 26 07 May 2016 20:57
| 68000 came in speeds of 8, 12,16 and 20 mhz, as it is possible that you can get 300 mhz, that is the question? Greetings
| |
| | Mercury Thirteen
Posts 9 07 May 2016 21:01
| Yes, that would be possible as long as Apollo is loaded onto an FPGA which can handle that speed.
| |
| | Martin Soerensen
Posts 232 07 May 2016 21:25
| It would not be possible to make a 300 MHz 68000 CPU even using today's fastest technology, and maybe it never will be due to physical limits (super-cooling might help get there). What you can do however, is to modify the structure so more things can happen in parallel and this is how you make faster CPUs. Thus, you could make a CPU which is almost 100% compatible with a 68000 and as fast as a 300 MHz 68000 would be, but it will not be 100% compatible. Some products use an overclocked Freescale (latest model) 68000 CPU clocked at ~50MHz.
| |
| | Thierry Atheist
Posts 644 07 May 2016 21:43
| I believe that the question is "this FPGA, if it was using only 68000 CPU instructions, what is the equivalent in 68000 MHz that it would be?".
| |
| | Lord Aga (Apollo Team Member) Posts 119 08 May 2016 00:09
| It looks more like: "Can someone make a 300MHz 68000 out of this carrot and fly swatter I'm holding?"
| |
| | Alex Kemmler
Posts 26 09 May 2016 11:09
| The 68,000-12 mhz few mips achieved? Certainly the 68060, can go beyond 90 mhz or explode I mean by the heat generated?Thank you for listening
| |
| | Alex Kemmler
Posts 26 09 May 2016 21:41
| Someone can help me in this doubt? Greetings
| |
| | Daniel Sevo
Posts 299 09 May 2016 22:01
| Alex kemmler wrote:
| Someone can help me in this doubt? Greetings
|
Hello Alex, 84 posts here: EXTERNAL LINK ...and you keep asking the same questions. Maybe if you tell wehat is your native language and someone who speaks it can explain it better..? So you dont have to use google translate?
| |
| | Lord Aga (Apollo Team Member) Posts 119 09 May 2016 22:06
| Guess I called it correctly. And I haven't even seen that thread before.
| |
| | Alex Kemmler
Posts 26 09 May 2016 22:16
| That person or member is not me
| |
| | Thierry Atheist
Posts 644 09 May 2016 23:01
| This is a difficult question to answer, as the CPU internally is 32 bits, and externally 16. 2 clocks to get an instruction, and 2 clocks to get one piece of data, and 2 clocks to push one out. If the 68000 was fully 32 bit alone, at 7.12 MHz the Amiga 500/600/1000/2000 would have been nearly twice as fast as they are!!!! Now THAT is TRAGIC!
| |
| | Nixus Minimax
Posts 416 10 May 2016 12:17
| Thierry Atheist wrote:
| This is a difficult question to answer, as the CPU internally is 32 bits, and externally 16. |
Actually the 68000 is also 16 bit on the inside, the only 32bit part were the registers. E.g. a 32bit ADD is fed to the 16bit ALU in two halves.
| |
| | Alex Kemmler
Posts 26 10 May 2016 17:34
| 32 bits works internally but would not be compatible with the following 68000
| |
| | Captain Zalo
Posts 71 10 May 2016 19:10
| Alex kemmler wrote:
| 32 bits works internally but would not be compatible with the following 68000
|
Hi Alex. I think a lot of the people reading this thread don't understand what you're asking or trying to say. Would you be so kind to post your question with some background in your native language? It seems a lot of what you're writing and/or asking is disappearing in translation. Please? Regards; Zalo
| |
|