An FPGA, for the most part, is a massive piece of silicon that contains a boatload of what are known as "Logic Elements", each of which can accept a set of inputs, and generate a set of outputs based on a lookup table, which are dynamically "wired" together by *software*.
-
Näytä tämä ketju
-
The bitstream that gets uploaded to an FPGA, unless it's concretely set in some ROM, is ostensibly software. It can be changed at run-time to configure the functionality of those logic elements. It is far more akin to software than even firmware.
2 vastausta 0 uudelleentwiittausta 8 tykkäystäNäytä tämä ketju -
Vastauksena käyttäjälle @TheMogMiner
When a CPU runs microcode is it hardware or software? If someone turned the verilog into an ASIC is it still software?
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjälle @Darren__O
If someone turns the Verilog into an ASIC, is it suddenly a more accurate representation of the hardware?
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
Vastauksena käyttäjälle @TheMogMiner
No. But if the verilog was derived from a chip decap it is.
2 vastausta 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @Darren__O ja @TheMogMiner
It's still a reimplementation. Take the 68000 for instance, all its internals uses latches while a fpga can only do flip-flops. fx68k is beautiful and externally identical to the real thing, but the internals are very different in practice.
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @o_galibert ja @TheMogMiner
Documenting those differences between emulation and the real thing is important. Many games in MAME are rendered using a frame buffer and blitting instead of line rasterization. The end result is identical from a player's perspective but not to someone that is fixing hardware
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @Darren__O ja @TheMogMiner
I really doubt people use verilog mister cores as reference instead of the infinitely more readable and patchable C++ of mame.
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @o_galibert ja @TheMogMiner
I completely agree. My point is that C++ code also may not be a good reference of how the hardware works.
1 vastaus 0 uudelleentwiittausta 0 tykkäystä -
Vastauksena käyttäjille @Darren__O ja @o_galibert
And *my* point is that anyone can (and should) submit improvements to MAME that address such issues. But it's a bit difficult when one set of developers seems interested in sharing no knowledge whatsoever, while others share it freely.
1 vastaus 0 uudelleentwiittausta 3 tykkäystä
It's not a matter of MAME "not being so". It's a matter of people abusing others' nostalgia to cop some bucks while saying with a straight face that MAME "can't" be so. It could, if information was shared, but it isn't. So at least have the balls to acknowledge the fuck-over.
-
-
If the MAME team, or the general public, apparently doesn't "deserve" information after 24 years of striving for accuracy, before some FPGA devs were ever born, say so. If it's because there's some perception that we haven't put enough collective dollars into the pot, say so.
1 vastaus 0 uudelleentwiittausta 1 tykkäys -
If it's that we haven't kissed enough collective ass, say so. If it's out of some Mafia-like desire that we all kiss the literal ring rather than the proverbial one, say so. But at least mount up the balls to make it clear what the situation is.
1 vastaus 0 uudelleentwiittausta 0 tykkäystä - Näytä vastaukset
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.