06-07-2014, 02:42 PM
Bigpet
06-07-2014, 02:56 PM
Don't we have the "No ETA questions" rules yet? If not, I would like to have one added.
flashmozzg
06-07-2014, 03:05 PM
It's not ETA, more like "ping". Falls into same category, in my opinion.
06-07-2014, 03:57 PM
(06-07-2014, 02:56 PM)AlexAltea Wrote: [ -> ]<[DH]> i have recompiler, but i try to optimize it and broke
<[DH]> i'll send recompiler at this month
(It's Xbyak based, btw)
Interesting. So is this a more advanced recompiler that not only directly translates ppc -> x86 assembler but also does advanced stuff such as eg loop optimizations for dat speed? (or whatever is appropriate, I am sure the the original binary is heavily optimized for the cell cpu, but different cache size or maybe even pipeline on x86 may mess up the optimizations done)
06-07-2014, 09:31 PM
(06-07-2014, 02:56 PM)AlexAltea Wrote: [ -> ]<[DH]> i have recompiler, but i try to optimize it and broke
<[DH]> i'll send recompiler at this month
(It's Xbyak based, btw)
(06-07-2014, 02:56 PM)Bigpet Wrote: [ -> ]Don't we have the "No ETA questions" rules yet? If not, I would like to have one added.
Yes we sort of have it, at least the FAQ insists on that regard.
Wow, nice one! Thank you for this information!
Waiting =)
P.S. sorry for my english(
hassanrazza
06-07-2014, 10:00 PM
Sweet! A recompiler coming soon!!
derpf
06-08-2014, 06:34 AM
(06-07-2014, 03:57 PM)ssshadow Wrote: [ -> ](06-07-2014, 02:56 PM)AlexAltea Wrote: [ -> ]<[DH]> i have recompiler, but i try to optimize it and broke
<[DH]> i'll send recompiler at this month
(It's Xbyak based, btw)
Interesting. So is this a more advanced recompiler that not only directly translates ppc -> x86 assembler but also does advanced stuff such as eg loop optimizations for dat speed? (or whatever is appropriate, I am sure the the original binary is heavily optimized for the cell cpu, but different cache size or maybe even pipeline on x86 may mess up the optimizations done)
There are lots of optimizations to be made, cross-architecture. I imagine he means he tried to make the recompiler itself faster and ended up breaking it instead, though.
Now you guys can quit asking about it.
hassanrazza
06-08-2014, 01:03 PM
(06-08-2014, 06:34 AM)derpf Wrote: [ -> ](06-07-2014, 03:57 PM)ssshadow Wrote: [ -> ](06-07-2014, 02:56 PM)AlexAltea Wrote: [ -> ]<[DH]> i have recompiler, but i try to optimize it and broke
<[DH]> i'll send recompiler at this month
(It's Xbyak based, btw)
Interesting. So is this a more advanced recompiler that not only directly translates ppc -> x86 assembler but also does advanced stuff such as eg loop optimizations for dat speed? (or whatever is appropriate, I am sure the the original binary is heavily optimized for the cell cpu, but different cache size or maybe even pipeline on x86 may mess up the optimizations done)
There are lots of optimizations to be made, cross-architecture. I imagine he means he tried to make the recompiler itself faster and ended up breaking it instead, though.
Now you guys can quit asking about it.
True, but he also said he will be releasing the recompiler at the end of this month, so the excitement is there.
Ekaseo
06-08-2014, 10:18 PM
Hmm, what are the benefits of a Xbyak-based recompiler and what are the benefits of asmjit recompiler?