|
and if you want to learn 3d modelling there are some really neat paleolithic cave paintings in the south of france you ought to check out
|
# ¿ May 4, 2016 11:15 |
|
|
# ¿ May 23, 2024 12:22 |
|
learning how a cpu is made is fine and all, but for the purposes of understanding the tradeoffs between two relatively closely related cpu implementations it is a very very long route to take. also, really, i don't think the information necessary to *really* understand the performance profile tradeoffs in e.g. skylake or twister or such is even at all publically available obsessing over pointless isa details is notably popular, being a part that is actually available for public viewing, but it really is of rather limited actual relevance Cybernetic Vermin fucked around with this message at 21:19 on May 4, 2016 |
# ¿ May 4, 2016 21:16 |
|
the stuff that could be done in parallel could already be done in parallel even without bringing a billion cores onto the same die, a lot of stuff cannot be done in parallel (most stuff in fact, in a theoretically sound but practically useless statement), and then only peak single-thread performance matters, and intel remains the only game in town there it is somewhat interesting to ponder the end-game for single-thread peak performance, since it will never stop mattering, but improving it further appears difficult in all directions
|
# ¿ May 9, 2016 09:40 |