You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Sounds reasonable. Re BFP, constructing a new packed value from components very rarely actually requires any masking. C programmers just write a<<n1 | b<<n2 | c<<n3 and ensure that works.
When C programmers write code like that, yeah. But consider something like struct { int a : 5; int b : 12; int c : 15; } s; s.b += foobar; Usually the compiler knows very little about the upper bits of s.b + foobar. For example, it may well be a negative value.pic.twitter.com/VbcbdXerSN
You use six instructions, two of them in B, and you got lucky with the li .. another struct or field might need lui/addi. You can do it with seven instructions from the base instruction set: s >> 27 << 27 | (s << 5) + (foobar << 20) >> 5
Oh, I didn't get lucky with the LI. The BFP instruction is designed so that you can always use ADDI here, never need LUI+ADDI.
Let's forget the +=. Simply doing s.b = foobar can be done with: (s << 17 >> 12 | foobar << 20) >> 5 | s >> 27 << 27 Which is 8 instructions. I guess your pack/bfp reduces that to 4, which is worthwhile. [oops I'm packing the fields in the opposite direction .. but its undef]
JFYI, the packing order is defined in the RISC-V calling convention.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
