sakman_ has quit [Remote host closed the connection]
Maylay has quit [Quit: Pipe Terminated]
brazuca has quit [Quit: Client closed]
sakman has joined #riscv
dlan has joined #riscv
Bluefoxicy has quit [Read error: Connection reset by peer]
Bluefoxicy has joined #riscv
handsome_feng has joined #riscv
jacklsw has joined #riscv
jacklsw has quit [Client Quit]
terminalpusher has joined #riscv
terminalpusher has quit [Remote host closed the connection]
jacklsw has joined #riscv
pavelow has quit [Quit: No Ping reply in 180 seconds.]
pavelow has joined #riscv
heat has quit [Ping timeout: 245 seconds]
ln5 has quit [Ping timeout: 252 seconds]
meta-coder has quit [Ping timeout: 244 seconds]
<sorear>
does -ffast-math actually authorize transformations that assume floats are less than INT32_MAX and doubles are less than INT64_MAX? (ceil on gcc-patches)
Leopold_ has quit [Remote host closed the connection]
Leopold has joined #riscv
<dh`>
I always assume that -ffast-math is incorrect and should never be used
<dh`>
nothing about it would surprise me and I kind of thing it ought to just be implemented by replacing all fp ops with raise(SIGFPE)
<dh`>
s/thing/think/
<sorear>
do you feel the same way about incompletely parenthesized expressions in fortran
meta-coder has joined #riscv
davidlt has joined #riscv
sakman has quit [Remote host closed the connection]
KREYREN_ has quit [*.net *.split]
rvalles has quit [Ping timeout: 240 seconds]
matoro has quit [Quit: No Ping reply in 180 seconds.]
matoro has joined #riscv
BootLayer has joined #riscv
foxbat has quit [Remote host closed the connection]
peeps[zen] has quit [Read error: Connection reset by peer]
peeps[zen] has joined #riscv
<knielsen>
sorear: It would seem so from the documentation of -funsafe-math-optimizations ("assume ... results are valid"). If you are talking about something like (int)ceil(x).
<knielsen>
sorear: But not for double n = ceil(x); obviously, there's no requirement that ceil() by itself fits in an int
<sorear>
as far as I can tell from reading the patch, it applies to ceil() and ceilf() without a preceding cast, although my track record with gcc code is fairly bad
handsome_feng has quit [Quit: Connection closed for inactivity]
<knielsen>
sorear: agree, that looks strange
billchenchina has quit [Ping timeout: 264 seconds]
billchenchina- has joined #riscv
foxbat has joined #riscv
foxbat has quit [Client Quit]
foxbat has joined #riscv
meta-coder has quit [Ping timeout: 264 seconds]
davidlt has quit [Ping timeout: 260 seconds]
handsome_feng has joined #riscv
billchenchina- has quit [Remote host closed the connection]
billchenchina has joined #riscv
billchenchina has quit [Remote host closed the connection]
grumbler has quit [Quit: It's time]
crabbedhaloablut has joined #riscv
davidlt has joined #riscv
alexghiti has joined #riscv
grumbler has joined #riscv
grumbler has quit [Quit: It's time]
MaxGanzII has joined #riscv
enthusi has joined #riscv
MaxGanzII has quit [Ping timeout: 252 seconds]
KREYREN has joined #riscv
xypron has quit [Ping timeout: 245 seconds]
crabbedhaloablut has quit [Read error: Connection reset by peer]
crabbedhaloablut has joined #riscv
markh has quit [Ping timeout: 240 seconds]
markh has joined #riscv
meta-coder has joined #riscv
meta-coder has quit [Client Quit]
grumbler has joined #riscv
grumbler has quit [Remote host closed the connection]
loki_val has joined #riscv
crabbedhaloablut has quit [Ping timeout: 240 seconds]
<knielsen>
sorear: Of course, if ceil(x) _doesn't_ fit in 32/64 bit, then x is already an integer. But like you, I don't see anything in the gcc patch that would skip the float->int->float conversion in this case