Defining the 128bit ints in gcc

white rabbit · Post by **white rabbit** » Sun Jul 10, 2005 4:21 am

I am trying to get the psp-gcc to accept the fact the PSP has 64-bit architecture, and can handle 128 bit integers using the __attribute__((mode(TI))) code, but the compiler is having none of it, saying it cannot emulate the TI mode.

This is due to me suspecting some data types are of this format, and aren't two (u)int64s bundled together - as that causes some... oddness.

Has anybody got a workaround for this, or got it working? I'm guessing it's some gcc switch, but I'm not very good with the MIPS architecture, or gcc switches.

Post by **mrbrown** » Sun Jul 10, 2005 4:23 am

I know about it, I haven't had the chance to fix it yet. BTW, the PSP is a 32-bit architecture (unless I misread what you were trying to say).

white rabbit · Post by **white rabbit** » Sun Jul 10, 2005 4:29 am

mrbrown wrote:I know about it, I haven't had the chance to fix it yet. BTW, the PSP is a 32-bit architecture (unless I misread what you were trying to say).

Nope you read right - I thought it was 64bit. Now feel pretty dumb :-S

Jst to make sure I'm not being uber-dumb. the int space is 32 bits? So longs are 64bit, and long longs are 128bit? If so, then maybe my suspicions of the 128bit numbers are just wrong (internal size maps of vectors and matricies).

ector · Post by **ector** » Sun Jul 10, 2005 4:44 am

long is also 32-bit, while long long is (emulated) 64-bit.

white rabbit · Post by **white rabbit** » Tue Jul 12, 2005 10:07 pm

:blush: right, it seems to be ok now I'm working in 32 bits rather than trying to get it to use 64.

Thanks!