Date: Thu, 23 Mar 2000 11:17:06 -0500 (EST) From: Andrew Gallatin <gallatin@cs.duke.edu> To: Ed Hall <edhall@screech.weirdnoise.com> Cc: obrien@FreeBSD.ORG, freebsd-alpha@FreeBSD.ORG Subject: Re: Compiler problems with -O2 (was Re: CVS Trouble, even under 4.0-RELEASE (alpha) HELP!) Message-ID: <14554.16893.126023.434069@grasshopper.cs.duke.edu> In-Reply-To: <14554.16137.955904.783199@grasshopper.cs.duke.edu> References: <gallatin@cs.duke.edu> <14553.19348.115781.273817@grasshopper.cs.duke.edu> <200003230820.AAA12969@screech.weirdnoise.com> <14554.16137.955904.783199@grasshopper.cs.duke.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Andrew Gallatin writes: > > (-stable and howardl@account.abs.net trimmed from the CC lines) > > Ed Hall writes: > > Andrew Gallatin writes: > > : I take it the O2 bugs are not unique to us, but rather they are > > : generic across all OSes that gcc version 2.95.2 runs on? Do the gcc > > : people know these problems exist? > > > > Just FYI, the Linux kernel is compiled with: > > > > -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -mno-fp-regs \ > > -ffixed-8 -fno-strict-aliasing > > > > on Alpha, and seems quite stable with these flags. Yes, this is with > > GCC 2.95.2. > > Any idea if the linux compiler uses the HAIFA optimizer? Perhaps that > is our problem. > > > On applications, I've seen gains of about 20% for the higher levels of > > optimization (for example, CSound processes a particular piece in 79 > > seconds when compiled with just -O, and 65 seconds with -O2). This is > > a useful improvement, though one that has to be balanced with the risks. > > To me, at least, this is not worth the risk. The Tru64 Compaq C > compiler & libraries have sped up a collegue's floating-point intensive > matrix manipulation code by a factor of 4. That's the [Oops.. Dropped my coffee mug on the right combo of keys to send the message prematurly. To continue:] To me, at least, this is not worth the risk. The Tru64 Compaq C compiler & libraries have sped up a collegue's floating-point intensive matrix manipulation code by a factor of 4. That's the sort of optimization level that I'd be willing to find & fix inherent bugs for, not just a piddling 20%. I really wish Compaq would introduce a native FreeBSD version of their compilers. My main motivation for suggesting temporarily disabling anything more than -O is that there are a lot of innocent people out there that expect the compiler to work. Given that it is horribly broken, they're going to be wasting a lot of time debugging their apps for no good reason. And they're going to be writing off FreeBSD as unstable, etc. I'd like to prevent that. Cheers, Drew ------------------------------------------------------------------------------ Andrew Gallatin, Sr Systems Programmer http://www.cs.duke.edu/~gallatin Duke University Email: gallatin@cs.duke.edu Department of Computer Science Phone: (919) 660-6590 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14554.16893.126023.434069>