Date: Tue, 29 May 2001 11:18:44 +0200 (CEST) From: Riccardo.Veraldi@fi.infn.it To: Yoriaki FUJIMORI <fujimori@grafin.fujimori.cache.waseda.ac.jp> Cc: <freebsd-alpha@freebsd.org> Subject: Re: tscp MIPS result really weird Message-ID: <Pine.NEB.4.33.0105291105240.1973-100000@nikita.fi.infn.it> In-Reply-To: <200105290758.QAA01412@grafin.fujimori.cache.waseda.ac.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi this is my case. I run again the compiler with proper options: DPW 500A (MIATA) 500MHz ev56 2MB BCACHE 96KB ACACHE 256MB RAM FreeBSD 4.3 GENERIC GCC: cc -O3 -funroll-loops -fexpensive-optimizations -fschedule-insns2 *.c -o tsc.gcc *.o -lcompat Nodes: 550822 Best time: 8335 ms Nodes per second: 66085 (449.2 MIPS) this with a few processes running on the system. The same benchmark when compiling in /usr/ports/x11/XFree86-4 gives me: Nodes: 550822 Best time: 11959 ms Nodes per second: 46059 (313.1 MIPS) the same benchmark in single user mode gives me 500MIPS Is it normal this big difference ?? I guess anyway the GCC is not as well optimized as COMPAQ cc. Anyone might run this test on his own Miata too ?? thanks Rick On Tue, 29 May 2001, Yoriaki FUJIMORI wrote: > Date: Tue, 29 May 2001 16:58:26 +0000 > From: Yoriaki FUJIMORI <fujimori@grafin.fujimori.cache.waseda.ac.jp> > To: Riccardo.Veraldi@fi.infn.it > Cc: freebsd-alpha@freebsd.org > Subject: Re: tscp MIPS result really weird > > I do not know the exact spec of your machine and how it is configured. > But, I will show you my result: > > (1) AlphaPC164 500MHz(2MB)/512MB [FreeBSD4.3] > (2) PC164LX 600MHz(4MB)/1024MB [Tru64 5.0] > (3) UP1100 600MHz(2MB)/768MB [FreeBSD4.2] > > (A) ccc -tune ev56 -O4 -fast -ifo *.c (Compaq ccc on linux-emulator) > (1) 550822, 7143ms, 77113 (524 MIPS) > (3) 550822, 4309ms, 127830 (869 MIPS) > (B) ccc -tune ev6 -O4 -fast -ifo *.c (Compaq ccc on linux-emulator) > (3) 550822, 4159ms, 132240 (900 MIPS) > (C) cc -nointrinsics -tune ev56 -O3 -ifo *.c (Compaq cc on tru64) > (2) 550822, 5554ms, 99175 (674.2 MIPS) > (3) 550822, 4354ms, 126509 (860 MIPS) > > I guess the amount of B-cache matters with respect to this kind of > small scale benchmarks. > > Yoriaki Fujimori > 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?Pine.NEB.4.33.0105291105240.1973-100000>