Date: Mon, 20 Feb 2012 13:50:39 -0500 From: Super Bisquit <superbisquit@gmail.com> To: "Kevin H. Patterson" <kpatterson.home@gmail.com> Cc: FreeBSD PowerPC ML <freebsd-ppc@freebsd.org> Subject: Re: FreeBSD 9.0-RELEASE on PowerMac Dual G5 Message-ID: <CA%2BWntOsjc6X329ujtxwgJP1=JceVu__=khntX9F2cVokKZq6eg@mail.gmail.com> In-Reply-To: <8830CCB0-1417-4678-A766-151080BF7796@khptech.com> References: <3934AD65-E01C-4DDD-8BDC-F52C6AE3655F@khptech.com> <CA%2BWntOtKvc6HMDoMNQj6AQtiA5kOXGrZqcnFcssuGs6nv_EimQ@mail.gmail.com> <4D7B47E5-6A33-4426-8186-D5E9D5C9B32E@khptech.com> <CA%2BWntOsYiKaygJBg_iQBkZ9YX1JjGF0_MnbzJwHqOxyK2uBnYg@mail.gmail.com> <8830CCB0-1417-4678-A766-151080BF7796@khptech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
. Try cd /usr/src and look at sys/dev at the codes there <- My opinion and the way I recently started looking at hardware errors, those mentioned in the list above arew better at it. On 2/19/12, Kevin H. Patterson <kpatterson.home@gmail.com> wrote: > I am interested in raw cpu integer and floating-point performance only, at > this point. This is *exactly* what simple synthetic benchmarks (like ubench) > are for. There is absolutely NO reason why performance should be 3x better > running under macosx than under freebsd, on the exact same hardware. Things happen that way. I've ran a firefox benchmark on different architecture running different systems and received different results. > > FWIW, I ran these benchmarks repeatedly, as root, after a clean boot, with > no window manager or other processes running. 'top' shows the cpus maxed out > on both macosx and freebsd. Under macosx, on the other hand, with the full > macosx GUI et. al. running, the numbers are about 3x better. > There's also the fact that Apple concentrated its efforts on the Power(PC) architecture. FreeBSD is not a single instance system- by this I mean an all-in-one hardware and software- such as MacOSX or Windows. > I have also run a number of other benchmarks, as well as compiling various > things like x11/gnome2, etc. It is all about 2-3x slower to build under > freebsd. > > At this point the strongest evidence points to the G5 clock / bus speeds not > being set up properly under freebsd. I am wondering who might have more info > on checking this out. Probably Whitehorn, HIbbits, and Grehan > > On Feb 19, 2012, at 1:58 AM, Super Bisquit wrote: > >> How about real world benchmarks of compiling and running programs? >> You're going to get different benchmarks on different operating >> systems on the same machine. OpenBSD runs better than FreeBSD or Linux >> but you need more memory and a processor of 500MHz or greater on the >> PowerPC architecture. Everything affects the outcome. >> >> If there is no window manager, the response is better. If the kernel >> and base system is not "testing," the response is better. Et al.... >> >> You'd be better off comparing the performance of two machines running >> a real world application such as S4P. >> >> >> >> On 2/19/12, Kevin H. Patterson <kpatterson.home@gmail.com> wrote: >>> I realize ubench is a far cry from comprehensive benchmarks. However, >>> although the absolute numbers are meaningless, it *is* designed to >>> compare >>> systems directly with each other. >>> >>> There is no good reason I can think of why the *exact same code* for >>> ubench >>> under macosx 10.5 and ubench under freebsd 9.0 would show a relative >>> speed >>> of 1/3 when running on freebsd. uBench is a simple program, written in C. >>> While the calculations involved might be senseless, there is, again, no >>> good >>> reason why freebsd should run this "senseless" code at 1/3 of the speed >>> that >>> it runs under macosx. >>> >>> This is not my imagination; I'd be happy to compare some other benchmarks >>> if >>> you like. >>> >>> On Feb 19, 2012, at 12:15 AM, Super Bisquit wrote: >>> >>>> http://www.freebsd.org/cgi/url.cgi?ports/benchmarks/ubench/pkg-descr >>>> >>>> Notice that the description uses the word "senseless" more than once. >>>> >>>> Try compiling a program or a set of programs. On a low end B&W, I was >>>> able to compile firefox plus a few other applications. POWER/PowerPC >>>> is known for performance. >>>> The architecture is used in gaming systems for reason that it can pass >>>> instructions through once they are "learned." This may be a simplistic >>>> explanation of load-store but it is one you should be aware of. >>>> >>>> Both the Power and Power64 releases are Tier 2 and are worked on by a >>>> small group within the FreeBSD community. >>>> >>>> You can change the flags of make in /etc/make.conf if you wish. >>>> >>>> On 2/18/12, Kevin H. Patterson <kpatterson.home@gmail.com> wrote: >>>>> Hello, >>>>> >>>>> I've taken an interest lately in running FreeBSD on the powerpc64 >>>>> architecture. I have access to a dual 2.5 GHz PowerMac G5, and I've >>>>> successfully got FreeBSD 9.0-RELEASE up and running on it. >>>>> >>>>> Only one thing seems amiss so far... it feels *very* SLOW. I realize >>>>> this >>>>> is >>>>> an older machine, but it feels much too slow for a dual G5. Compiling >>>>> seems >>>>> to take forever, and top shows ~50% or more "system" CPU usage when >>>>> doing >>>>> almost anything other than sitting idle. Furthermore, the system fans >>>>> never >>>>> speed up, but run at the lowest speed even when the system is under >>>>> full >>>>> load. I have tried both enabling and disabling powerd support, with no >>>>> effect. >>>>> >>>>> For a quick sanity check, I installed ubench (0.32) from ports. The >>>>> numbers >>>>> were quite disappointing: 109870 CPU / 50527 MEM multiprocessor, and >>>>> 55433 >>>>> CPU / 30863 MEM single-processor. >>>>> >>>>> For comparison, I ran ubench (0.32 from MacPorts) under Mac OS X 10.5.8 >>>>> on >>>>> the same machine. This time, the fans do ramp up, and the numbers are >>>>> *WAY* >>>>> better: 277207 CPU / 317119 MEM multi-processor, and 141021 CPU / >>>>> 284113 >>>>> MEM >>>>> single-processor. >>>>> >>>>> As you can see, all is not well. I am wondering what is slowing FreeBSD >>>>> down >>>>> on this machine. I have tried both GENERIC and my own kernel config. It >>>>> feels like the CPU and or bus speed is clocked down perhaps to the most >>>>> energy-saving level. Maybe this is where openfirmware leaves it after >>>>> boot? >>>>> Also interesting is to note the drastic *single-processor* ubench >>>>> difference >>>>> between macosx and freebsd. To me this looks like a low clock-speed >>>>> smoking >>>>> gun. >>>>> >>>>> I also noticed that the kernel build includes flags like -msoft-float >>>>> and >>>>> -mno-altivec... >>>>> >>>>> I am interested in any build or config tweaks that might be in order. I >>>>> am >>>>> also more than happy to debug and get to the bottom of this. Any ideas? >>>>> >>>>> Sincerely, >>>>> >>>>> Kevin H. Patterson >>>>> KHPtech >>>>> >>>>> kevpatt@khptech.com >>>>> >>>>> _______________________________________________ >>>>> freebsd-ppc@freebsd.org mailing list >>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-ppc >>>>> To unsubscribe, send any mail to "freebsd-ppc-unsubscribe@freebsd.org" >>>>> >>> >>> > >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CA%2BWntOsjc6X329ujtxwgJP1=JceVu__=khntX9F2cVokKZq6eg>