More G5 Benchmarks (NASA)

[blackhorse95]

pc reality check

I have only recently joined the ranks of Mac owners. I did this because I wanted to see for myself if all the smug bragging that I had seen and heard from Mac users was founded in reality. I had always assumed that it was just a huge load of B.S. Admittedly, I started out my mac experience at the lower end of the performance pool; a G4 400 mhz. However, as I begin to learn the workings of the mac architecture and OS X, I finally feel almost qualified to post some unbiased remarks concerning the performance of mac hardware vs. x86 pc hardware.

I also read the nasa performance ratings that are mentioned here; however, I did not get the same impression that most of the previous posters have. The bottom line that I got from the nasa benchmarking was that the 2 GHz G5 was roughly equilavent to a P4 2.6 GHz. I was somewhat surprised at this, since before this I had been seeing Apple G5 advertisements claiming the G5 was the fastest pc on the planet. I had assumed that the G5 was actually much faster than it tested out by NASA.

Anyway, what we all need to do is to quit speculating and assuming results that have not been independently and fairly tested. G5 processors MAY show higher performance when benchmarks are recompiled for this architecture, but until we actually see this demonstrated, it is extremely arrogant to assume so. Also, do not always make assumptions that place higher performance on Mac hardware and lower performance on X86 hardware. It is assumed here that X86 machines perform better running linux than Windows XP. Why would anyone assume this? While linux is at its core a unix derrivative, as is OS X; the comparison is hardly apples to apples. Pun intended. If we wanted to have a fair test between a P4 and a G5, we would use linux as the OS for BOTH machines. Using OS X gives a clear advantage to the G5. Why not see the results of the P4 running XP? I hope everyone here realizes that the Apple G5 tests and the associated publicity was entirely a marketing tool used by Apple to promote the G5. Do not get me wrong, I applaud Apple for the G5. I believe it is a great piece of hardware. But realistically I believe the hardware performance results to be pretty close to what was reported in the NASA test.

It seems like many Mac fans are obsessed with patting themselves on the back and socializing in a mutual appreciation society. I find this entirely similar to pro sports fans who watch their teams play on TV and say to each other how "we" sure kicked their ass Monday night!!!

I hope to continue to learn more about my Mac and OS X unix. I doubt that I will ever abandon my X86 systems in favor of them. So far, I see a nice platform with no real advantage. I would encourage everyone to take a first hand look at the "other team" before dismissing it entirely.

 

[legion]

Totally agree with you. I find the "self congratulatory" attitude to be a hindrance to the Apple/Mac platform realizing its potentional... and that's just it; at the heart of the Apple/Mac machine is tremendous potentional, it's just poorly realized and poorly marketed. If most people involved with Macs would get beyond the "whining" stage of anti-Mac consipiracies and "we're the best at ..." mentality and instead just produce great results, the platform and its merits would speak for itself.

Also, realizing that computers are tools (and not surrogate mothers, fathers, lovers, children, and best friends), there would be the epiphany that there are different tools for different jobs and that not one is inherently bad.

Some how along the way, these things have been lost and generations have been brainwashed by marketing departments.

 

[DrBoar]

With IBM linux blade servers soon using both IBM 970 AKA G5 and Intel Xeon we will have more valid comparisons of how well the G5 stacks up against the competition.

There still is the problem of bad programing and compiling but here there are less vested interests. I do not think anyone expect Quiktime to perform better in Windows than OS X or some applications ported from windows to run better in X than windows. Linux ought to be more "agnostic".

And the bottom line is if it feels fast it is fast. With small subsets of tests anything can be proven Client speed RC5-72 G4 1.25 MHz 13millon keys/s a Pentium 4 at 3.2 GHz only score 4.5 million and the AMD Opteron 8 million. This only proves that the G4 is way faster than any Intel CPU in this particular applcation thanks to AltiVec. In most application the Intel stomp on the G4.

 

[bousozoku]

Considering some of the idiots who work for NASA here in Floriduh, it's a wonder they get anything to compile at all. I think they're just glad to see it work and don't worry a whole lot about optimisation. That's where Apple's version of gcc, prior to version 3.3 has always broken down. It was terrible at optimising code and the PowerPC line is not good at handling bad code.

That said, I don't wish to try the "other team" at home. I've used it extensively for work at various places. Even the latest Windows XP is more stressful than Mac OS X 10.2 and above. With Mac OS X, though, many developers have had to discard their old "truths" because they just don't work anymore. I've seen several products which have become much slower on Mac OS X because they thought that they could simply add a Carbon resource file and re-compile, as they've been told could work. Of course, they didn't read through Apple's skimpy documentation to find that it doesn't work well to do that.

I'm sure Apple people working with Adobe and Macromedia find themselves frustrated each time.

At least, the new developers don't have any System 6 legacy to drag Mac OS X down.