PowerMac G5 1.6GHz PSBench 7 Results

[MacRumors]

One Mac User has finally posted a set of benchmarks from PSBench7 using a PowerMac G5 1.6GHz with 1GB of RAM, using the Advanced (50MB) test file.

Raw numbers are availalble in this thread, and for comparison, a link to Ace's Hardware numbers are provided.

At table with all results is available at Chaosmint.com. Fastest numbers are in red.

[Kenndac]

Woah... While the G5 isn't the fastest in all tests (in fact, most of them) I'm quite impressed... seeing as this is only the 1.6 G5 vs 3+Ghz machines...

Can't wait to see Dual 2.0 results! :-)

[tizza]

Can't wait to see results with Panther - should be even better!!

[Freg3000]

Great results. A 2.0 Ghz machine would win about half of the tests (my rough estimation). A DUAL 2 Ghz PowerMac is going to be smokin'.

[zigzag]

Originally posted by tizza
Can't wait to see results with Panther - should be even better!!

Amen!

And photoshop 8!

[Macmaniac]

Very nice for a "slow" machine!! the 2 is going to be mad awsome!

[myrdred23]

This still doesn't say anything about the "strange" Xbench results. The test compares PCs to the new G5, which is not to say is not useful, it is very much, but doesn't draw a bottom online on G4 vs G5 performance.

Anyone with a MDD 2x1.42 care to run that test so we can have a COMPLETE set of results?

[marco114]

One thing i didn't like about the test is that it did start to use the disk at some periods. Once it does that, it really becomes an unfair test and does not show the true power of the processor.

You should allocate as much RAM as possible to Photoshop and not let it use disk cache. Otherwise, it will really slow down the results.. A 50GB file however with 300GB of RAM, it should really never use disk cache in my opinion... but.. oh well.

It does however show the speed in a basica real-life situation.. so it's not too bad.

[myrdred23]

Another thing to note, is although the G5 only wins a minority of the tests, if you actually look at how well each machine came, you would notice that the G5 scored SECOND out of the seven machines, with only the P4 3.06 HT beating it.

[Powerbook G5]

Well, considering this this the *low end* machine against the PC industry's *high end* machines, the 1.6 definitely holds its own and bests them in a few spots, so not bad. I can't wait to see a dual 2 GHz G5 running Panther in a rematch...then by Christmas/MWSF when they up the clock speeds, it will be even more amazing. Just imagine what the G5s will be able to do by this time next year.

[punter]

I made a graph of the results, but I'm afraid it didn't turn out as I had hoped.

What it does show is that (at this scale) the G5 1.6 is right on the money.

It also shows is that a lot of these tests are now millisecond shootouts, and that all the modern computers are easily quick enough for these tasks.

Finally it shows that I need to upgrade

*edit: I'd like to see the results with over 100mb files.
They'd probably just get even better!

[macrumors12345]

These scores look good. Keep in mind that PSBench calculates the overall score in a mindless way (summing the individual times) that puts a vast majority of the "total score" weight on just four tests: accented edges, pointillize, water color polar, and radial blur. Coincidentally (or maybe not...conspiracy theory?), these are the 4 tests in which the G5 does particularly poorly, so the fact that it beats the Athlon 2200+ in the overall score is actually quite impressive. Basically, you should think of the overall score as just consisting of those 4 benchmarks I mentioned.

A much better "overall" benchmark is to look at the number of tests each machine won (though this is still imperfect - in fact, there is no perfect benchmark in this case unless you know with what frequency you do each type of operation). Ignoring the lone DP machine, the P4/3.06 HT won the most tests (10) - no surprise there. The G5/1.6 won the second most (5), and the Athlon 3000+ won the fewest (4). I would not be surprised if overall the G5/1.6 could match the Athlon 3000+, though still fall somewhat short of the P4/3.06 HT (but then you are comparing bottom of the line Mac vs. top of the line PC, so what do you expect?).

[Stella]

I think the results are quite impressive, considering the 1.6Ghz is the very low end processor.

Be interesting to see the1.8 and especially the Dual 2Ghz beasty..

[RalphNumbers]

They say:

With Photoshop 7 running, we set 300-340 MB physical RAM memory (out of 512 MB) was used, so disk swapping did occur on occasion, though not often.

Anyyone see a good reason to set the RAM for photoshop so low?

And why does he say he has 1GB of RAM at the begining and only 512MB here?



I get the feeling the G5 would have scored somewhat better if it had a bit more ram allocated to photoshop.

[Stike]

With those scores at hand... would you like to imagine a 1.2 GHz G5 Powerbook? Would people still want it after seeing those scores?

[Mav451]

Originally posted by marco114
One thing i didn't like about the test is that it did start to use the disk at some periods. Once it does that, it really becomes an unfair test and does not show the true power of the processor.

You should allocate as much RAM as possible to Photoshop and not let it use disk cache. Otherwise, it will really slow down the results.. A 50GB file however with 300GB of RAM, it should really never use disk cache in my opinion... but.. oh well.

It does however show the speed in a basica real-life situation.. so it's not too bad.

This is a good point since "most" PShop users would have 1 gig open (i.e. not running warcraft 3 in the background while burning some cd's, while watching finding nemo lol)

HD I/O can vary greatly! But the Apple has SATA. I'm not sure the other test systems would had PATA's or SATAs or even SCSI's!!

Additionally, the 1.6G5 is tested with 1Gb, while the Ace hardware results are from 512, that's actually "300-340MB".

The motherboards that the Athlons were tested on aren't even mentioned--Nforce2 or Via? Via tends to underperform particularly in memory intensive applications and overall is considerably slower.

http://www.hardocp.com/article.html?art=MzY0LDI=
pay attention to the Content Creation Benchmark (yes the 2.8 is the old B model running 533fsb, but again the diff between via and nforce)

Regardless, the optimizations are amazing (the 3 test specifically: 10 Despeckle, RGB-CMYK, and Reduce Size 60%)...the PC's aren't even close--clearly some good tweaks there

[bousozoku]

I was pleasantly surprised to use PS7 on a 1.6 GHz G5 tonight.

This one did not have the G5 plug-in but was surprisingly quick. It was ready to go and I was still waiting for it. I made a layer copy of a 16 MB file and ran some filters on the copy, then reduced the opacity. All of it was so quick it made me wonder why my dual 800 is so slow.

[RalphNumbers]

nevermind...

[MrMacMan]

Yeah this is a decent result of the testing, the G5 scores okay for the single.

I want to see a single 2.0 vs. 3 GHZ P4.

That would be a good test.

[jettredmont]

Originally posted by RalphNumbers
They say:


Anyyone see a good reason to set the RAM for photoshop so low?

And why does he say he has 1GB of RAM at the begining and only 512MB here?



I get the feeling the G5 would have scored somewhat better if it had a bit more ram allocated to photoshop.

That note was for the PC tests, not the G5 test (he didn't say what memory allocations he used there). It is a quote from the Ace Hardware article.

[jettredmont]

Originally posted by Mav451
The motherboards that the Athlons were tested on aren't even mentioned--Nforce2 or Via? Via tends to underperform particularly in memory intensive applications and overall is considerably slower.

I despise Ace's layout ... makes it nearly impossible to find the "beginning" of an article you're reading. However, if you follow the links from the front for that article (click the "Barton: 512 KB Athlon XP Reviewed" link on the right), you will come across this page, which gives the specifics on the PC configurations:

http://www.aceshardware.com/read.jsp?id=50000366

nForce 2 was the Athlon's chipset. Granite Bay was the P4 chipset.

[AidenShaw]

Originally posted by Mav451
But the Apple has SATA. I'm not sure the other test systems would had PATA's or SATAs or even SCSI's!!

With a single disk, there would be very little difference between a 7200 RPM SATA disk and a 7200 RPM PATA disk.

The difference between 150 MB/sec and 133 MB/sec is noise when the disk's transfer rate is around 50 MB/sec.

SCSI would be about the same - the disks are usually a bit faster, but SCSI has more command overhead.

SATA's big advantages are the smaller cables, and the fact that no drive is a slave. A two-disk master/slave PATA setup can really get bogged down. SCSI and SATA handle multiple drives better.

[Catfish_Man]

It's weird how the G5 *crushes* the competition in a few tests, and gets absolutely squashed in others. I wonder why that is.

[jettredmont]

Not to rain on the parade, but ...

The P4 configurations tested, using the Granite Bay chipset, are far from top-of-the-line today. They were current in Feb, 2003, not August 2003!

In particular, the FSB was significantly slower for those P4's than one you would buy today (today = Canterwood/800MHz FSB, not Granite Bay/533MHz).

So, as always, take benchmarks, especially against non-contemporaries, with a grain of salt.

Ref: Dell has a P4/3.06/800MHz bus with a workstation video card for $2000 after rebate:

Dell Precision™ Workstation 360 Desktop
Intel® Pentium® 4 Processor, 3.00GHz, 512K / 800 Front
Side Bus
Qty: 1
Price: $2,160.00

(rebate is $150)

[AidenShaw]

Originally posted by Catfish_Man
It's weird how the G5 *crushes* the competition in a few tests, and gets absolutely squashed in others. I wonder why that is.

And I wonder which tests were used in the Stevenote