Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

PowerMac G5 1.6GHz PSBench 7 Results

Re: Re: Overall times

Originally posted by macrumors12345
Please do not quote aggregate PSBench scores as a performance indicator. As I explained earlier, the aggregate scores are VERY heavily weighted towards just four filters, so they are not at all indicative of overall performance. It is like quoting RC5 scores to show how "fast" the G4 is compared to the P4 - it's just a single benchmark, so it's not meaningful.

If you want to scale, you should reduce the time of each benchmark by 20% and compare the two machines. I calculate that a SINGLE G5/2.0 would win in 9 or 10 of the 21 benchmarks then versus the P4/3.06 HT, so it would just fall short in PSBench. But don't worry, it will do even better in more comprehensive PS benchmarks.
Click to expand...

When have I ever said I was going to use these as a performance indicator?

Whether or not those filters play a big role in the overall time is irrelevent. You simply can't choose to "see" what a particular CPU is faster at and what that CPU is slower at without considering the extent. Your win-lose perspective is too black and white to accurately compare either cpu, if the P4 beats the G5 slightly, it's considered a "win", if the G5 beats the P4 significantly, it's also considered a "win", there is nothing in the details and thats what I think the overall score offers. Just because alot of the time is spent on those four specific filters doesn't refute the validility of the time it took to actually finish with everything!

Now, you can argue that this benchmark as a whole doesn't represent a typical workload in Photoshop, that it consists of four "very heavily weighted" that probably aren't going to be seen in a typical workload but again thats not going to invalidate the overall score.
 
Originally posted by hvfsl
In the original article it said that the 1.6Ghz G5 is about the same as a Athlon XP2400. Well the Athlon XP 2400 is clocked at around 1.7Ghz with a 266Mhz bus, so this does not look so good for the G5.
Click to expand...

No,the 2400 is clocked at 1.93 ghz.
 
Re: 1.42 results

Originally posted by synthetickittie
Here are the results from my 1.42ghz dual g4 powermac (fw800), 2gb of ram, GeForce4 Ti 128mbs, running 10.2.6.

90 Degree Rotate-1.0
9 Degree Rotate-3.6
.9 Degree Rotate-7.4
1 Gaussian-5.5
3.7 Gaussian-6.7
85 Gaussian-8.1
Unsharp 50/1/0-5.3
Unsharp 50/3.7/0-10.1
Unsharp 50/10/5-7.9
Despeckle-5.5
RGB-CMYK-7.6
60% Reduction-2.3
LensFlare-9.9
Color Halftone-11.2
NTSC Color-10.8
Accent Edge-18.2
Pointilize-19.6
Watercolor-37.8
Polar Coordinates-8.4
Radial Blur-32.6
Lighting-7.6
Click to expand...

There is something bad wrong with these scores.It is slower than a single 1.33 ghz G4 upgrade card. Check this out:
http://www.xlr8yourmac.com/G4CARDS/giga_designs_125Ghz/index3.html

there are some psbench scores on the page above.

Beware of making bold assumptions based on ANY benchmark.I have yet to see one I would put faith in. In this case remember just hom many different filters and actions there are in photoshop-and many parameters can be changed for each.Choosing a different set of operations and a different file size will give vastly different results.So dont waste time saying the G5 looks to be as fast as some so and so clocked x86 processor based on this or any other bench. Such a statement will only apply to that small particular test and real world useage will say something different.
 
It's not low end vs. high end

While I realize that 1.6GHz is the low end of the G5 chip, the 1.6GHz PowerMac G5 isn't a "low end" computer and it's not being compared to a "high end" P4 based computer. What makes a complete package high or low is its price. The AthlonXP 2400+ is definitely low end (roughly the price of an eMac) while the 3GHz P4 apparently costs the same as the 1.6GHz G5 making them both mid-range. The high end would be a dual Xeon versus a dual G5. We love our Macs, but let's be fair in our comparisons.
 
Actually, the 2400+ is the last of the 266 fsb processors in the thoroughbred B line.

So, with that in mind, the 2400+ is rated at 2ghz. To be more exact, with the 133 fsb, mulitplied by 15, it is around 1995mhz, + or -.

Hardocp did a OC review on the processor back in september:
http://hardocp.com/article.html?art=MzUw

Now the 2500+ barton, on the other hand, is clocked at 1.83ghz. That may be where the confusion is. Additionally, the 2000+ model palomino is clocked at 1.67ghz.
 
Re: Re: 1.42 results

Originally posted by moped
-----------------------------------------------
Synthetickittie,

I don't know what happened to your 1.42 dual but my 1.25 dual
gave me the following result.

90 Degree Rotate-0.9
9 Degree Rotate-2.8
.9 Degree Rotate-2.6
1 Gaussian-0.7
3.7 Gaussian-2.3
85 Gaussian-3.3
Unsharp 50/1/0-1.2
Unsharp 50/3.7/0-2.7
Unsharp 50/10/5-3.2
Despeckle-0.8
RGB-CMYK-3.5
60% Reduction-0.6
LensFlare-5.9
Color Halftone-6.2
NTSC Color-7.3
Accent Edge-22.4
Pointilize-19.3
Watercolor-48.1
Polar Coordinates-4.8
Radial Blur-34.5
Lighting-3.3
Click to expand...


It's nice to know that my G4 800 DP is the fastest mac to rotate 90°. :)

This are my result on PPCG4 800DP:

Rotate 90 0,8
Rotate 9 9,7
Rotate .9 8,8
Guassian 1 4,6
Guassian 3.73 7,9
Guassian 85 10,3
Unsharp 1 6,6
Unsharp 2 9,5
Unsharp 3 10
Despeckle 6,5
RGB CMY 8,9
Reduce 50 3
Lens 13
Color 14,4
NTSC 15,4
Accent 31,3
Pointillize 30
Water 67,2
Polar 11,9
Radial 52,5
Light 10
 
Re: It's not low end vs. high end

Originally posted by Bregalad
While I realize that 1.6GHz is the low end of the G5 chip, the 1.6GHz PowerMac G5 isn't a "low end" computer and it's not being compared to a "high end" P4 based computer. What makes a complete package high or low is its price. The AthlonXP 2400+ is definitely low end (roughly the price of an eMac) while the 3GHz P4 apparently costs the same as the 1.6GHz G5 making them both mid-range. The high end would be a dual Xeon versus a dual G5. We love our Macs, but let's be fair in our comparisons.
Click to expand...

I agree but most people miss out on the fact that the Mac comes with far better apps and the full version of the os, which matters pricewise. But Macs also are very reliable, use better quality hardware and are easy to work on, something that doesn't show up in the numbers game. I also think OSX is the best os, in my opinion of course:D
daniel
 
Re: Re: One big note ...

Originally posted by theRebel
The price that I came up with was $2261 before the rebate and that did not include any FireWire and only had a basic CD-RW drive (no DVD capabilities at all; no read and no write). What did you take off the config in order to lower the price?
Click to expand...

I looked back over my "Dell Config" and found that I had selected slower memory (512MB of DDR333, not DDR400 ... what a crappy ordering interface!). Still, changing to DDR400 brought it up to $2210 before rebate ... I'm not sure where your other $51 came from; I just used the default options except for CPU and memory going in (using the "Small Business" side of Dell's store, selected the left-most "Workstation" model to start from ...)


Also you should be aware that you can buy a 1.8Ghz G5 PowerMac with a CD-RW/DVD Combo drive for $2199 from the Apple Store. Thus, when you figure in that Dell also charges at least $75 for shipping (whereas the PowerMac ships for free), your Dell Workstation costs more than the 1.8Ghz PowerMac.
Click to expand...

Absolutely. I wasn't hoping to begin a "Dell is cheaper than Apple" contest. Note first that the Dell only had 512MB of RAM, a smaller HD by half, and CD-RW only (although they give you the option of nixing that for $79 back ... which I personally would take since I've got four computers with DVD burners in my office already ...)

On the other hand, it's hard to compare memory quantitatively between Dell and Apple as OSX generally requires much more memory to operate than Windows (due to its windowing server ... which has nice benefits; I'm not complaining! But still, a 512MB P4 system is relatively spacious for development work whereas a 512MB OSX machine is constantly running out of physical RAM ...)

Were I ordering a 1.8GHz G5 I would be ordering one with at least 1GB RAM, which adds $250 to the price.

On the other hand, ordering the Dell, I'd have gone for a larger HD (80GB on the Dell by default, not 160GB as on the G5) ... which would add $100 (160 isn't an option with Dell; only 120GB ... why? No idea!).

And, no, I can't seem to find the OSX option on the Dell web site either. You might have to call in and request that special :)
 
Re: Re: Put "bottom of the line" in context

Originally posted by theRebel
How do you figure the single 1.6Ghz G5 PowerMac with a slower bus, slower RAM, etc, to be 80% of the speed of a Dual 2Ghz PowerMac?
Click to expand...

If you kept a single proc, I could see his point.

Everything in the 1.6GHz system is 20% slower than a theoretical single 2.0GHz G5 machine. The buses and memory are scaled (more or less) linearly with the CPU speed. Granted, PCI vs PCI-X might make a difference, but not on these benchmarks. On the other hand, HDD access is likely to be just as fast on the 1.6GHz G5 as on the dual 2.0's.

However, you are quite correct: throw in dual CPUs each with their own FSB, and ... well, we'll have to see those particular benchmarks when they come out.

BTW, the difference between February's P4 and today's P4 is about 5-10% generally on benchmarks. I expect a significantly larger benchmarking difference between the 1.6 and 2x2.0 GHz G5's!
 
Re: Re: Re: 1.42 results

Originally posted by macrumors12345
That said, the fact that these numbers are varying so wildly between similar machines throws into question whether this is even a remotely reliable benchmark.
Click to expand...

The numbers vary wildly between multiple runs of the SAME machine! You have to run the benchmark 5-10 times, throw out outliers, and average the rest to get reproducible and meaningful results!
 
Originally posted by theRebel
2 x 2.0Ghz = 4Ghz = 2.5 x 1.6Ghz

I know that a Dual 2Ghz is not actually equal to a 4Ghz system, but the Dual 2Ghz G5 PowerMac's PSBench numbers could very well be over twice as fast as the 1.6Ghz G5 PowerMac. The 2Ghz G5 is obviously faster than the 1.6Ghz G5 and the PowerMac has 2 of them. Plus the Dual 2Ghz system has slightly faster memory than the 1.6Ghz system.
Click to expand...

Look at the posted results again.

How many of them improved between a single Athlon and dual Athlons? Not all that many, I'm afraid.

PSBench is not multiple-proc friendly.
 
Re: Re: Re: Overall times

Originally posted by Cubeboy
Whether or not those filters play a big role in the overall time is irrelevent. You simply can't choose to "see" what a particular CPU is faster at and what that CPU is slower at without considering the extent. Your win-lose perspective is too black and white to accurately compare either cpu, if the P4 beats the G5 slightly, it's considered a "win", if the G5 beats the P4 significantly, it's also considered a "win", there is nothing in the details and thats what I think the overall score offers.
Click to expand...

I agree. There is in fact no perfect methodology unless you know exactly how much time you will be spending in each filter when you use Photoshop. I am simply pointing out that comparing wins vs. losses is a much better (though still very imperfect) test of "overall" performance.

Just because alot of the time is spent on those four specific filters doesn't refute the validility of the time it took to actually finish with everything!
Click to expand...

It invalidates it as a meaningful test of general Photoshop performance. It does not invalidate it as a meaningful test of those 4 specific filters.
 
Re: Re: Re: Re: 1.42 results

Originally posted by jettredmont
The numbers vary wildly between multiple runs of the SAME machine! You have to run the benchmark 5-10 times, throw out outliers, and average the rest to get reproducible and meaningful results!
Click to expand...

My point exactly!
 
Re: It's not low end vs. high end

Originally posted by Bregalad
While I realize that 1.6GHz is the low end of the G5 chip, the 1.6GHz PowerMac G5 isn't a "low end" computer and it's not being compared to a "high end" P4 based computer. What makes a complete package high or low is its price. The AthlonXP 2400+ is definitely low end (roughly the price of an eMac) while the 3GHz P4 apparently costs the same as the 1.6GHz G5 making them both mid-range. The high end would be a dual Xeon versus a dual G5. We love our Macs, but let's be fair in our comparisons.
Click to expand...

As pointed out by others, the 3.0 GHz P4 and 1.8 GHz G5 are closer in terms of cost than the 3.0GHz P4 and 1.6GHz G5.
 
Those PCGaming Opteron Photoshop numbers don't tell us which filters they are using, nor which file they are using.

If we don't know that info, those numbers are meaningless when compared to the 50 meg file PSBench7 numbers.
 
Macrumors12345:

Looking back, I see your point, so I'm going to do another comparison which compares the CPUs in terms of how they perform in each filter.

Filter--------------P4---G5----Result
Rotate 90---------0.2---0.7---P4:250% faster
Rotate 09---------2.7---2.8---P4:4% faster
Rotate 0.9--------2.5---2.7---P4:8% faster
Gaussian Blur 01---0.8---0.7---G5:14% faster
Gaussian Blur 3.7--2.0---2.4---P4:20% faster
Gaussian Blur 85---2.3---3.4---P4:48% faster
Unsharp 50/1/0/---0.9---1.2---P4:33% faster
Unsharp 50/3/7/0--2.1---2.9---P4:38% faster
Unsharp 50/10/5---2.2---3.5---P4:59% faster
Despeckle---------2.2---0.8---G5:175% faster
RGB CYMK---------7.3---4.2---G5:74% faster
Reduce60---------0.9---0.8---G5:13% faster
Lens Flare---------2.5---6.1---P4:144% faster
Color Half-Tone----2.2---4.3---P4:95% faster
NTSC Colors-------2.4---4.2---P4:75% faster
Accented Edge----10.9--16.3--P4:50% faster
Pointillize----------12.1--25.0--P4:107% faster
WaterColor--------26.4--35.7--P4:35% faster
Polar Coordinates--7.0---5.0---G5:40% faster
Radial Blur---------33.1--53.9--P4:62.8% faster
Lighting Effects----1.9---3.5---P4:84% faster

Total filters won by 3.0 GHz P4: 16/21
Total filters won by 1.6 GHz G5: 05/21

filters 3 Ghz P4 outperformed 1.6 GHz G5 by 10%+: 14/16
filters 1.6 GHz G5 outperformed 3 Ghz P4 by 10%+: 4/5

filters 3 GHz P4 outperformed 1.6 GHz G5 by 25%+: 13/16
filters 1.6 GHz G5 outperformed 3 GHz P4 by 25%+ : 3/5

filters 3 GHz P4 outperformed 1.6 GHz G5 by 50%+: 9/16
filters 1.6 GHz G5 outperformed 3 GHz P4 by 50%+: 2/5

Assuming linear scaling (25% faster in every filter) from the 1.6 GHz G5 to the 2 GHz G5, it is fairly reasonable to say that a single 2 GHz G5 will outperform the Pentium 4 in any filter in which the Pentium 4 did not perform 25% better than the 1.6 GHz G5.

Speculation:

Total filters won by 2.0 GHz G5: 8/21
Total filters won by 3.0 GHz P4: 13/21

filters 2 GHz G5 outperformed 3 GHz P4 by 10%+: 7/8
filters 3 GHz P4 outperformed 2 GHz G5 by 10%+: 11/13

filters 2 GHz G5 outperformed 3 GHz P4 by 25%+: 5/8
filters 3 GHz P4 outperformed 2 GHz G5 by 25%+: 8/13

filters 2 GHz G5 outperformed 3 GHz P4 by 50%+: 3/8
filters 3 GHz P4 outperformed 2 GHz G5 by 50%+: 5/13
 
Exactly how much a 3 GHz Pentium 4 desktop would cost would depend largely on the other components, these days, you can configure a computer with the top of the line processor and still price it significantly cheaper than a high end/midrange desktop.

Dell Dimension 4600:
Pentium 4 3.2 GHz
512 MB Dual Channel DDR-400
80 GB Ultra ATA/100 HD
48x CDROM
48x CD-RW (Free second slot)
Basic Mouse+Keyboard
No Monitor
Radeon 9800
Integrated Sound Card
Altec Lansing ADA215 Speakers
Windows XP Home Edition (bleh)
WordPerfect Productivity Pack
56K Data/Fax Modem

Total Price: $1528


Dell Dimension 8300:
Pentium 4 3.2 GHz
512 MB Dual Channel DDR-400
200 GB Ultra ATA/100 HD w/Databurst Cache
16X Max DVD-ROM
4x DVD+RW/+R Drive w/CD-RW (savings)
Basic Mouse/Keyboard
No Monitor
Radeon 9800 Pro
SoundBlaster Live! 5.1 Digital Sound Card
Harman Kardon HK-395 Speakers with Subwoofer
Windows XP Professional
Word Perfect Office 11
56K Data/Fax Modem
Integrated Intel PRO 10/100 Ethernet

Total Price: $2487
 
The 1.6

I got to play with one today at CompUSA. First off, there is no computer that I have seen before made like this one. The finish is extraordinary. Lots of Wintel folks came up, touched it, said wow and started asking questions of the Apple folks in that section of the store. Internally, it is, again, distinctive: solid, curved lines on the airflow deflectors, NO WIRES or crappy cables to be seen (except at the very bottom of the casing). This is made like a German scientific instrument or a medical instrument... One problem: the Superdrive door drops down with a real snap! and the tray pops out... or at least it should but at first the tray got stuck in the drive aperture. this was fixed by actually pushing downwards on the whole drive and it suddenly clicked into place... this dislodgement may have been due to shipping etc... who knows?.. It was easily fixed. Overall sense of performance? Can't tell. Seemed as good as my dual 1Gig G4 machine but then again, I wasn't benchmarking. Most things launched quickly. One bounce two bounce... oh iDVD was 6 bounces but I don't use it so I have no reference point for that app.
Anyway I was impressed that Apple has produced a machine of exemplary PHYSICAL qualities... After looking at the Sonys, Dells and HPs... well, they look like cheap plastic and fake metal... the nouveau American car!! So stunningly........... well, let's drop that one.
 
where's your compusa? I don't think there's any at Maryland yet...so if you're west coast, then you probably get them earlier--so i guess a few more weeks over here then.
 
CompUSA

This one is in Reno NV. They have a superb Apple section (thank you Sheila and Nate!) and Apple is thriving here in the upper portions of the Silver State... now back 'home'-- in OZ, it's not as good but even there, under new management Apple is making some progress... it just has to keep going and not lose direction as it has done before. Australia deserves better... they are, after all, decent folk!
 
benchmarks, percentages, times, etc

some of those %'s are such low numbers, a 250% difference in speed is pointless..

.2 seconds vs .4 seconds, etc.

total time for the test would be a more accurate judge, imho.

I am just not motivated enough to even try and compare.

heh
 
Originally posted by eatme8888
Those PCGaming Opteron Photoshop numbers don't tell us which filters they are using, nor which file they are using.

If we don't know that info, those numbers are meaningless when compared to the 50 meg file PSBench7 numbers.
Click to expand...

I was beginning to wonder if anyone was going to catch on to that. :rolleyes:
 
Shaktai: I posted those numbers/urls ONLY because barefeats included some of those numbers in its news bulletin earlier today.

They say that "the dual 2ghz G5 (without G5 plugin for Photoshop) is 44% faster than the Dual 2.4GHz Xeon.

However, they base their comparisons on a low-end model, the 242, which was reviewed 3 months ago.

At the SAME site, barefeats.com NEGLECTS to bring up a review made yesterday, 8/26/2003 of the faster 244 opteron model.

Again, as i had said before, the 244, with Barefeats' logic is 47% faster than the dual 2.4ghz xeon. All of a sudden, because the dual 2ghz G5 beats the 242, the

"G5 should easily beat the Dual 1.6GHz Opteron. The Dual 2GHz Opteron might be another story."


They put forth a confusing blanket statement that b/c a low-mid range 242 is beaten by the 2ghz G5 (w/o plugin :) ), that the "G5" (which one) easily beats the opterons. They don't emphasize that they are comparing the 2ghz G5's results to those of the 1.6 opteron!

Nonetheless, they admit that the dual 2ghz opteron will be a different story--indeed it should :)
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back