Forums | MacLife

This is stoopid.

This morning at my shop, I ran the same 115MB psd file I used previously through a client's Celeron 2GHz with 512MB ram.


Rotate 90: 2 sec
RGB>CMYK: 6.6 sec
Unsharp Mask (200% 5 pixel): 6.5 sec
Gaussian Blur 1 pix: 2.2 sec
Gaussian Blur 25 pix: 7.2 sec

shoot,

I'm getting a PC. Cancel the G5 and the extra ram, take the money and pick up a Dell and use the extra cash to buy the Adobe Design suite and call it a day.

At least that is what I feel like. I can't get times like that Celeron on my G4 933.



(checks the Dell site) OK, well those boxes are a little pricey. But what the heck, I can just get a Celeron box add a few peripherals and there you go! At least that is what it seems like. Hmm... I will keep the G5 on order but it had better run these tests in microseconds or its off to Ebay!

It's a 2GHz celeron, so I wouldn't expect the 933 to outperform it. But I bet there are plenty of photoshop actions where a 933MHz G4 would be competitive with a 2GHz Celeron and possibly beat it.

As far as the Macaddict chart goes; if it's accurate, then it makes a laughingstock of G4 and G5 Photoshop speed performance.

...it can't be accurate.

Either the scale is screwed up, or, the .psd file being used is 115MB on the hard drive, and when opened up in photoshop is much larger because it's uncompressed. If that's the case (and it could be!), then Rik should be aware that when reporting a file size in Photoshop, one always reports the file size from within Photoshop. IE, one of my 115MB test file uses up about 30 MB as it sits on the disk and uncompresses to 115MB when opened in Photoshop, another uses 111MB on the disk and uncompresses to 115MB.

... or perhaps Adobe.

Either or both may suggest a test file to use that is on their site(s) somewhere such that across-the-boards testing can be common.  Because there is just no chance that an Apple G5 is going to lose in any comparison clock-to-clock, let alone when a 64-bit Photoshop app appears along with a G5 optimized OS X.

Otherwise, readers are going to get the wrong idea, quickly!

Do keep in mind, however, that in the comparisons of the two G4s and two G5s, I was testing stock systems -- that is, ones with a mere 512MB of DDR SDRAM.  With that amount of memory, Photoshop is, of course, going to disk quite heavily -- which will, of course, depress performance.

Rik, it should hardly disk at all if you are only running one test and then quitting and relaunching.

I still maintain that the best sort of test is where a long sequence is performed so that all factors( bus speed, memory amount and speed, disk performance, processor(s), memory settings) come into play.  The beauty of the action that I posted yesterday is that it creates its own file so that there can be no quibbling over the characteristics of the file.

MA, why don't you give that a try?

Oh, Ribtorus, thanks for pointing out that the Celeron was a 2ghz - phew.

MA Rik et al, back to the drawing board on this one. I, for one, am going to completely disregard your results

The chart's scale almost looks to be off by an order of magnitude. I could believe it if the scale were in tenths of seconds, particularly the rotate and Gaussian blur times.

It's wierd.

I stand behind the tests -- I performed them myself, using, as stated, a 115MB Photoshop file (a rather complex raw image from a high-end digicam), OS 10.2.7, Photoshop 7.0.1 with the G5 patch, repeating each test over and over and over and over again.  What you see is what happened.

Do keep in mind, however, that in the comparisons of the two G4s and two G5s, I was testing stock systems -- that is, ones with a mere 512MB of DDR SDRAM.  With that amount of memory, Photoshop is, of course, going to disk quite heavily -- which will, of course, depress performance. 

When I bumped the Dual 2GHz G5 up to 2GB of SDRAM, Photoshop flew -- not surprisingly, as Photoshop loves scads of memory.  I look forward to stuffing 2GB into the 1.8GHz G5 and seeing how it performs, and also incrementing the total SDRAM from 512MB to 1GB, to 1.5GB, to 2GB, to 2.5GB, and seeing where the performance boost kicks in.

Also, I look forward to running far more tests on the Dual 2GHz G5.  As I mentioned, we only received it in its stock configuration late Friday morning (and in the tricked-out config on Thursday) -- many more tests are to come, using not only Photoshop (with smaller and larger files), but also with a raft of other apps.  I look forward also to frame-rate comparisons of the Radeon 9600 vs. the 9800 (which, by the way, has its own fan, bringing the total number of fans in the Dual 2GHz G5 to 10).

-snip-

But as to the Photoshop testing...  Hey, if you can enlighten me about how you think those times might be wrong, I'd be more than happy to listen -- but, as I said, all tests were repeated many times, and all results were within repeatable range.

Ric, I'm sure your times are correct, but from my POV, the tests aren't that helpful in that they aren't really pitting the processor & memory bus only, on the various systems, against each other.  IMO, you might begin repeating them using a small enough file that will fit easily in the 512MB of ram that each system has OR post charts comparing machines that have enough ram to hold the 115MB file in memory.   Isn't the rule of thumb roughly 5-6X the file size?

I've also heard on other forums that the G5's Energy setttings should be set to "Highest" and I also read that there's a setting somewhere to uncheck that keeps the hard disk noise level at "quiet."   I think this is the default setting, and I suppose it might slow down disk reads (though hopefully with enough ram this wouldn't be an issue).

Thanks for doing all the work of benching.  If nothing else, as you said in the article, the last chart shows the real world value of having plenty of ram!

Guys, I just ran these tests on my SE/30 with 4MB of RAM and slow HD.

Rotate 90: 9.6 s
RGB -> CMYK: 18.3 s
Unsharp Mask: 19.6
Gaussian 25 px: 12.9
Gaussian 1 px: 8.6

I think Rik's tests must be flawed 

It be my understanding that OS X does not use Ram allocation. Photoshop and other programs only grab as much Ram as they need.

The results shown in both Photoshop charts are dismal.

They show a group of crippled workstation-class machines that are easily outclassed by almost any modern PC (and dare I say mac).

The second Photoshop chart is even worse; it shows a G5 MP with 2GB ram that is performing at about the level of a single processor 2GHz Celeron that has one quarter the amount of ram.

This is likely the result of very selective benchmarking. It's almost as if the trial was designed to illustrate the benefits of buying 2GB of ram. But from the text, we know that the extra ram was merely serendipitous. I wonder why, if it is known that 512MB of ram cripples a G4 or G5, the decision was made to use such a relatively large file. What would it have cost to use a smaller file? Why not illustrate the processor and ram speed by taking the disk sub-system out of the equation?

I look forward to further results.

You are all judging the G5s total worth as a processor over a quick off the cuff test done at the last minute on a Friday afternoon. Personally I have more intelligence to wait for full tests to be done on these machines before I start touting which processor is better, etc.

Besides, it is NO SURPRISE to me that the G5 had an incredible boost in performance when more RAM was added. The beauty of the G5 (computer, not processor) is the fact that EVERY part of the system is fast. The bus, memory, CPU, etc. are all fast. You can't say that about most P4 systems.

Later, Frank

You are all judging the G5s total worth as a processor over a quick off the cuff test done at the last minute on a Friday afternoon. Personally I have more intelligence to wait for full tests to be done on these machines before I start touting which processor is better, etc.

Besides, it is NO SURPRISE to me that the G5 had an incredible boost in performance when more RAM was added. The beauty of the G5 (computer, not processor) is the fact that EVERY part of the system is fast. The bus, memory, CPU, etc. are all fast. You can't say that about most P4 systems.

Later, Frank

No one is judging the G5 - just the test results versus other machines. My P4 with LESS RAM and slower clock speed still beat the G5 with 2GB of RAM in most areas when duplicating the MacAddict test. My 933 G4 with half the RAM also comes very close to the G5/2GB results.

Something is very wrong with the optimization of Photoshop for the G5 running OSX 10.2.7 or with 10.2.7 itself. Can someone confirm that the unit was tested with 10.2.7 as the machine is intended to run? Plus, as has been suggested, was the processor performance set on "highest" in the System Preferences instead of "automatic"?

This may have started as with people questioning the tests but it has quickly evolved to people bashing the G5 based on these tests alone. Go back and read the thread and you will see what I am talking about.

Statements like the following:

If the P4 is outclassing the G5...

OS X and Photoshop for OS X suk donkey nuts.....they're not ready for the hardware we're throwing at them.

The bottom line is that if you are getting faster times using a 400MHz Celeron then you are not performing the tests in the same manner with the same file that MacAddict used. I love how you all think you can recreate the same tests. I admit that MacAddict messed up by putting up tests that were done at the last minute, etc. They very well should have prepared much more and simply waited until the tests were ready for the magazine.

However, don't try to claim that every other computer on earth is faster than the G5. It simply isn't going to hold true in the real world. Making false claims based on partial information is no better than the results that you are already bashing. Get a grip and be patient. Real results will be forthcoming soon enough.

Later, Frank

Go find a G5. Play with it for 30 minutes. Get back to me then. 

I already have but I'm intelligent enough NOT to base my opinion of a machine on "playing" with it for 30 minutes. I know that it isn't how fast a machine launches a couple of apps or does a few things that you judge its overall speed from. You won't be able to judge it until you use it to do the things you do every day on a computer and you do that for at least a week. Until people do that and beging to post everyone needs to chill out.

The level of whining and bitching in this post is getting out of control when we are talking about the short few tests that MacAddict did. Get a grip on yourself.

Later, Frank

Rik responed to my concerns in this thread, and he assured me that he felt there was nothing odd about the tests, and that the times he posted were in fact correct.

There has been no subsequent responses from Rik, and the charts have remained unchanged. I can only conclude that Rik and Macaddict are satisfied with those results.

Personally, I'm pretty sure that when properly tested against a similarly configured 1.42 MP, the 2GHz G5 MP is about 20% to 30% faster on average in Photoshop. As far as I'm concerned, it's a so-so improvement. Ideally it should be at least a 49% improvement due to the clockspeed alone, never mind the enhanced memory architecture.

I'm also pretty sure that in a proper test involving Photoshop actions that go beyond rotating, gaussian blur, unsharp mask and RGB>CMYK conversion (something more like the PS7Bench trials), The G5 MP will show itself to be at about the level of a dual Xeon 2.4GHz-2.8GHz. There's no shame in that, but you'd never know it from the charts that Macaddict has posted so far.

I think the published charts are designed to demonstrate the need to run the G5s with at least 2GB ram.

That is exactly what I am saying. The tests were bad. Yet, even though we agree on this fact many people in this post are still acting like the G5 is crap based solely on these tests.

See my point?

Later, Frank