'Real World' Photoshop Benchmark - Speed Test

[Horst]

At least according to the results posted her, I think it's amazing how a moderatly pimped out G5 DP from 2 years ago is still among the fastest Mac offerings for Photoshop .
And will be for another year till the arrival of CS3, when probably an even more advanced Rev B MacPro w/o possible initial bugs will be available !

For me, that's the first time Apple get's their hardware releases perfectly timed ...

 

[worriedmac]

There are plenty faster speeds but I'm very happy with my machine

Powerbook 1.5GHZ with the stock 4200 harddrive and 1GB of RAM
Time with 80% memory used and 20 history = 7min50
Time with 100% memory used and 1 history = 4min40

I'm upgrading the ram and harddrive soon, I'll test again after that.

 

[Solar Core]

Time: 9 minutes 53 seconds

Test Speed: 9 minutes 53 seconds
System Specs: Dual 2 GHz G5 tower with 1 GB RAM, Main Hard Drive is OEM, Radeon X800XT Running dual displays. 160GB, plus 2 external 250 GB (1 is internal SATA, the other is 250GB USB). Photoshop CS 1

Probably the reason it was so slow is that I have a few apps idling in the background.

I was wondering what would happen if I quit one of the apps. One that I had noticed was particularly cpu intensive, even in the background. It helped, or so it seemed as my new reults are over a minute faster

Old Results: 9 minutes 53 Seconds
New Results: 8 minutes 35 Seconds

I'll test again with that app running again to see if that made the difference or it was just the cache (although I did restart photoshop last time).

 

[timskrastins]

I was wondering what would happen if I quit one of the apps. One that I had noticed was particularly cpu intensive, even in the background. It helped, or so it seemed as my new reults are over a minute faster

Old Results: 9 minutes 53 Seconds
New Results: 8 minutes 35 Seconds

I'll test again with that app running again to see if that made the difference or it was just the cache (although I did restart photoshop last time).

You might want to check how many history states you have (and reduce all the way down to 1 for some comparisons with other posts here) and also if Photoshop or the scratch disk is running off the USB drive instead of the internal I would imagine that could slow it down too.

 

[RichP]

I was wondering what would happen if I quit one of the apps. One that I had noticed was particularly cpu intensive, even in the background. It helped, or so it seemed as my new reults are over a minute faster

Old Results: 9 minutes 53 Seconds
New Results: 8 minutes 35 Seconds

I'll test again with that app running again to see if that made the difference or it was just the cache (although I did restart photoshop last time).

My G5 was really slow too, I found that reinstalling the OS REALLY helped out. Does the machine seem rather slow otherwise?

 

[Solar Core]

Ok unfortunately my slow speed wasn't because of the idling apps, or at least the evidence is less conclusive.

Time 1: 9 minutes 53 seconds (6 apps idling)
Time 2: 8 minutes 35? (5 apps idling)
Time 3: 8 minutes 49 (same 6 apps idling)

My G5 was really slow too, I found that reinstalling the OS REALLY helped out. Does the machine seem rather slow otherwise?

Wow, that is a lot of work. I just installed Tiger about a month ago. Of course it was an upgrade install, not a brand new install.

The machine has been slower lately but I didn't think it was that slow.

You might want to check how many history states you have (and reduce all the way down to 1 for some comparisons with other posts here) and also if Photoshop or the scratch disk is running off the USB drive instead of the internal I would imagine that could slow it down too.

I have 20 history states. I wouldn't actually want to run photoshop with just 1 history state though, so it wouldn't actually do me any long term good (since I often go back). I suppose I could sleep better knowing that my G5 isn't actually crippled, its just differently configured, but at 9 minutes for every iteration I'm already tired of running the test.

 

[CTYankee]

Dual 2Ghz Rev A G5, 3GB RAM, 74GB Raptor + seperate scratch disk

CS1 - 2m03s

CS2 - 5m00s

And that is why I am still using CS1 - CS2 was a watse of money for me.

something is wrong with your system. People using the same setup getaround 2 minutes.

 

[CTYankee]

OK, I have put together some averages:
PSCS2, 20+ history states

dual 867 G4 351
Dual 1.4 G4 266
Mini 1.42 578
PB 1.0 G4 548
PB 1.33 G4 297
PB 1.5 G4 470
PB 1.7 G4 185
MB 2.0 413
MBP 2.16 385
single 1.8 G5 261
dual 1.8 G5 196
dual 2.0 G5 139
dual 2.3 G5 133
dual 2.5 G5* 117
Dual 2.7 G5* 129
Quad 2.5 G5 103
Quad 2.6 Xeon 213
Quad Opteron 94

Dual 2.5 and 2.7 data limited and has an outlier in each. One is a system using a Raptor scratch disk (dual 2.5) and the other is a 2.7 that ran the test in 153 seconds. Remove these two and the averages would be 2.5=130s, 2.7=117s.

 

[timskrastins]

1:50

PowerMac G5 dual 2.7GHZ, 4.5GB RAM
SATA 250GB boot drive
CS2
99 history states

I was also screwing around in other apps (Entourage, Safari, timer program) and running 2 huge Finder copies between attached USB2 and Firewire drives....in other words, typical everyday usage for me while I'm doing Photoshop. Wow...I sure hope Rosetta is slowing those Xeons waaaaaay down because I'd expect a machine so much newer than mine to stomp the crap out of mine.....I mean like 20-30 seconds at most. Can't wait to see if that's the case with the Universal Binary of Photoshop.... c'mon Adobe, you can do it!
-------------

Heh, ran it again with 1 history state (so the scratch disk wouldn't get used and indeed it ran at 100% efficiency this time, meaning it wasn't having to rely on the hard disk), still with Safari, Entourage, the timer and the 2 big Finder copies going in the background....this time 1:09

Now I've been able to run it off a fresh restart with no other apps running, no finder copy, etc. Three attempts at each setting and it's interesting to see the first run seems to be slower (then I close the file and re-open it).

1 history state - 01:20, 00:50, 00:50

20 history states - 2:14, 1:28, 1:25
indeed, efficiency dropped as low as 15% on this setting, indicating it was using the scratch (boot) disk.

 

[the kid 05]

Yeah, it is just as noisy as the maxtor - you can hear it read and write. not worse than the maxtor though.

never heard the disk in any other computer though.

maybe due to the tower design? (all the holes?)

 

[Anonymous Freak]

Rosetta vs. Parallels.

Okay, so I decided to run a test to compare Rosetta emulation speed vs. Parallels virtualized speed.

My config:
15.4" MacBook Pro 2.0 GHz, 2 GB RAM, 7200 RPM 100 GB hard drive.
OS 10.4.7, heavily used (the computer has been my primary machine since I got it, and it was in the first batch to be sent out.) It has the last 6 GB lopped off for a Boot Camp partition, and the primary partition only has 9.4 GB free. So from a disk perspective, this is horribly un-optimized. Windows in Parallels is a fairly 'fresh' install, with only Microsoft's 'OneCare' anti-virus program installed.

Common settings: Photoshop CS2 30-day tryout version downloaded from Adobe's website, 1 History level, memory usage to 90%

Differences: OS X has access to both 2.0 GHz processors, and the full 2 GB of RAM. But its running through Rosetta. Windows (via Parallels) has access to only one processor, and I only have 1 GB of RAM assigned to it. So Windows thinks it's a single-processor, 1 GB system. But it's running 'natively'. (Parallels has an 8GB 'hard drive' that is really just a file in the OS X world.)

Results:
Rosetta: 7 minutes, 14 seconds.
Parallels: 6 minutes, 20 seconds.

Again, this is horribly unoptimized. But even that shows that running the test through Parallels produces a faster result. So, if you happen to own a spare copy of Windows, and Photoshop for Windows, running through Parallels would be a good way to go until Adobe releases a Universal version of Photoshop.

 

[Mitthrawnuruodo]

[Experiment] So, if you happen to own a spare copy of Windows, and Photoshop for Windows, running through Parallels would be a good way to go until Adobe releases a Universal version of Photoshop.

Well, I got the opposite result running that other, smaller test using Photoshop CS. On that Rosetta beat Windows 2000 under Parallels 1.06 to 1.47.

 

[Grimace]

2:27 PM G5 dual 2.7 CS
1:12 PM G5 Quad (fastest of several runs) CS2

I just got 2:49 on my new Mac Pro (stock video card and 3GB RAM). These things aren't too slow with Photoshop! I really don't notice very much difference between my previous dual-core 2.3Ghz G5 with 3.5GB RAM. My new Mac Pro has less RAM but seems to keep up at the same speed. I guess I don't do hardcore Photoshopping like everyone else

 

[Horst]

Great test , thanks for setting it up !
It got me to play around a little with my PS settings, with interesting results.

My rig : DP 2.0 Ghz, Rev. B, 4.5 GB Ram, internal SATA scratch disk (half full), Radeon X800XT .

Photoshop CS2 / 9.01, Cache: 2


History 15, Ram 70% : 1.58

History 1, Ram 85% : 1.17

History 1, Ram 85%, 'Bigger Tiles' Plugin enabled : 1.06

History set to more realistic 10, Ram 85%, 'Bigger Tiles' Plugin enabled : 1.40

My results with Photoshop CS / 8.01, Cache levels: 2
CS2 results see above.


History 1, Ram 95% : 1.10

History 10, Ram 95% : 1.40

History 20, Ram 95% : 1.53

I upped the Ram allocation a little for CS, as it isn't able to use more than 2GB, as opposed to the 3GB CS2 is capable of using.

 

[Grimace]

My results with Photoshop CS / 8.01, Cache levels: 2
CS2 results see above.


History 1, Ram 95% : 1.10

History 10, Ram 95% : 1.40

History 20, Ram 95% : 1.53

I upped the Ram allocation a little for CS, as it isn't able to use more than 2GB, as opposed to the 3GB CS2 is capable of using.

What kind of machine are you using?

 

[tomhayes]

Okay, so I installed BootCamp on my Mac Pro (1gb, ATI 7300, 250GB) a few times.

Each time I installed from my original Windows disk and updated drivers for Nvidia, Intel 5000 Chipset and ethernet drivers manually.

I installed a copy of Adobe Photoshop, downlaoded and installed all updates from Adobe and left all settings as default (20 history states, 6 cache and 55% memory.)

Well, something is wrong. It takes over 15 minutes to complete the full test and over 3 minutes to do the first two steps.

Same result each install. On the last install I put history states to 1, Cache to 1 and Memory to 100% (888 megs) and it made no differnce.

I installed BootCamp with a partition of 50GB each time.

So something is wrong with this picutre.

Maybe Windows 64 would help? I don't own a copy of Windows 64 so it won't be me who cracks this nut.

 

[RichP]

Wow, that is a lot of work. I just installed Tiger about a month ago. Of course it was an upgrade install, not a brand new install.

The machine has been slower lately but I didn't think it was that slow.

I was averaging around 6 minutes, with some runs over 7, before I reinstalled. After, I was in the 3 minute range.

And yes, I know its a lot of work, lol.

Okay, so I installed BootCamp on my Mac Pro (1gb, ATI 7300, 250GB) a few times.

So something is wrong with this picutre.

Maybe Windows 64 would help? I don't own a copy of Windows 64 so it won't be me who cracks this nut.

Hmm..that shouldnt matter; Adobe isnt an app that is 64 bit anyway. (I think I am correct on this) At most, you wouldnt be optimized, but in no way as slow as your machine is currently running.

 

[macintoshse1]

G5 Dual 2.7
2 GB
Photoshop 7.0

Time 2:45

 

[Horst]

What kind of machine are you using?

It's in the quoted text of my post.

 

[greenmac]

I decided I wanted to see what a G3 could do! So I turned on our old iMac(summer 2000) and installed cs. This is a 400MHz, 192MB RAM, 8MBVRAM, 10GB HD. well after a blistering 34mins the HD ran out of space on the Gaussion blur. History states 20 55%memory usage.

eMac 1.25GHz, 256MB RAM, 80GB HD 32MB VRAM, running cs 11m 21s
Pentium4 3.2 GHz 512MB RAM, 256MB VRAM, running cs2 8mins 7s

 

[greenmac]

Got a Mac Pro, but only have Elements. I can't seem to get the action to work

Kimo

you can get a free trial download off the adobe website, the trial lasts 30 days (I think).

 

[Jinorasa]

10min 47sec; 12" PowerBook G4 867Mhz 1152MB RAM 60GB RPM7200 HDD, Photoshop CS2 20 history states and memory usage 70%.

Quite satisfy with the results actually....

 

[Anonymous Freak]

I decided I wanted to see what a G3 could do! So I turned on our old iMac(summer 2000) and installed cs. This is a 400MHz, 192MB RAM, 8MBVRAM, 10GB HD. well after a blistering 34mins the HD ran out of space on the Gaussion blur. History states 20 55%memory usage.

Oooh, what an idea! I'll have to try it on my original Bondi iMac G3/233 MHz with 96 MB of RAM! (1 history state, that way I don't run out of HD space.)

 

[Chuck]

These Mac Pros don't seem as fast as they should be.

 

[Seinfeldt]

Okay folks,

I read (almost) the whole tread now and it seams like an nice test, so I dont bother testing it in my PB 17", beacuse it seams like that part is covered.

So, i went home to my gaming rig which is an:

Dual Dual Core Opteron 280 2,4GHz
2GIG ram
4xSeagate Cheethas 36,6 gig's in SCSI 320 Raid 0 on an Adaptec card.
1xWestern Digital Raptor 76gig as an SWAP disk for PageFile
2 x Nvidia 7900GTX SLI
WinSHIT XP SP2
PS CS2

And did ran the test:

1.35m

I'd really hoped for an better score here, but I see it is quite memory intensive, so i guess if I had 4 gig's it would be better, also if I could have booted on the Raptor an used the SCSI array as an scratch disk.

my 2 cents


(want an MacPro) with bootcamp for gaming


peace