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

'Real World' Photoshop Benchmark - Speed Test

jerryrock said:
OK I ran it again using my iPhone as a timer this time!

1st run: 9.10 seconds

2nd run: 9:08 seconds

I believe the variable is due to reaction time hitting the stop button on the timer. Can't this test be rewritten to use Photoshop's built in timing function?

MacPro 6,1 3.5GHz 6-Core Intel E-5, 64GB 1866MHz DDR3 (OWC) RAM, dual AMD FirePro D500's, Apple 256GB PCI SSD.
Adobe Photoshop Version: 14.1.2 (14.1.2 20130923.r.427 2013/09/23:23:00:00) x64
Operating System: Mac OS 10.9.1
Click to expand...

That's within a 10th of a second of what I got on the same setup (although I have a 32GB/1TB FWIW).
 
I just tested at the Apple Store... (without a system restart, but made the PS preference changes.)

Base model nMP, current version of PS CC.

Tested 3 times, 7.8 sec, 10.9 sec, 11.2 sec (a friend timing on his iPhone, so some variability there).

I'd wager that the current test isn't pushing the new hardware enough to see differences between the Quad or Hex systems.
 
ozbimmer said:
My MacPro5,1 6c(W3680)/48GB/512GB PCIe SSD/Quadro4000 with Photoshop CC 14.1.2 gives a time of 17.7 seconds. Compare to a Dell T7600 with dual 8c(E5-2680)/256GB/1TB PCIe SSD/Quadro K5000 the time was a shocking 6 seconds.

I notice radial blur, which is used in this test, uses all CPU cores.
Click to expand...

Hello,
I'm the person who has hosted the radial blur test for the past three years.
Your time of 6 seconds seems about right for the computer you mentioned.
The current fastest time on the test is is 8.3 seconds that was run on a workstation with 2 X 8 core Xeon CPUs running at 2.3 GHz, with turbo boost to 3.1 GHz.

Your/the listed Dell workstation is using 2 x 8 core Xeon CPU's running at 2.7 GHz, with turbo boost to 3.5 GHz.

The test uses all cores and seems to run them at the turbo boost speed.

As your test result would now rate as the fastest listed, I would hope you would leave a comment on the test page with data so I can add your time to the top of the chart.

Thanks,
Keith
 
Mac Pro 5.1 Hex 3.33GHz
HD 7970 reflashed
12GB 1333 Memory
Samsung Pro 512GB on Sata3
Photoshop CC

I tested using the "Radial Blur Speed Test" and got a consistent 17.5 seconds. So the result of 14 seconds from the nMP 6 core is what I would expect from a newer CPU clocked at 3.5GHz .

I saw no difference in time with the GPU activated in Photoshop.

The real world difference between the 2010 MP and the nMP using GPU is confirmed by the "Bare Feats Final Cut Pro Shootout" in my mind
see http://www.barefeats.com/tube05.html

I will wait a year before I buy to let things settle down on a new product.
 
Mac Pro 3,1
2 x 2.8 gHz quad core Xenon
8GB Ram
Nvidia GeForce 8800GT 512Mb
OSX 10.9.1
Photoshop CC

15.45 Seconds for the speed test listed at the beginning of this thread

17 seconds for the Radial Blur speed test
 
Last edited:
Mac Pro 1,1 OS X 10.6.8
Dual-Core Intel Xeon 3GHz
Quad Core
14GB
NVIDIA GeForce 7300 GT

26 Seconds


Posting this so I can compare the nMP 3.5GHz 6 Core 32GB 1TB Flash D500s when it arrives.
 
results for MP3,1 & iMac (late 2013)

Thought it would be interesting to compare times with various versions of PS while I'm awaiting my nMP/MP6,1.

Mac Pro 3,1 (early 2008) 2x2.8 Xeon, 16GB 800MHz DDR2, ATI 5870 1GB, Samsung 840 Pro 256GB (MP3,1 backplane is SATAII) running OSX 10.8.5.

I tested with CS5, CS6 and CC 14.2. Results are averaged over 3 runs.

21.8, 18.3, 16.8 (CS5, CS6, CC)

iMac 27" (late 2013) Core i7 3.5, 16GB 1600MHz DDR3, NVIDIA GTX 780M 4GB, Apple PCIe SSD Samsung 512GB running OSX 10.9.1.

12.6, 11.7, 11.3 (CS5, CS6, CC)
 
PreFlexoPress said:
Mac Pro 1,1 OS X 10.6.8
Dual-Core Intel Xeon 3GHz
Quad Core
14GB
NVIDIA GeForce 7300 GT

26 Seconds


Posting this so I can compare the nMP 3.5GHz 6 Core 32GB 1TB Flash D500s when it arrives.
Click to expand...

I just got everything loaded and set up on my nMP. As quoted above it's the 3.5GHz 6 Core 32GB 1TB Flash D500s.

7.7 seconds :eek:

I thought something might be a miss so I trashed the speed test folder and downloaded it again. Reinstalled the SpeedTest Action. Set all the PS prefs. Restarted and ran it again.

7.6 seconds

I'm one of the nMP owners that is loving it's performance since I'm coming off a 1,1.

Excuse while I go clean out my shorts.
 
PreFlexoPress said:
I just got everything loaded and set up on my nMP. As quoted above it's the 3.5GHz 6 Core 32GB 1TB Flash D500s.

7.7 seconds :eek:

I thought something might be a miss so I trashed the speed test folder and downloaded it again. Reinstalled the SpeedTest Action. Set all the PS prefs. Restarted and ran it again.

7.6 seconds

I'm one of the nMP owners that is loving it's performance since I'm coming off a 1,1.

Excuse while I go clean out my shorts.
Click to expand...
which version(s) of PS?
 
analog guy said:
which version(s) of PS?
Click to expand...

CS6 (13)

I'll run the test with the cloud version whenever I get over being pissed off at Adobe for such an asinine decision. IMO of course.

Illustrator and InDesign are lightening quick, also.

nMP is being used for in-house prepress flexographic print.

I have my old 1,1 sit up and running right next to the nMP to manually transfer files. (I don't like migration assistant) So I've had the luxury to compare numerous tasks that are performed throughout the day.

One particular customer's files are large (10 to 40mb) Illustrator vector art files. I've opened a couple of their most recent larger files with the nMP and compared the frustration factor with my 1,1 machine which is sitting right next to me. Night and day difference in speed/performance.

There's a folder that comes with Ill. CS6 within the application folder titles "Cool Extras" > en_US > Sample Files > Sample Art > Modern-Day Venus.ai That files is a 39.3mb vector file and is similar to the art files I have to manipulate/fix.

I remember feeling this way 7 years ago when I bought these 1,1 MPs. It's a great feeling. Almost better than new car smell.
 
analog guy said:
Thought it would be interesting to compare times with various versions of PS while I'm awaiting my nMP/MP6,1.

Mac Pro 3,1 (early 2008) 2x2.8 Xeon, 16GB 800MHz DDR2, ATI 5870 1GB, Samsung 840 Pro 256GB (MP3,1 backplane is SATAII) running OSX 10.8.5.

I tested with CS5, CS6 and CC 14.2. Results are averaged over 3 runs.

21.8, 18.3, 16.8 (CS5, CS6, CC)

iMac 27" (late 2013) Core i7 3.5, 16GB 1600MHz DDR3, NVIDIA GTX 780M 4GB, Apple PCIe SSD Samsung 512GB running OSX 10.9.1.

12.6, 11.7, 11.3 (CS5, CS6, CC)
Click to expand...

adding this one:
nMP 6,1 6-core 3.5, 64GB 1867 DDR4, D500, Apple PCIe SSD Samsung 512GB running 10.9.1

10.6, 10.5 and 8.4 for CS5, CS6 and CC 14.2.

----------
PreFlexoPress said:
CS6 (13)

I'll run the test with the cloud version whenever I get over being pissed off at Adobe for such an asinine decision. IMO of course.
Click to expand...

you can do the free trial just as a test. if you did, indeed, get 7.7, based on my comparisons of cs5, 6 and CC on 3 different systems you should see a result that is up to 20% faster.
 
PreFlexoPress said:
I just got everything loaded and set up on my nMP. As quoted above it's the 3.5GHz 6 Core 32GB 1TB Flash D500s.

7.7 seconds

7.6 seconds

CS6 (13)
Click to expand...

analog guy said:
6-core 3.5, 64GB 1867 DDR4, D500, Apple PCIe SSD Samsung 512GB running 10.9.1

10.5 CS6
Click to expand...

My 6-core 3.5, 32GB, D500, 1TB did..

9.33 seconds on CS6

So on the same hardware, and same version of CS6, we have results ranging from 10.5 seconds, to 7.6 seconds. WTF?! :confused:

wildmac said:
I just tested at the Apple Store... (without a system restart, but made the PS preference changes.)

Base model nMP, current version of PS CC.

Tested 3 times, 7.8 sec, 10.9 sec, 11.2 sec (a friend timing on his iPhone, so some variability there).
Click to expand...

And here, same machine, same version of PS, results ranging from 11.2 seconds to 7.8 seconds.

I think all we can conclude from all this, is that this benchmark is a complete waste of time. There's way too much variability for it to be useful. A useful benchmark needs to be repeatable. This is anything but. :rolleyes:
 
VirtualRain said:
My 6-core 3.5, 32GB, D500, 1TB did..

9.33 seconds on CS6

So on the same hardware, and same version of CS6, we have results ranging from 10.5 seconds, to 7.6 seconds. WTF?! :confused:



And here, same machine, same version of PS, results ranging from 11.2 seconds to 7.8 seconds.

I think all we can conclude from all this, is that this benchmark is a complete waste of time. There's way too much variability for it to be useful. A useful benchmark needs to be repeatable. This is anything but. :rolleyes:
Click to expand...

too much is based on when one stops and starts their timer. the chance for error is high and that error represents a large portion of the benchmark.

if one is a second off with the LR benchmark (times in the 4-minute range) it doesn't make much difference.
 
analog guy said:
too much is based on when one stops and starts their timer. the chance for error is high and that error represents a large portion of the benchmark.

if one is a second off with the LR benchmark (times in the 4-minute range) it doesn't make much difference.
Click to expand...

Perhaps. But there's almost a 3-4 second variance (even by the same person). That means user error can account for around 30-40% of the result. Horrible.

No matter what the cause, the lack of repeatability makes it useless as a system benchmark.

If a Photoshop benchmark is important, then the medium test at Mac Performance Guide seems a lot better since no stop watch is required and it appears much more taxing.
 
VirtualRain said:
Perhaps. But there's almost a 3-4 second variance (even by the same person). That means user error can account for around 30-40% of the result. Horrible.

No matter what the cause, the lack of repeatability makes it useless as a system benchmark.

If a Photoshop benchmark is important, then the medium test at Mac Performance Guide seems a lot better since no stop watch is required and it appears much more taxing.
Click to expand...

yep--that's exactly the point. even a 1s error is significant, so when that user reported 3-4" difference per run i discounted the results.

my own results were quite repeatable. tend to be skeptical of the results that are significantly faster (or slower) for the same system.
 
Haswell i5 2.4 Ghz rMBP 13", 4gb RAM, 128SSD, Mavericks with PS CS6

32.4 seconds timed with iPhone.

Dell Inspiron 15r SE, i5-3210M 2.5 Ghz, 8gb RAM, 750GB 7200RPM, Win 8.1, PS CS6

24.13 seconds timed with iPhone.

This is odd because the rMBP just feels MUCH faster than the Dell.
 
MP 1,1 x5355 x 2

27.51 seconds
Mac Pro 1,1 installed 2 x5355's 2 years ago, runs great
Graphics 8800 and 6870
10.7.5
12GB RAM
CS 5.1

Big Mac MP 2 x 2.0
CS3 - does not recognize speedtest.atn. In the folder with other .atn's, but PS CS3 just doesn't see it, even after restarting.
Only the Default Actions folder shows up - molten lead, wood frame, etc
 
nMP 4 core... is the CPU really upgradeable?

I am buying a 4 Core nMP because it's all I can afford. Upgraded to 512 SSD and 16GB RAM. I heard that they will be moving to a new CPU for this in just a few months. Should I put my order on hold and get the newer CPU... or is the current 4 core chip well up to LR5/PS CC6 tasks?:confused:
 
AppleUSA said:
Haswell i5 2.4 Ghz rMBP 13", 4gb RAM, 128SSD, Mavericks with PS CS6

32.4 seconds timed with iPhone.

Dell Inspiron 15r SE, i5-3210M 2.5 Ghz, 8gb RAM, 750GB 7200RPM, Win 8.1, PS CS6

24.13 seconds timed with iPhone.

This is odd because the rMBP just feels MUCH faster than the Dell.
Click to expand...

Dell has twice the RAM.

When I went from 4 to 8Gb RAM in my 2010 MBP it made a big difference to the Photoshop benchmark time.
 
AppleUSA said:
Good point. I will try allocating only 4GBs of RAM to CS6 on the Dell and see what kind of time I get from that :D Thanks!!
Click to expand...

Might be worth checking what the Photoshop preferences are set to on the two machines, especially the Performance section.
 
Mac Pro 2009 (Flashed 5,1), 3.46GHz Hex core (W3690)
32G RAM 1066MHz non ECC, HD7950 Mac Edition, 840 Evo 1T SSD via Tempo SSD (SATA 3)
OSX 10.9.2, CC 14.0

About 10.5s with 2 profile logged in and little bit other softwares running at the background (e.g. mail, safari, hardware monitor...)

Before I upgrade my CPU (the previous one is 2.66GHz Quad core W3520), the best I can do is about 15s.
 
Last edited:
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