Welcome to the MacNN Forums.

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

You are here: MacNN Forums > Hardware - Troubleshooting and Discussion > Mac Desktops > New PSBench7 Tests

New PSBench7 Tests
Thread Tools
jrod7350
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 06:01 PM
 
Okay, I've used a G5 1.6 with 1GB of RAM, ran PSBench7 Advanced (50MB) file, with 1 History state and the Adobe optimization plugin installed. Here are my results (BTW, I can't seem to get it to format in a nice table, but time is in seconds for each of the three runs):

90 Degree Rotate .6 1.1 .4
9 Degree Rotate 3.0 2.8 2.7
.9 Degree Rotate 2.6 2.6 2.8
1 Gaussian .8 .6 .6
3.7 Gaussian 2.5 2.3 2.3
85 Gaussian 3.5 3.3 3.5
1 pix USM 1.3 1.2 1.1
3.7 pix USM 2.9 2.9 2.8
10 pix USM 3.6 3.4 3.4
Despecle .8 .8 .8
RGB-CMYK 4.3 4.2 4.2
60% Reduction .8 .8 .8
LensFlare 6.1 6.1 6.0
Color Halftone 4.2 4.3 4.3
NTSC Color 4.2 4.2 4.3
Accent Edge 16.3 16.3 16.2
Pointilize 25.1 25.0 25.0
Watercolor 35.6 35.7 35.8
Polar Coordinates 5.1 5.0 4.9
Radial Blur 53.8 53.6 54.3
Lighting 3.5 3.5 3.4

In comparing these results to others from
Ace's Hardware, it seems as thought he 1.6 G5 is roughly about as fast overall to an Athlon 2400+, but if you notice, on some of the benchmarks, the G5 creamed even the P4 3.0Ghz with HT on. So I think with the dual 2Ghz and a lot of RAM, the G5 is going to cream everything that's out their right now. I know I've been doing my fair share of whining on these boards over the last couple of days from poor results during testing, but today after using PSBench7 (BTW, it was recovering the file-going back to the original state-in 1.3 seconds!), I'm fully convinced that the G5 is for real and is a great, great future for the Mac platform.
( Last edited by jrod7350; Aug 26, 2003 at 06:07 PM. )
     
daydreamer
Mac Enthusiast
Join Date: Jan 2001
Location: Belgium
Status: Offline
Reply With Quote
Aug 26, 2003, 06:07 PM
 
Cool, and it's only going to get better from now on, starting with Panther...
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 26, 2003, 06:07 PM
 
Fixed:
Code:
90 Degree Rotate .6 1.1 .4 9 Degree Rotate 3.0 2.8 2.7 .9 Degree Rotate 2.6 2.6 2.8 1 Gaussian .8 .6 .6 3.7 Gaussian 2.5 2.3 2.3 87 Gaussian 3.5 3.3 3.5 1 pix USM 1.3 1.2 1.1 3.7 pix USM 2.9 2.9 2.8 10 pix USM 3.6 3.4 3.4 Despecle .8 .8 .8 RGB-CMYK 4.3 4.2 4.2 60% Reduction .8 .8 .8 LensFlare 6.1 6.1 6.0 Color Halftone 4.2 4.3 4.3 NTSC Color 4.2 4.2 4.3 Accent Edge 16.3 16.3 16.2 Pointilize 25.1 25.0 25.0 Watercolor 35.6 35.7 35.8 Polarize Coordinates 5.1 5.0 4.9 Radial Blur 53.8 53.6 54.3 Lighting 3.5 3.5 3.4
Here is another set of benches, but with only 256 MB and a different file, so it's probably not that useful nor is it comparable.
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 26, 2003, 06:26 PM
 
Ace's numbers vs. the G5 numbers. (I'm assuming it's accurate to compare them.)
Code:
Ath Ath P4 P4 Ath 2xAth G5 Test 3000 2700 3.06 2.8 2200 2200 1.6 1 Rotate 90 0.2 0.2 0.2 0.2 0.3 0.3 0.4 2 Rotate 9 2.8 2.9 2.7 3.3 3.5 2.4 2.7 3 Rotate .9 2.6 2.6 2.5 3.2 3.3 2.2 2.8 4 Gaussian Blur 1 0.8 0.8 0.8 0.9 1.1 1 0.6 5 Gaussian Blur 3.7 3 2.9 2 2.1 3.8 2.4 2.3 6 Gaussian Blur 85 3.4 3.5 2.3 2.6 4.6 2.9 3.5 7 Unsharp 50/1/0 1 1.1 0.9 1.2 1.3 1 1.1 8 Unsharp 50/3/7/0 3.1 3.1 2.1 2.4 4 2.5 2.8 9 Unsharp 50/10/5 3.1 3.1 2.2 2.4 4 2.5 3.4 10 Despeckle 2.6 2.7 2.2 4 3.3 1.9 0.8 11 RGB-CMYK 8.1 8.2 7.3 9.4 10 5.4 4.2 12 Reduce Size 60% 1.2 1.2 0.9 1 1.5 1.1 0.8 13 Lens Flare 3.9 4 2.5 3.8 4.3 3.1 6.0 14 Color Halftone 2.8 3.3 2.2 2.4 4.3 4.5 4.3 15 NTSC Colors 2.3 2.4 2.4 2.6 3.1 3.1 4.3 16 Accented Edges 10.4 10.9 10.9 12 13.7 13.9 16.2 17 Pointillize 17.5 17.7 12.1 18.8 21.3 12 25.0 18 Water Color 22.6 23.6 26.4 29.2 29.4 29.9 35.8 19 Polar Coordinates 8 8.3 7 8.5 10.2 6.1 4.9 20 Radial Blur 46.6 46.9 33.1 43.6 62.7 34.4 54.3 21 Lighting Effects 2 2 1.9 2.1 2.5 2 3.4 Total 148.0 151.4 124.6 155.7 192.2 134.6 179.6
It looks like G5 1.6 performs at about the level of an Athlon 2400+ (1.93 GHz), as jrod7350 said.

It also looks like a P4 3.06 would probably best a single G5 2.0, which would contradict Apple's numbers (although they used a different benchmark setup):



Whatever the case, it's clear that even the "lowly" 1.6 does quite well in Photoshop. This should make a lot of Mac creative types happy. Also, at least for the Athlon, dual CPUs was not such a huge help. It would help for multitasking of course, but indeed, it does seem the G5 1.8 would be adequate for many potential new Power Mac G5 buyers (unless they're already running dual G4 1+ GHz machines or something).
( Last edited by Eug; Aug 26, 2003 at 07:17 PM. )
     
idyll
Senior User
Join Date: Mar 2002
Location: Sarasota, Florida
Status: Offline
Reply With Quote
Aug 26, 2003, 06:32 PM
 
This is what we have been waiting for .
     
squish
Junior Member
Join Date: Jan 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 06:40 PM
 
the G5 was a lot slower on radial blur, water color, pointillize and accented edges. why?
     
Groovy
Mac Enthusiast
Join Date: Apr 2001
Status: Offline
Reply With Quote
Aug 26, 2003, 06:48 PM
 
Originally posted by idyll:
This is what we have been waiting for .
yup and a dual 2Ghz with plenty of ram is going to trounce all those scores
     
Brazuca
Mac Elite
Join Date: Oct 2000
Status: Offline
Reply With Quote
Aug 26, 2003, 06:49 PM
 
Originally posted by Groovy:
yup and a dual 2Ghz with plenty of ram is going to trounce all those scores
Can someone post the results for the dual Xeons?
"It's about time trees did something good insted of just standing there LIKE JERKS!" :)
     
Groovy
Mac Enthusiast
Join Date: Apr 2001
Status: Offline
Reply With Quote
Aug 26, 2003, 06:50 PM
 
Originally posted by squish:
the G5 was a lot slower on radial blur, water color, pointillize and accented edges. why?
as Chris Cox from adobe said the plug in only opts some basic stuff for G5. PS 8.0 is when everything will be full speed.
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 06:53 PM
 
Originally posted by squish:
the G5 was a lot slower on radial blur, water color, pointillize and accented edges. why?
I'm not sure about the first couple, but it still beat the Athlon 2200 on the radial blur by quite a bit, placing it somewhere inbetween an Athlon 2700 and 2200, which for a 1.6Ghz machine isn't bad at all (I was quite pleased). And basically for all the things I use photoshop for (creating graphics, digital photography on hi-res digital photo files), the 1.6 G5 flew. Things like resizing, incremental rotation (to level a picture out), color space conversion (when needed for print), and USM are my bread and butter for digital photography, and those were some of the fastest tests on the 1.6 G5 (even getting close to and beating a 3Ghz P4 with HT), so you can just guess what a dual 2Ghz could do. Yah, it didn't do quite as well on some of those heavy-computation filters, but maybe there's something that photshop isn't taking advantage of yet (I know that I installed the optimizing plugin, but you also have to consider that it's just a small upgrade, Photoshop 8.0 may have some more in store for us, especially under Panther-and even if it doesn't, this is a very respectible score-especially if we're going to 3Ghz in a year and Intel's having trouble reaching 4Ghz due to heat problems).
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 07:13 PM
 
Originally posted by Eug:
It also looks like a single P4 would probably best a single G5 2.0, which would contradict Apple's numbers
At least on PSBench that might be the case. But remember, that was a 45 filter test (you can read what they ran in the white paper on the Apple site). Depending on what size the file was, and what they were doing with the filters, it would only take one or two filters to throw the whole test off in favor of the G5's. I mean a 600MB file is HUGE. Gosh, one RGB-CMYK conversion (the 1.6 was almost twice as fast as the 3Ghz P4) could probably tip the scales in Apple's favor. Also with that size of a file, you're getting a lot of disk-swapping, where the increased bandwidth of the G5's is really going to shine. So I guess it just depends on what you're doing. With PSBench, in many of the tests, the single-proc G5's are comprable to the fastest wintel systems out there. On a 600MB file (not 50MB), that may all change drastically.

I think something else you're also going to have to notice about Apple's testing procedures. They're not really "fudging" scores, since these filters are absolutely essential to the real-world print workflow, and things such as RGB-CMYK cannot be avoided, but you'll notice in the WWDC bakeoff that basically the Xeon got stuck on one operation. Once it completed that operation, it was basically moving just as fast (or a little slower) than the dual G5. The part where Phil Schiller said that the Dell "hadn't crashed". The thing is though, that you do those types of filters over and over again in a given day (such as resizing, reducing size and changing color space (for the web, etc.), and it adds up big time. In the WWDC bakeoff, it meant that the G5 had the file finished before the Xeon got started on the rest of the file. If you could skip that step, then I'm sure the G5 would have only been a couple of seconds faster than the Xeon. But you can't skip these steps in our current print workflow. And so I don't think it's fudging when Apple posts these kind of scores that cause the Xeonto get choked up on one or two tests that happen to be essential to the print workflow. But it can be misleading when you try to apply that to every operation in Photoshop, and expect the G5 to always be twice as fast or whatever. That just won't happen. It's an overall workflow issue, and I think it's nice to actually see that Apple concentrates on that-not just saying that my machine can do a blur faster than yours. Cause the truth is I do WAAAY more than apply blurs all day. I want my WORK to be faster-that's many, many things, not just one or two operations. And if the Xeons can fly on a buch of stuff but choke on one or two things that make my WORKflow crawl to a halt, then there's a definte problem. So I think these are the things to keep in mind when comparing Mac's to PC's in Photoshop. The nice thing now is that the Mac can also keep up and beat wintels in all those "little" filters that it was lagging behind in before.
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 07:16 PM
 
Originally posted by Eug:
(unless they're already running dual G4 1+ GHz machines or something)
Sorry, don't mean to pick on you

But I also PSBenched my dual 1Ghz Mac at work (has 1.5GB of RAM), and the 1.6 G5 ate it for lunch
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 26, 2003, 07:26 PM
 
In the WWDC bakeoff, it meant that the G5 had the file finished before the Xeon got started on the rest of the file. If you could skip that step, then I'm sure the G5 would have only been a couple of seconds faster than the Xeon. But you can't skip these steps in our current print workflow.
Is your experience representative of the general pro Photoshop population? (I'm asking because I only use Photoshop for light editing.)

But I also PSBenched my dual 1Ghz Mac at work (has 1.5GB of RAM), and the 1.6 G5 ate it for lunch
Post the numbers, and I'll add it to the table.

Anyways, what I was suggesting was just that if you already have a dual 1.25 it might make sense just to hang on to it for a while and upgrade to the rev. B G5 2.25 GHz machine or something.

EDIT:

Others are saying that a single G5 1.6 is not that much faster a dual 1 GHz G4 in PSBench. Maybe it feels faster for you (eg. OS responsiveness, hard drive speed, or choice of filters), but for PSBench with adequate memory it may be a different story. If you can post your numbers, I'll add it to the table.
( Last edited by Eug; Aug 26, 2003 at 07:41 PM. )
     
Tiauguinho
Fresh-Faced Recruit
Join Date: Oct 2002
Status: Offline
Reply With Quote
Aug 26, 2003, 07:26 PM
 
Those are some fantastic results! Thanks a lot for posting them jrod7350!

I love my QuickSilver DP 1Ghz but now im seriously considering getting a G5 DP 2Ghz. I need the speed, need the bandwidht and i need a good excuse to get a new Mac!
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 07:46 PM
 
Originally posted by Eug:
Is your experience representative of the general pro Photoshop population?
Well, you can consider when I have 500 photographs (which is typical for a concert shoot), and I have to basically batch resize those for the web, do some color manipulation (levels, etc.), convert their color space from the expanded AdobeRGB to sRGB (for the web again), and possible upres them for large-scale print, converting those large-scale ones into CMYK. You also have to touch up any photos that are particularly important to you. Also you may have to make low-res contact sheets for client approval. The point is, with a couple of files that's no big deal, but with 500 files at a time, that can be a headache. And as you can see from above, most of the processing time is spent resizing and converting-the other amount of time is what I would consider "creative time" where I'm touching up a photograph, and in that case, both platform seem fast enough in general that I'm not waiting around forever. But sometimes on a really large file, you get bogged down again. I mean I don't think it's a mistake that the G5's smoke a Xeon with a 600MB file. That's a BIG file. And when you're creating a photograph for a poster, it's not unusual to have a file that big (especially when you start throwing in layers, even if the file itself isn't huge resolution-wise). So these are some real-world examples to consider that benchmarks like PSBench can't really tell you too much about, but that you quickly learn when you're under fire.
     
ASIMO
Mac Elite
Join Date: May 2002
Location: SoCal
Status: Offline
Reply With Quote
Aug 26, 2003, 07:49 PM
 
Damn speed junkies.

Maybe this is a good time for me to jump in the water. From a G3 Powerbook straight to a G5. I have resisted the future for too long.
I, ASIMO.
     
arn
Junior Member
Join Date: Dec 2001
Status: Offline
Reply With Quote
Aug 26, 2003, 07:51 PM
 
Hope no one minds, I threw the numbers into a "real table" for ease of reading

http://www.chaosmint.com/benchmarks/...c-g5-ps7bench/
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 07:51 PM
 
Originally posted by Eug:
Others are saying that a single G5 1.6 is not that much faster a dual 1 GHz G4 in PSBench. Maybe it feels faster for you (eg. OS responsiveness, hard drive speed, or choice of filters), but for PSBench with adequate memory it may be a different story. If you can post your numbers, I'll add it to the table.
When you start including dual processor-aware operations, then yes the dual's catch up. But things like cloning, painting, smudging, dodging, resizing, color-space conversion, moving layers, don't seem to be dual processor aware, and that's where the 1.6 seems to eat my G4' s for lunch. I guess that's where my perception is a bit skewed, since that's what I normally do with my time in photoshop, and consequently those were the operations I noticed the most "snappiness" in.
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 26, 2003, 08:01 PM
 
Yeah, a lot of the filters don't seem to benefit much from dual CPUs.

With regards to your RGB-CMYK though, it seems that BOTH the G4 and G5s are fast, compared to PCs. Thus, it's not the G5 that is providing the major advantage per se, but PPC in general in this particular case. Furthermore, it seems that duals help a lot here, and indeed the dual 1.25 G4 is significantly faster than the G5 1.6 at this as well (2.9 vs. 4.2). If this is a hugely important action in one's work, then maybe a dual G5 machine is advised for that person.
     
tazznb
Fresh-Faced Recruit
Join Date: Aug 2003
Location: NJ
Status: Offline
Reply With Quote
Aug 26, 2003, 08:36 PM
 
Originally posted by Eug:
Ace's numbers vs. the G5 numbers.
It also looks like a P4 3.06 would probably best a single G5 2.0, which would contradict Apple's numbers (although they used a different benchmark setup):



You have to consider this tiny fact tha may put the dual 2Ghz over the P4 3.06; The FSB increases with the clock speed.... it's not only the clock speed of the CPU increasing.
Huh?
     
clarkgoble
Mac Elite
Join Date: Mar 2001
Location: Provo, UT
Status: Offline
Reply With Quote
Aug 26, 2003, 08:42 PM
 
Originally posted by tazznb:
You have to consider this tiny fact tha may put the dual 2Ghz over the P4 3.06; The FSB increases with the clock speed.... it's not only the clock speed of the CPU increasing.
Photoshop isn't bandwidth limited though. That's why Athalons and Pentiums are often so comparable.

Now if you are doing video now...
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 26, 2003, 08:54 PM
 
Originally posted by tazznb:
You have to consider this tiny fact tha may put the dual 2Ghz over the P4 3.06; The FSB increases with the clock speed.... it's not only the clock speed of the CPU increasing.
True, but the numbers Apple presents definitely show a dramatically different picture than the numbers posted here.

ie.

PSBench
G5 1.6 roughly the speed of an Athlon 2400+.
P4 3.06 GHz destroys a G5 1.6. (G5 is 44% slower.)

Apple bench
G5 1.6 destroys a dual Xeon 3.06. (36% faster)
Dual G5 2.0 is twice as fast as a dual Xeon 3.06.

Note though that a dual G5 2.0 is 2.2X as fast as the baseline PC according to Apple. The single G5 1.6 is 1.5X as fast. Thus, using Apple's numbers are that the dual 2.0 is 47% faster than the single 1.6. This means that if we were to extrapolate the numbers from the benches posted in this thread, the dual 2.0 would only be slightly faster than a P4 3.06. Thus, it makes sense that by these numbers a single 2.0 would be slower than a P4 3.06.

In other words, the new G5s are fast, but you may be in for a big surprise if you actually believe the Apple benchmarks.

I think probably in real life, both a 2.53 GHz P4 and a 1.8 GHz G5 will be damn fine in Photoshop. In some situations the P4 will be faster, and in others, the G5 will be faster. If you prefer OS X then get the G5.

Now, it will be interesting once more bandwidth limited apps enter the picture.
     
christom
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 26, 2003, 11:02 PM
 
Why are you guys ignoring single and dual Opteron workstations? With SSE2, 64 bit opterating systems available now, and the advantage of dual channel DDR and an ondie memory controller it is clear that Opteron is 20-30% faster than Athlon, faster than a 3.2GHz P4 and has no bottlenecks in dual operation unlike the Athlon MP or Xeon MP. Wouldn't this make it a leader in Photoshop performance? I'm certain it would contend, and it has been available for a few months.
     
Eug Wanker
Posting Junkie
Join Date: Jun 2003
Location: Dangling something in the water… of the Arabian Sea
Status: Offline
Reply With Quote
Aug 26, 2003, 11:11 PM
 
Originally posted by christom:
Why are you guys ignoring single and dual Opteron workstations? With SSE2, 64 bit opterating systems available now, and the advantage of dual channel DDR and an ondie memory controller it is clear that Opteron is 20-30% faster than Athlon, faster than a 3.2GHz P4 and has no bottlenecks in dual operation unlike the Athlon MP or Xeon MP. Wouldn't this make it a leader in Photoshop performance? I'm certain it would contend, and it has been available for a few months.
Because we don't have any dual Opterons? If you have one, feel free to do the bench and post the numbers.

And welcome to MacNN.
     
Hydra
Senior User
Join Date: Aug 2001
Location: NY
Status: Offline
Reply With Quote
Aug 26, 2003, 11:37 PM
 
Yes please. Gimme a dual Opteron and I'll bench it til it cries. Dual Opterons are pretty nice machines. In the meantime I'm waiting for my dual G5.

Jerry C.

DIsclaimer:

Any machines given to Jerry C. for the express purpose of performing benchmarks become the property of Jerry C.
     
Spliffdaddy
Posting Junkie
Join Date: Oct 2001
Location: South of the Mason-Dixon line
Status: Offline
Reply With Quote
Aug 26, 2003, 11:51 PM
 
aw hell. let me test a dual Opteron and I'll give it back when I'm done.

     
PantherPPC
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 12:33 AM
 
The Opteron would not be the leader in PS benchmarks. Photoshop is only available for Mac and Windows. Windows doesn't support 64 bit code, making the Opterons power useless. It's a server chip anyway, and I know it can be used in a desktop system, but AMD lists it on their site as a server chip, so it's probably much better suited for Linux anyway.
     
stingerman
Fresh-Faced Recruit
Join Date: Mar 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 12:36 AM
 
The Opteron's would run slower right now since they could only run photoshop in 32-bit compatibility mode which is slower than its native mode. Opteron's are at Microsoft's mercy right now unles you are using them for Linux servers. $1000 / processor is pretty expensive too, so you'll not see one in a dual G5 desktop price range anytime soon; though they may call it a desktop but charge 5K for it.
     
Eug Wanker
Posting Junkie
Join Date: Jun 2003
Location: Dangling something in the water… of the Arabian Sea
Status: Offline
Reply With Quote
Aug 27, 2003, 12:38 AM
 
Hmmm... Popular thread I guess. Three new members specifically for this thread. And another that has come out of lurk mode for his second post in just half a year.
     
Thinine
Mac Elite
Join Date: Jul 2002
Status: Offline
Reply With Quote
Aug 27, 2003, 12:44 AM
 
Remember, Apple's Photoshop benchmarks used a 600MB file. And each filter probably added to that size. In the end, it was the G5's massive bandwidth that let it beat the P4 and Xeon so badly.
     
action snake
Fresh-Faced Recruit
Join Date: Aug 2003
Location: jacksonville, fl
Status: Offline
Reply With Quote
Aug 27, 2003, 01:02 AM
 
the g5's will rule the universe one day.
be patient my fellow mac friends.
my mom thinks i'm cool
     
Brazuca
Mac Elite
Join Date: Oct 2000
Status: Offline
Reply With Quote
Aug 27, 2003, 01:12 AM
 
jrod, can you run SPEC tests on that machine?
"It's about time trees did something good insted of just standing there LIKE JERKS!" :)
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 01:30 AM
 
Originally posted by Brazuca:
jrod, can you run SPEC tests on that machine?
Uh, yeah

Sorry, not to be rude or anything, but you're gonna have to leave that to someone else, cause while I know my way around the terminal and GNU, I have no idea how to compile stuff (I would like to learn though).
     
jrod7350  (op)
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 01:37 AM
 
Originally posted by Thinine:
Remember, Apple's Photoshop benchmarks used a 600MB file. And each filter probably added to that size. In the end, it was the G5's massive bandwidth that let it beat the P4 and Xeon so badly.
I have to second this again too. I repeat from earlier, Apple's test was NOT PSBench. It was not a 50MB file. It was 600MB. There's no way that you can operate on a file like that without disk caching. Unless of course those 2Ghz G4's had the full 4 or 8GB of RAM. Also the I'm not surprised that the 2Ghz OVERALL is only 50% faster, since many things in Photoshop are not dual processor aware. And again, Apple ran different stuff, and a lot more filters (over twice as many) as PSBench does. Those things can add up. So this is really (no pun intended) apples or oranges. PSBench is a good indicator of general day-to-day use. Remember is was also written back in the Photoshop 4.0 days, and when I bought a computer back then, 128MB of RAM was considered a ton of RAM. So I don't think the test has grown with the years, if you know what I mean. Today 50MB is nothing, but 600MB, if you're doing some heavy Photoshop lifting, is not necissarily uncommon. You're going to have to consider the Apple numbers in this light before you outright dismiss them compared to what I benched on PSBench.
     
3.1416
Dedicated MacNNer
Join Date: Feb 2001
Status: Offline
Reply With Quote
Aug 27, 2003, 01:52 AM
 
Originally posted by Groovy:
yup and a dual 2Ghz with plenty of ram is going to trounce all those scores
Hmm, let's try some extrapolation. Making the conservative assumptions that a single 2.0 G5 would be 10% faster than the 1.6, and that a dual G5 will scale the same as the dual Athlon, we get this:

Code:
Ath Ath P4 P4 Ath 2xAth G5 | G5(*) 2xG5(*) Test 3000 2700 3.06 2.8 2200 2200 1.6 | 2.0 2.0 1 Rotate 90 0.2 0.2 0.2 0.2 0.3 0.3 0.4 | 0.4 0.4 2 Rotate 9 2.8 2.9 2.7 3.3 3.5 2.4 2.7 | 2.5 1.7 3 Rotate .9 2.6 2.6 2.5 3.2 3.3 2.2 2.8 | 2.6 1.8 4 Gaussian Blur 1 0.8 0.8 0.8 0.9 1.1 1 0.6 | 0.5 0.5 5 Gaussian Blur 3.7 3 2.9 2 2.1 3.8 2.4 2.3 | 2.1 1.3 6 Gaussian Blur 85 3.4 3.5 2.3 2.6 4.6 2.9 3.5 | 3.2 2.0 7 Unsharp 50/1/0 1 1.1 0.9 1.2 1.3 1 1.1 | 1.0 0.8 8 Unsharp 50/3/7/0 3.1 3.1 2.1 2.4 4 2.5 2.8 | 2.5 1.6 9 Unsharp 50/10/5 3.1 3.1 2.2 2.4 4 2.5 3.4 | 3.1 1.9 10 Despeckle 2.6 2.7 2.2 4 3.3 1.9 0.8 | 0.7 0.4 11 RGB-CMYK 8.1 8.2 7.3 9.4 10 5.4 4.2 | 3.8 2.1 12 Reduce Size 60% 1.2 1.2 0.9 1 1.5 1.1 0.8 | 0.7 0.5 13 Lens Flare 3.9 4 2.5 3.8 4.3 3.1 6.0 | 5.4 3.9 14 Color Halftone 2.8 3.3 2.2 2.4 4.3 4.5 4.3 | 3.9 4.1 15 NTSC Colors 2.3 2.4 2.4 2.6 3.1 3.1 4.3 | 3.9 3.9 16 Accented Edges 10.4 10.9 10.9 12 13.7 13.9 16.2 | 14.6 14.8 17 Pointillize 17.5 17.7 12.1 18.8 21.3 12 25.0 | 22.5 12.7 18 Water Color 22.6 23.6 26.4 29.2 29.4 29.9 35.8 | 32.2 32.8 19 Polar Coordinates 8 8.3 7 8.5 10.2 6.1 4.9 | 4.4 2.6 20 Radial Blur 46.6 46.9 33.1 43.6 62.7 34.4 54.3 | 48.9 26.8 21 Lighting Effects 2 2 1.9 2.1 2.5 2 3.4 | 3.1 2.5 Total 148.0 151.4 124.6 155.7 192.2 134.6 179.6 | 162.0 119.1
Not bad, not great. Slightly faster than the P4, although they're up to 3.2GHz now, and we don't have scores for dual Xeons. Still it's good to be in the same ballpark.
     
klinux
Senior User
Join Date: Jul 2002
Location: LA
Status: Offline
Reply With Quote
Aug 27, 2003, 02:31 AM
 
Originally posted by jrod7350:
But I also PSBenched my dual 1Ghz Mac at work (has 1.5GB of RAM), and the 1.6 G5 ate it for lunch
Yup, I think the whole G5 vs x86 chips will go on for the next 12 months so I look forward to to the whole slew of benchmarks yet to come.

However, I hope people will now be convinced how shitty G4 is (sorry G4 owners!) and its role in Apple's losing of market share. I mean, G3 was competitive, sometimes bettering its x86 counterparts when it was introduced. G5 too is competitive and I bet will better x86 counterparts in certain operations. Sure, G4 is more efficient in some operations but overall, it simply cannot make up for the higher clockspeed and faster buses of its x86 counterparts. Still, so many people were so convinced by the mantra of megahertz myth and cannot admit that G4 by Motorola ultimately was bad for Apple. I mean, if G4 was already vastly superior to its P4/AMD counterparts why the joy over G5?

The fact is G3 could and G5 can compete. G4 cannot.

Okay, rant over. Please continue with G5 discussion.

Disclaimer: I use a iBook/700 and may pick up a cheap G4 i.e. eMac soon!
     
sparkplug
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 03:59 AM
 
2x Xeon 3.06 Ghz 2 GB Ram, 533 Mhz bus,50 MB Test.
Photoshop 7.01

90 degree clockwise rotation | 0.3| 0.3| 0.3|
9 degree clockwise rotation | 2.1| 2.1| 2.1 |
.9 degree clockwise rotation | 2.0 | 2.0 | 2.0 |
1 pixel gaussian blur | 0.6 | 0.6 | 0.6 |
3.7 pixel gaussian blur | 1.3 | 1.3 | 1.3 |
8.5 pixel gaussian blur | 1.6 | 1.6 | 1.6 |
50%, 1 pixel, 0 level unsharp mask | 0.6 | 0.7 | 0.6 |
50%, 3.7 pixel, 1 level unsharp mask | 1.3 | 1.4 | 1.3 |
50%, 10 pixel, 5 level unsharp mask | 1.3 | 1.3 | 1.4 |
despeckle filter | 1.3 | 1.3 | 1.3 |
RGB to CMYK | 5.3 | 5.5 | 5.3 |
60% image reduction | 0.5 | 0.5 | 0.5 |
lens flare filter | 1.4 | 1.4 | 1.4 |
color halftone filter | 2.2 | 2.2 | 2.2 |
NTSC colors filter | 2.9 | 2.9 | 2.9 |
accented edges brushstrokes | 11.8 | 11.8 | 11.7 |
pointillize filter | 6.7 | 6.7 | 6.8 |
watercolor filter | 28.1 | 28.2 | 28.4 |
polar coordinates filter | 3.3 | 3.3 | 3.3 |
radial blur filter | 17.3 | 17.2 | 17.3 |
lighting effects filter | 1.3 | 1.2 | 1.3 |

Total Time: ~93.4 seconds.

PantherPPC writes
----snip----
The Opteron would not be the leader in PS benchmarks. Photoshop is only available for Mac and Windows. Windows doesn't support 64 bit code, making the Opterons power useless. It's a server chip anyway, and I know it can be used in a desktop system, but AMD lists it on their site as a server chip, so it's probably much better suited for Linux anyway.
----snip----

Photoshop is a 32bit app, does this make the G5's power useless??
Last I checked 64bit windows was alive and well. Server chip? AMD's website? funny when I look at AMD's webite I see this.

amd products page

And this

Opteron Workstation Processor

Odd
     
Hash
Mac Elite
Join Date: Apr 2001
Status: Offline
Reply With Quote
Aug 27, 2003, 05:04 AM
 
Sorry to disappoint AMD opteron fans, but Opteron is not something we are discussing here. We are talking about Photoshop test.

AND __ People use Pentium/Athlon PC and Windows as well as Macs and OS 9/X for Photoshop work (probably, majority of pros Macs). I never heard about anyone working on Photoshop on Opteron server. So shut up with your Opteron arguments, they are not relevant here__

End of Opteron discussion
Return to PSBench discussion
     
Shaktai
Mac Elite
Join Date: Jan 2002
Location: Mile High City
Status: Offline
Reply With Quote
Aug 27, 2003, 05:36 AM
 
Originally posted by 3.1416:
Hmm, let's try some extrapolation. Making the conservative assumptions that a single 2.0 G5 would be 10% faster than the 1.6, and that a dual G5 will scale the same as the dual Athlon, we get this:

Code:
Ath Ath P4 P4 Ath 2xAth G5 | G5(*) 2xG5(*) Test 3000 2700 3.06 2.8 2200 2200 1.6 | 2.0 2.0 1 Rotate 90 0.2 0.2 0.2 0.2 0.3 0.3 0.4 | 0.4 0.4 2 Rotate 9 2.8 2.9 2.7 3.3 3.5 2.4 2.7 | 2.5 1.7 3 Rotate .9 2.6 2.6 2.5 3.2 3.3 2.2 2.8 | 2.6 1.8 4 Gaussian Blur 1 0.8 0.8 0.8 0.9 1.1 1 0.6 | 0.5 0.5 5 Gaussian Blur 3.7 3 2.9 2 2.1 3.8 2.4 2.3 | 2.1 1.3 6 Gaussian Blur 85 3.4 3.5 2.3 2.6 4.6 2.9 3.5 | 3.2 2.0 7 Unsharp 50/1/0 1 1.1 0.9 1.2 1.3 1 1.1 | 1.0 0.8 8 Unsharp 50/3/7/0 3.1 3.1 2.1 2.4 4 2.5 2.8 | 2.5 1.6 9 Unsharp 50/10/5 3.1 3.1 2.2 2.4 4 2.5 3.4 | 3.1 1.9 10 Despeckle 2.6 2.7 2.2 4 3.3 1.9 0.8 | 0.7 0.4 11 RGB-CMYK 8.1 8.2 7.3 9.4 10 5.4 4.2 | 3.8 2.1 12 Reduce Size 60% 1.2 1.2 0.9 1 1.5 1.1 0.8 | 0.7 0.5 13 Lens Flare 3.9 4 2.5 3.8 4.3 3.1 6.0 | 5.4 3.9 14 Color Halftone 2.8 3.3 2.2 2.4 4.3 4.5 4.3 | 3.9 4.1 15 NTSC Colors 2.3 2.4 2.4 2.6 3.1 3.1 4.3 | 3.9 3.9 16 Accented Edges 10.4 10.9 10.9 12 13.7 13.9 16.2 | 14.6 14.8 17 Pointillize 17.5 17.7 12.1 18.8 21.3 12 25.0 | 22.5 12.7 18 Water Color 22.6 23.6 26.4 29.2 29.4 29.9 35.8 | 32.2 32.8 19 Polar Coordinates 8 8.3 7 8.5 10.2 6.1 4.9 | 4.4 2.6 20 Radial Blur 46.6 46.9 33.1 43.6 62.7 34.4 54.3 | 48.9 26.8 21 Lighting Effects 2 2 1.9 2.1 2.5 2 3.4 | 3.1 2.5 Total 148.0 151.4 124.6 155.7 192.2 134.6 179.6 | 162.0 119.1
Not bad, not great. Slightly faster than the P4, although they're up to 3.2GHz now, and we don't have scores for dual Xeons. Still it's good to be in the same ballpark.
Those are very conservative extrapolations. About half of what can be realistically expected. Funny thing is some folks in other forums are already quoting those as truth. I'll wait for actual tests on a real dual 2.0.
     
nil
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 05:52 AM
 
Originally posted by PantherPPC:
The Opteron would not be the leader in PS benchmarks. Photoshop is only available for Mac and Windows. Windows doesn't support 64 bit code, making the Opterons power useless. It's a server chip anyway, and I know it can be used in a desktop system, but AMD lists it on their site as a server chip, so it's probably much better suited for Linux anyway.
For months Boxx has been selling DP Opterons that run Win2K Pro and XP Pro. Opterons 32 and 64 bit code natively ergo any DP aware Windows OS will run them just dandy. Same as how the G5 runs non 64 bit OSX.


You're welcome to look here where an Opteron 144 (1.8GHz) goes toe to toe with a 3Ghz P4 and XP3200 running XP Pro. They list the price of the 144 box at $1990.


If you really want to illustrate the Mhz myth, run most any G4 (heck a 700 Athlon might do as well) with
these, nothing like a 400% speed increase.
     
Marco Fanciulli
Fresh-Faced Recruit
Join Date: Aug 2003
Location: Italy
Status: Offline
Reply With Quote
Aug 27, 2003, 07:11 AM
 
Originally posted by squish:
the G5 was a lot slower on radial blur, water color, pointillize and accented edges. why?
Hi, Photoshop and other application developed with G4 and G4's Altivec in mind, have several coding techniques that are strongly inefficient on G5. If you can follow the program execution at low level, you can see that an high number of DST and ST derivative instructions are dispatched and the same can be said about DCBZ. The problem whith DCBZ is that it works on 32 byte chunks while the G5 operates on the entire 128 bytes cache with as much as four times the effort.

There are a number of other underperformant code lines and several branch misprediction which cause a drop down in performances.

You can read more about G5's performance dont's at http://developer.apple.com/technotes/tn/tn2087.html

Marco
     
Chryx
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 07:25 AM
 
Originally posted by sparkplug:

Last I checked 64bit windows was alive and well.
64Bit IA-64 (read 'Itanium') is available, 64Bit x86-64 (read 'Opteron / Athlon64') has, as far as I'm aware, not entered beta stage yet, it certainly isn't available to buy.

so no, 64bit windows, in the context of the Opteron, is NOT alive and well.
     
Chryx
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 07:34 AM
 
Originally posted by Hash:
Sorry to disappoint AMD opteron fans, but Opteron is not something we are discussing here. We are talking about Photoshop test.
And they were talking about how the Opteron would cope with Photoshop..


TBH, the current SSE2 implementation on the Opteron is somewhat weak (they've apparently improved it in the C stepping chips, but I've yet to see one of them benchmarked), and it can't be an x86-64 version of photoshop as of yet, which basically means I would expect it behave like.. well, a similarly clocked AthlonXP with a meg of L2 cache and a memory controller on crack.

AND __ People use Pentium/Athlon PC and Windows as well as Macs and OS 9/X for Photoshop work (probably, majority of pros Macs). I never heard about anyone working on Photoshop on Opteron server.
Opteron based workstations are available... some would argue that they're _as_ available as G5s are (although I'd argue that they aren't, but I'm in the UK, they seem to be fairly widespread in the USA), ergo it's valid to mention them in a discussion of graphics workstation performance.

Now, does anyone have access to an Opteron so they could throw PSBench at it?
     
Bill Harrison
Fresh-Faced Recruit
Join Date: Sep 2002
Status: Offline
Reply With Quote
Aug 27, 2003, 07:36 AM
 
Originally posted by stingerman:
The Opteron's would run slower right now since they could only run photoshop in 32-bit compatibility mode which is slower than its native mode. Opteron's are at Microsoft's mercy right now unles you are using them for Linux servers. $1000 / processor is pretty expensive too, so you'll not see one in a dual G5 desktop price range anytime soon; though they may call it a desktop but charge 5K for it.
This is not true. The Opteron runs 32 bit code faster than an equivilant Athlon XP speed wise. There is no "emulation" and it runs 32 bit apps just as fast as it does 64 bit.
     
Chryx
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 07:47 AM
 
Originally posted by Bill Harrison:
This is not true. The Opteron runs 32 bit code faster than an equivilant Athlon XP speed wise. There is no "emulation" and it runs 32 bit apps just as fast as it does 64 bit.
You're mostly right, x86-64 adds registers, which means that in 64bit mode code needs to hit cache/memory less, which means that an Opteron running 64bit code is faster than it is in 32bit mode.

not because 64bit is inherently faster (because it isn't) but because the implementation of the Opteron is faster on 64bit code.

in 32bit mode it runs the code faster due to its memory controller and extra L2 cache, the core is effectively the same as an AthlonXP if you switch off the nifty stuffs.
     
johneee
Fresh-Faced Recruit
Join Date: Aug 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 08:53 AM
 
I guess what it comes down to is that the top end chips of all the major vendors come out pretty close to the same... Just like it always has, with some shuffling here and there as to who is on top, but nobody really having enough of a lead to make it something that is best for everyone.

It's all just marketing. Bah Humbug I say.
     
videian28
Mac Elite
Join Date: Apr 2003
Location: fredericksburg va
Status: Offline
Reply With Quote
Aug 27, 2003, 09:09 AM
 
someone needs to throw a dual 1.42 into that table
     
legacyb4
Mac Elite
Join Date: May 2001
Location: Vancouver
Status: Offline
Reply With Quote
Aug 27, 2003, 09:21 AM
 
Arn,

Thanks. Having the highlights really makes it easy to get a feel for results.

Cheers.

Originally posted by arn:
Hope no one minds, I threw the numbers into a "real table" for ease of reading

http://www.chaosmint.com/benchmarks/...c-g5-ps7bench/
Macbook (Black) C2D/250GB/3GB | G5/1.6 250GBx2/2.0GB
Free Mobile Ringtone & Games Uploader | Flickr | Twitter
     
skwerrl
Fresh-Faced Recruit
Join Date: Mar 2003
Status: Offline
Reply With Quote
Aug 27, 2003, 11:11 AM
 
noone has even mentioned the gfx cards being used in this thread...

THAT can impact photoshop by a LOT
     
Eug
Clinically Insane
Join Date: Dec 2000
Location: Caught in a web of deceit.
Status: Offline
Reply With Quote
Aug 27, 2003, 11:23 AM
 
2x Xeon 3.06 Ghz 2 GB Ram, 533 Mhz bus,50 MB Test.

Photoshop 7.01Total Time: ~93.4 seconds.
Yeah, I have a tough time believing that a dual 2.0 will reach that Xeon time. I do believe the 119 seconds quoted elsewhere in this thread might be overly conservative though. So, my guess is that dual G5 will perform somewhere in between 93 and 119 seconds. How about a guestimate of 106?

Hi, Photoshop and other application developed with G4 and G4's Altivec in mind, have several coding techniques that are strongly inefficient on G5. If you can follow the program execution at low level, you can see that an high number of DST and ST derivative instructions are dispatched and the same can be said about DCBZ. The problem whith DCBZ is that it works on 32 byte chunks while the G5 operates on the entire 128 bytes cache with as much as four times the effort.

There are a number of other underperformant code lines and several branch misprediction which cause a drop down in performances.
It's not the same as designing filters from the ground up for the G5, but don't forget that the numbers posted are with the G5 optimization plug-in. I do suspect that both X.3 and Photoshop 8 will speed up the G5 even more, but we'll see those speeds when they're out. Right now OS X.2.7 with Photoshop 7.01 and the G5 optimizations will have to do.

One thing I'd be interested in is the numbers with and without the G5 optimizations, on both the G4 and the G5. ie. How much do the optimizations speed up the G5, and how much do they slow down the G4?
     
 
Thread Tools
 
Forum Links
Forum Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Top
Privacy Policy
All times are GMT -4. The time now is 08:42 AM.
All contents of these forums © 1995-2017 MacNN. All rights reserved.
Branding + Design: www.gesamtbild.com
vBulletin v.3.8.8 © 2000-2017, Jelsoft Enterprises Ltd.,