Blue Velvet said:
Yeah awesome - also the test on the front page for the MBP benchmarks are done with a free download - check it out i forgot what its called.
either way - this is gaining pace
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Establishing Standard Benchmarks?
- Thread starter arn
- Start date
- Sort by reaction score
Yeah awesome - also the test on the front page for the MBP benchmarks are done with a free download - check it out i forgot what its called.
either way - this is gaining pace
A batch Photoshop Action could be set up as droplets for various Photoshop versions in a folder with some images and zipped for download...
If one of our photographers can supply a nice big high-quality and vaguely interesting TIF I don't mind coming up with a rigorous and demanding set of actions that will bring everybody's Mac to its knees.
If one of our photographers can supply a nice big high-quality and vaguely interesting TIF I don't mind coming up with a rigorous and demanding set of actions that will bring everybody's Mac to its knees.
Blue Velvet said:A batch Photoshop Action could be set up as droplets for various Photoshop versions in a folder with some images and zipped for download...
If one of our photographers can supply a nice big high-quality and vaguely interesting TIF I don't mind coming up with a rigorous and demanding set of actions that will bring everybody's Mac to its knees.
We must be pretty much in sync - Im in the middle of doing it using one of my images - which actions should i include?
My current batch is
1, resize
2, gaussian
3, median
4, layer copy x 5 and merge visible
5, radial
This is a great idea, i thought the original photoshop idea was great too (that large test image that has a radial blur applied to it) These are much more real-world tasks than benchmarks, and, for all intents, processor/OS independent.
You really need to consult an expert on this; you want some tests that tax the whole system (FSB/drives/memory/video) or those that just tax one aspect (CPU like the pshop test)
You really need to consult an expert on this; you want some tests that tax the whole system (FSB/drives/memory/video) or those that just tax one aspect (CPU like the pshop test)
BakedBeans said:
All of them... why not? If you've got that far, then I won't stand in your way.
Blue Velvet said:
anything yu would like to see included? The reason ive put those in is because it seems a broad range of things that can be quite intensive with a large image - also, they are used fairly often.
but anything else you think should be in there?
Depending on the file size, maybe convert to CMYK and then to 16bit before doing a blur, then convert back to RGB and 8bit.
In fact, I would start with a greyscale image so it's easier to download then action to convert to RGB, then apply some Hue/Saturation to colorise it, then some curves to vary the colour across the tonal range, then your set of actions...
It has to be a test that even the Quads won't get through too quickly otherwise it could be outdated within 12 months when the new PowerMacs come out which is why I suggested the CMYK 16bit conversion and blur.
It has to be a test that even the Quads won't get through too quickly otherwise it could be outdated within 12 months when the new PowerMacs come out which is why I suggested the CMYK 16bit conversion and blur.
For the game benchmark. What about using X-Plane? The demo and the full version are the same. The only difference is that if you don't have the CD in it runs in demo mode and puts a 6min cap on the controls.
I know it isn't what people usually think of when they test games but it does push systems a lot if you turn up graphics all the way (they were getting about .5 fps on a dual G5 with highest settings) and has a built in system for displaying and recording system performance.
I know it isn't what people usually think of when they test games but it does push systems a lot if you turn up graphics all the way (they were getting about .5 fps on a dual G5 with highest settings) and has a built in system for displaying and recording system performance.
lol, we are actually on the same page.
ive included an rgb - cmyk already (the image was already 16bit)
this is the action so far (like you i didnt want to make it to easy for quads but it cant be to hard for ibooks/powerbooks)
Image Resize - 2110x3380 Px to 4220x6760 px (40.8m to 163m)
Gaussian Blur - 100px
Mode - RGB to CMYK
Invert
Posterize
Layer x 5 and then merge visible
Desaturate
Unsharp mask
ive included an rgb - cmyk already (the image was already 16bit)
this is the action so far (like you i didnt want to make it to easy for quads but it cant be to hard for ibooks/powerbooks)
Image Resize - 2110x3380 Px to 4220x6760 px (40.8m to 163m)
Gaussian Blur - 100px
Mode - RGB to CMYK
Invert
Posterize
Layer x 5 and then merge visible
Desaturate
Unsharp mask
Blue Velvet said:
Just what I had in mind. Then add to iTunes, right click, convert selection to AAC. Rip to 128kbps error protection off AAC? Thats seems to be a normal bitrate.
Handbrake would be a good one, but currently its not fully optimised for Intel. It really stresses out the processor, and it is easy to get a frames per second. We would need some Video_TS folder for dl I guess though.. As you can't just use a standard DVD.. not everyone would have it.
Looking good people
AIFF/Lossless to AAC/MP3
OK, while in the bath I've had an idea for producing the aiff for an iTunes encoding test and a foolproof way to avoid any copyright problems.
Without going into too much detail, I've got a 24-bit audio editing app here and will do some processing on a generated file using base material that is in the public domain.
Leave it to me... give me a 4-5 days to put it all together in my spare time and get it into a shape that's ready for uploading to somewhere, probably as a .dmg of a folder containing a Lossless file.
It'll take a bit of messing around to achieve a balance between a file of sufficient complexity and dynamic range -- one that gives a meaningful task to a Quad and future Macs -- and yet is not too outrageously large for downloading.
The main thing is that I'm not going to be aiming for a file that will necessarily be tested on the grounds of audio quality and fidelity, rather one that is just chunky enough to pose a challenge.
Watch this space.
OK, while in the bath I've had an idea for producing the aiff for an iTunes encoding test and a foolproof way to avoid any copyright problems.
Without going into too much detail, I've got a 24-bit audio editing app here and will do some processing on a generated file using base material that is in the public domain.
Leave it to me... give me a 4-5 days to put it all together in my spare time and get it into a shape that's ready for uploading to somewhere, probably as a .dmg of a folder containing a Lossless file.
It'll take a bit of messing around to achieve a balance between a file of sufficient complexity and dynamic range -- one that gives a meaningful task to a Quad and future Macs -- and yet is not too outrageously large for downloading.
The main thing is that I'm not going to be aiming for a file that will necessarily be tested on the grounds of audio quality and fidelity, rather one that is just chunky enough to pose a challenge.
Watch this space.
Blue Velvet said:
Um. Sorry but I'm not going to have time to get this done for a while, even though I made a start and only just remembered about it today.
I'm just a time-waster.
I'm supposed to get my DTK replacement iMac today.Tomorrow I'll stick 2 gigs in it.
Thursday evening I'll put it through the paces using all native apps like iLife '06. paying close attention to iMovie and iTunes.
I will not run any tests under Rosetta that involve Altivec translations because most of the apps the currently run under Rosetta will be native by April..
"correct" results will be posted here later Thursday evening..
Thursday evening I'll put it through the paces using all native apps like iLife '06. paying close attention to iMovie and iTunes.
I will not run any tests under Rosetta that involve Altivec translations because most of the apps the currently run under Rosetta will be native by April..
"correct" results will be posted here later Thursday evening..
Although I love the idea of these benchmarks, reading them in threads is going to be a nightmare...
Now what I'd love to see is an affiliated website that allows the user to select their model and config (like Crucial) and state the RAM and graphcs card... then they enter their results...
this means that we have a one-stop shop for all would be purchasers. It needs to be tidy and apple like...
Now I am willing to donate into a kitty required to do this and am also willing to do some of the front end work if I can..
Now what I'd love to see is an affiliated website that allows the user to select their model and config (like Crucial) and state the RAM and graphcs card... then they enter their results...
this means that we have a one-stop shop for all would be purchasers. It needs to be tidy and apple like...
Now I am willing to donate into a kitty required to do this and am also willing to do some of the front end work if I can..
benchmarking hard to measure
To do a fare benchmark then you need decent compilers to build the bench programs. GCC not up to that level. That a performance for processor benchmark.
Or do do we to see who fasts certain apps take to run.
To do a fare benchmark then you need decent compilers to build the bench programs. GCC not up to that level. That a performance for processor benchmark.
Or do do we to see who fasts certain apps take to run.
I actually got interested in the Photoshop benchmark, want to see how big of a difference a seperate scratch disk drive and more ram makes... Working with files that take 22 min to open on a Dual 2.0 G5, so the test should have a pretty big file to it.
Wish I had my G5 here with me, Ill put a reminder to work on a benchmark for you guys though.
Wish I had my G5 here with me, Ill put a reminder to work on a benchmark for you guys though.
Zbrush render please
I'd really like to see a render out of Zbrush. I'd be happy to provide the file. I'm sure there are a few people here who use it, and it would complement the other tests nicely.
As for photoshop images you can't go wrong with hubble. Try this link that has a choice of big files in tiff and jpeg or you can go all the way to an 18000x18000 pixel image here
I'd really like to see a render out of Zbrush. I'd be happy to provide the file. I'm sure there are a few people here who use it, and it would complement the other tests nicely.
As for photoshop images you can't go wrong with hubble. Try this link that has a choice of big files in tiff and jpeg or you can go all the way to an 18000x18000 pixel image here
i'm very interested in a new gaming benchmark. there should be a game engine or something that can utilize the full power of any mac, like intel, g5, g4, and g3, and have no framerate cap. it should also not lean torward nvidia, or ati. unreal tournament 2004 demo's graphics engine is ok, but i'm not sure if it's realiable to test for performance.
do you think porting 3dmark is a good idea ?
any opensource 3d game that we can use?
do you think porting 3dmark is a good idea ?
any opensource 3d game that we can use?
How about this?
http://www.xtremesystems.org/forums/showthread.php?t=84660
Simple, works on Mac and it's cross platform.
Regards,
JM
http://www.xtremesystems.org/forums/showthread.php?t=84660
Simple, works on Mac and it's cross platform.
Regards,
JM