Jump to content

the new mac pro


Recommended Posts

I ran Cinebench R15 on my 4-year-old 17" MacBook Pro and got the following results:

OpenGL 10.72 fps

CPU 136 cb

The OpenGL figure seems in line with Jim's test measurements, the CPU one completely at odds with them.

Looking at the OpenGL result only it looks like I need a new computer!

A very helpful thread, thank you.

Link to comment
  • Replies 82
  • Created
  • Last Reply

Top Posters In This Topic

  • Vectorworks, Inc Employee

For any users that were interested in the benchmarking topic:

I just benched my own machines again with R11 and R15 of Cinebench. The OpenGL metric seems to have stayed very similar, I got an OpenGL score of 36 in r11 and 38 in r15.

However the CPU scores were very different and marked with a new type of unit. r11 scored my iMac at 3.92 "pts" while r15 used a new metric and scored the same machine 344 "cb"

So it appears they have indeed changed how they rate CPUs.

Link to comment

bcd, what was that comment about a mac mini? that is my original question,

get a top of the line mac mini, which i only keep for a year

or

low end mac pro

there is about a $2000 difference and if a mini can get me through another year why spend the extra money..

if i get the mac pro am i going to see $2000 worth of speed?

Link to comment

From what I know (and I may be wrong) is that yes the OSX can be handed off like you say. but that they have opened the door for developers to do the same. and that the system will feed overtaxing data to the graphics cards. The use of GPU in a computer is the only way to massively boost the performance of these workstations. the primary cores feed data to the GPUs, this may take some time for all third parties to write in to code (if truly needed) but this is the same route Cray Research took with the super Computers. And yes I know this is not a Super Computer but years ago when I was at Cray the Y-MP (was round just ALOT BIGGER) had 2,4 or 8 processors and a performance of 333 megaflops and max 512 MB of ram this was 1992 and cost $15,000,000.00. In 2002 the X1 would process to 50 teraflops and be sold in excess of $50,000,000.00 and is still used today.

Here is a system that can process 7 TF at a price (guessing as it is not released) of around $10,000.00. Apple has moved from desktop as we know it today to workstations that are designed to be workhorses. And the implementation of GPU processing takes up the gap in speed improvements that the chip manufacturers(intel) are lacking on, as the overall processing of the Intels have not gained that much lately. And yes the dual GPUs should work with Renderworks as they are bridged and process jointly.

Sorry for the long rambling, I am quite excited as I saw years ago Cray doing unbelievable things and now I get a "Mini Cray" just fun!

Link to comment
And yes the dual GPUs should work with Renderworks as they are bridged and process jointly.

Isn't Renderworks good old Intel CPU based rendering, albeit multithreaded, so graphics card isn't doing much other than displaying the rendered pixels?

OpenGL there will be a difference, but that isn't Renderworks.

That's my understanding.

Link to comment

ok ok ok ok?

what would really be useful is this:

1. the vw web site have a file that we can all download.

this file has a lot of renderings and section vps

2. we open it and do "up date all viewports"

this file has a way of recording the time it takes and collects some computer info

3. this info is sent back to the web site and listed for all to see.

NOW we have a common reference point for something real. we can all compare our render time to everyone else.

so instead of "i got a Cinebench # of this or that" we can see that you took 5 min to render the file, a file that i am familiar with, and it took me 10 min.

this is what we would see:

file test 1

5min 45sec

4 cores, 8 threads 2.70 GHz

Intel Core i7 - 3820QM CPU

Nvidia GeForce GT 650M

OS X 10.9.0

THAT'S what we need..

Link to comment

then we can calculate the value for our own offices.

like this:

20 hr work week (modest)

times 50 weeks (2 weeks vacation)

hourly rate $35/hr (modest)

$2000 difference in mac mini & mac pro

so 20hrs x 50wks = 1000hr for the year of working with vw

$2000 / 1000hrs = 2hrs per 20hr work week

so if i spend the extra $2000 will the mac pro shave off 2hrs of work? updating viewports each week?

2hrs = 120min

so 120min / 20hrs = 6min per hour

so this 6min per hour is a number that an office can really use in making a judgment in how much to spend on hardware.

Link to comment
  • Vectorworks, Inc Employee
ok ok ok ok?

what would really be useful is this:

1. the vw web site have a file that we can all download.

this file has a lot of renderings and section vps

2. we open it and do "up date all viewports"

this file has a way of recording the time it takes and collects some computer info

3. this info is sent back to the web site and listed for all to see.

NOW we have a common reference point for something real. we can all compare our render time to everyone else.

This would be great for rendering, but the other VW functions are just as (if not more) important. Since most VW functions are not multithreaded, they would most likely benefit from different hardware configurations than Renderworks does (faster processors vs more threads). For many of us, we want things as snappy and fluid as possible while we're doing the actual work. Things like waiting for complex PIO's to regenerate (or open their settings dialogs) interrupt the workflow much more than Renderworks speed.

So maybe this "benchmark file" we download could also have a way to test many other VW functions. The information this would provide users could save users hundreds or thousands of dollars in making hardware purchases. It could even prompt many users to hold off on new hardware if the speed difference is only minimal.

Link to comment
  • Vectorworks, Inc Employee
And yes the dual GPUs should work with Renderworks as they are bridged and process jointly.

Isn't Renderworks good old Intel CPU based rendering, albeit multithreaded, so graphics card isn't doing much other than displaying the rendered pixels?

OpenGL there will be a difference, but that isn't Renderworks.

That's my understanding.

Correct, video cards will currently only process the OpenGL rendering mode and a bit of wireframe/hidden line. The video card will not currently affect any of the full Renderworks rendering modes.

Link to comment
  • Vectorworks, Inc Employee
ok ok ok ok?

what would really be useful is this:

1. the vw web site have a file that we can all download.

this file has a lot of renderings and section vps

2. we open it and do "up date all viewports"

this file has a way of recording the time it takes and collects some computer info

3. this info is sent back to the web site and listed for all to see.

NOW we have a common reference point for something real. we can all compare our render time to everyone else.

so instead of "i got a Cinebench # of this or that" we can see that you took 5 min to render the file, a file that i am familiar with, and it took me 10 min.

this is what we would see:

file test 1

5min 45sec

4 cores, 8 threads 2.70 GHz

Intel Core i7 - 3820QM CPU

Nvidia GeForce GT 650M

OS X 10.9.0

THAT'S what we need..

The problem I see with this is that you would have to fully install Vectorworks on the machine in order to benchmark it in this manner. So you couldn't simply go around benching multiple machines, since you would rapidly run into the max activations limit installing Vectorworks on so many machines.

That system would work great for judging machines that already had Vectorworks on them, but would not be so useful for testing prospective new computers.

I was attempting to devise a procedure that would allow a user to drop a small utility or two onto a machine and test it, without the entire process of installing Vectorworks itself.

Edited by JimW
Link to comment
  • Vectorworks, Inc Employee

Check out how maxwell does it with a test file and a nice post page.

http://www.maxwellrender.com/index.php/benchwell

One thing I find useful with this method is to confirm if my system is at least running at par with others.

Excellent! This is exactly what I am already working on here with all of the hardware available to me.

Link to comment

Jim W, I understand the issues with full VW on a machine but with as many users out there and being able to benchmark computers would aide in many users purchasing decisions. What I mean is in my case I would like to go to a new mac pro, I will take the chance that it is a superior machine and spend the money. I will then benchmark this machine and post. Then when John Doe (imaginary person sorry if your name is John Doe) is looking to buy a new machine, he can look at the posted computers and make his selection based on his workflow and budget. I just spoke with a user that renders custom furniture pieces and does not need a workhorse of a computer so he can find the computer that gives him the results he wants at a price he wants. For me if I can spent $2000 more and save 10-20 min a day then money well spent.

Dont look at it as testing all new models but a general benchmark for users by users to make better decisions for their business.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...