hi,

All Vienna Instruments available from Vienna Symphonic Library were installed on this machine and the Data was spread over two 750 GB internal sATA Drives.

Every half an our the setup was changed from one song to another.
We did not use the technics Nick mentioned but the VSL server was filled with 2,86 GB on RAM.
Some Instruments in the song were optimized to be able to load all Samples needed.

I will give you some more precise testresult with this machine after the weekend.
(max out cpu with VI count, max out HD´s with voices....)

best
chris K

Grazie Maestro [:)]

@Sonrise said:

Is that easy to do? Wouldn't mind knowing the process since Apple is taking their time of Lepoard.

There is a detailed article on how to do this in a recent Virtual Instruments Magazine (along with tons of other informative articles including reviews of VSL.) Nick above could tell you what issue has the standalone/ram article.

http://www.virtualinstrumentsmag.com/

Chris K,

Would it be possible to continue this thread with the update you offered?
Would be greatly appreciated.


basically a macpro in that configuration is equal to the cost of three self assembled shuttle pc's.

Would the macpro equal the performance of the three shuttles?
Looks like it might and thats a trade off I'd go for.

@Maya said:

Hi Pavel_k,

we made quite good experiences with Mac Pro so far. A Mac Pro 2.66 GHz/9 GB RAM with OS 10.4.9 and Logic Pro 7.2.3 was used at NAMM 2007 and Musikmesse Frankfurt at our booth for presentations.

Is it possible to for us to have this Logic file please? I'd like to do some testing on my new dual Quad, it would be interesting to try this arrangement for comparison

@chriskard said:

I will give you some more precise testresult with this machine after the weekend.
(max out cpu with VI count, max out HD´s with voices....)

Hi all,

the Mac Pro was away from the office for a while (for presentations), that´s why we could not yet finish our tests. "Mac is back [:)]" and I´ll proceed with testing and post the results as soon as possible.

Best regards,

Maya

Mac Mini's looks to be discontinued
http://www.appleinsider.com/articles/07/05/24/closing_the_book_on_apples_mac_mini.html

I had been waiting and hoping for the new generation of mini's but alas it looks not to be.

Maya - would it be possible to do some deep testing of your Macpro as a VSL Farm? (as in not also hosting the sequencer - just using it as a slave farm PC).
It looks like the macpro is going to be our only sample farm option if we want to stay mac only.
THANKS!

They're fun to read sometimes, but you have to remember that Apple rumor sites are wrong far more often than they're right.

My fantasy about the type of person who runs a site like that is probably nastier than warranted, but I just can't help picturing this dorky nerd who desperately wants to feel part of his "celebrity" company. You know, the type of person who would stand in line around the block all nght to get into a fricking *Apple Store opening.* Yes, there really are people who do that! It's frightening.

But I am being very nasty, so you should also take anything I say with a grain of salt. [6]

Appleinsider had been more right than wrong for awhile now.
Mostly because the don't go out on any limbs.

macosrumors is nearly dead and mostly wrong.
macrumors makes few predictions themselves as they mostly copy and link any dumb ass thing posted anywhere on the web.

Nick - please stop bestowing us with "nasty" posts and focus on giving us (your subscribers) 12 full issues each year!
The friggin wait in between issues is driving me crazy - especially now that keyboard mag is only about 10 pages cover to cover these days. (I play the b'way pits for a living and the lastest keyboard issue didn't even get me through one damn show - I ended up actually having nothing to do but pay attention the that guy waving that stick).[[:|]]

Okay Ed, we'll do that right away. No problem.

[[;)]]

In all honesty our plans have changed. The first goal is to increase the size of the magazine every other month rather than to go monthly - i.e. to increase the number of advertising pages so we can afford more editorial pages. More than one advertising client has said they they'd prefer that, and it makes a lot of sense.

And it'll happen. We're getting there slowly but surely.

Not all that slowly, as a matter of fact.

Hang in there with it, Nick. VI Mag is a great concept and is wonderfully just ahead of the curve as far as the growing numbers of VI users are concerned. The internet has taken a bite out of the magazine industry, as you know, but I'd rather see VI Mag every other month than not at all.

Keep up the good work!

Thanks JWL, I appreciate it.

But you know, the magazine sells well on the newsstand, so it's probably not as far ahead of the curve as one might think.

@edhamilton said:

Mac Mini's looks to be discontinued
http://www.appleinsider.com/articles/07/05/24/closing_the_book_on_apples_mac_mini.html

I had been waiting and hoping for the new generation of mini's but alas it looks not to be.

Maya - would it be possible to do some deep testing of your Macpro as a VSL Farm? (as in not also hosting the sequencer - just using it as a slave farm PC).
It looks like the macpro is going to be our only sample farm option if we want to stay mac only.
THANKS!

Bump...

Did you already perform a stress test of the 8-core Mac Pro? How many instances of VI can you host with, say, Plogue Bidule? Significantly more than with a quad-core Mac Pro or not...?

From all reports is seems that Ram will limit you before the CPU will.
So until 64 bit osx (and shitloads of ram) it looks like the octocore is overkill for sample farm use.
I haven't heard of anyone maxing out the CPU on a quad core macpro.
Scary.

i didn't have a chance to stress-test an octo core so far (neither mac nor pc, which have pretty much the same design regarding chipsets and other components), but looking at the design it could probably happen that the bus to the processors turns up as bottlenck (it is of the same width as for the quad core)

much if not everything would depend if an application is using data from the L1 or L2 cache of the processor for heavy number crunching or relies on *fresh data* to be transported from the memory to the processor (and back to some other busses like PCI)

in the past practice has often shown that larger caches are affecting performance more than number of processors, because especially with audio applications you have to take a _very_ close look which process can run *threaded* (i.e. use 1,2, more processors) and which not.

i personally am considering 8 cores as overkill currently ...
christian