This one is 1000 bucks :
http://www.hyperossystems.co.uk/07042003/products.htm#hyperosHDIIproduct

But with the SATA bus bottleneck...
Is this bottleneck critical ?
If it is, then, end of the discussion.

When you say that ‘the amount of samples to be streamed parallel is not the issue’, does that mean that for sampling technology it’s the IOPS that maters rather than the bandwidth ? If it is, then the bus bottleneck is not so critical, or I miss something ?

I’m a bit confused about measures.

The actual best seek time for hard drives is around 5 milliseconds (ms), right ?

The seek time for affordable SSDs (my link) is around 50 microseconds (µS) (but with bus bottleneck).

What is exactly the seek time of our actual ‘on mother board’ RAM ? Nanoseconds ?

(Sorry for my ignorance, I can’t find the magic formula to google it...)

thanks for the link to the product, i've ordered (for another purpose) the GigaByte iRAM ...
ok, 64 GB = 4 x 16 GB = 4 X $ 1.000.- (for the board) + 32 x $ 200.- = $ 10.400 (excl. VAT) for a poor-man's SSD (4 free sATA ports needed, system may remain 160 minutes without power, otherwise ... total data loss).
no need to mention for the symphonic cube you'd need 6-7 of them (besides about 50 free sATA ports of course) and i'm not sure this would be very welcome by users who sometimes consider the SC for $ 10.000.- to be expensive ...

technical details: this page says to have tested the hyperdrive on random read (without providing too many details like size of requested data packets) and mentions a latency of 0,2 ms. this would be a factor of 20 compared to the fastest sATA and SCSI drives and so calculating some overhead (but not other system latencies)buffer size could be reduced to 1/16 to about 4 kB. so buffering all 800.000 samples from the SC would need a computer with 3 GB accessible RAM - nice, should work. but - oops - what is this? average CPU usage 31% (on a not specified CPU) with a maximum of 67% ... seems to be unusable for a sampling computer.
the data thoughput is IMO not the real issue, because an average 40 MB/s like with many current harddrives would allow about 230 voices (stereo, already taking the realtime-decompression into consideration) to be streamed.

so all in all - to go this route seems to be not very practicable for me.
christian

And thanks for iRAM , I didn’t know it. Those ‘large public’ products are rare, I can’t find any catalogue.

I’ll buy one as soon as possible for my OS, anyway.


Your projection about RAM to fill up the HyperDrive IV is even optimistic, it needs ECC registered.

But, hey...

@Another User said:

buffer size could be reduced to 1/16 to about 4 kB. so buffering all 800.000 samples from the SC would need a computer with 3 GB accessible RAM

This is the very good part of the demonstration...

Cause the whole point is to (only) buffer, of course. I don’t expect those SSDs to stream, for the VSL, it’s economically absurd, for now, as you show it.

But CPUs usage maters. They look weird. It’s only an HyperDrive II test. Can’t find any HyperDrive IV test...

@Another User said:

This will take a combo of things:

1. Apple will have to make faster machines than they do even now, Mac Pro included

2. Hard drive transfer rates will have to increase as standard, as well as seek times

3. VSL will perhaps find even more ingenious ways of optimizing its software for greater efficiency

4. Leopard and the other 'felines' will have to become a way of life with bit addressing

5. DAW hosts will have to break their RAM access limits

Hi, JWL. ....About RAM limitation, I though buffering on a SSD (Solid State Drive) could be enough (200 times faster than hard drives), and permanent (but as I’ll get ‘no comment’...). There is now even some solutions without bus bottleneck, for industries. Why not for studios ?

Again, this library is special. It cost the price of a car and doesn’t deserve the actual hardware.

Multiple machines solutions (and loading times) are not productive for a pro musician.

I won't disagree with you on this one.

The term "prosumer" still applies to what's publically available in the way of certain features. VSL is indeed a special bundle, so custom upgrades to hardware are in order. No doubt, manufacturers have built systems more powerful than they will release to the general public, and I wouldn't be surprised at all if the cleverest bits of these systems are available to industries.

Then again, I'm reminded of this photo showing 600 G5's networked together for the sake of mastering the Star Wars DVDs!!!
http://www.apple.com/pro/film/lowry/starwars/lowryqtvr.html

@Crystal said:

Multiple machines solutions (and loading times) are not productive for a pro musician.

What makes you sure of such a statement?

Most (if not all) of the "tech-oriented" big-time composers and arrangers (music and film music industry) use multi-computer setups.

Gigastudio is a the best example of that trend.

I am myself right now setting up a studio for my boss where the VI library is being splitted between 8 Mac Mini's. And I know three other composers who will be following us...

Jerome

Crystal, you're trying to find the end of a rainbow! Everyone has wanted a 1-machine solution since the beginning of the digital age. But as soon as one comes, developers take advantage of the extra power and we're chasing the next one.

Two 64-bit operating systems are coming out next year, and as soon as the software we use is adapted, the memory access issues that have caused us to use multiple machines will disappear. But will something else bring the machine to its knees?

@Nick Batzdorf said:

Crystal, you're trying to find the end of a rainbow! Everyone has wanted a 1-machine solution since the beginning of the digital age. But as soon as one comes, developers take advantage of the extra power and we're chasing the next one.

Two 64-bit operating systems are coming out next year, and as soon as the software we use is adapted, the memory access issues that have caused us to use multiple machines will disappear. But will something else bring the machine to its knees?

I hope *my* point wasn't missed. The apparent impracticality seems to be rooted in the fact that demands for one machine always seem to bring it to its knees no matter how fast it is. All too often when a new machine is announced the discussion resumes as to whether The End-All Dream Machine has finally been created. No doubt, networking is going to be here for a while, but perhaps more will be done on fewer machines without today's brick walls.

I'm quite excited about what's in store for 2007 as far as VI users are concerned-- but the techno power chase is never-ending. What distinguishes this discussion from others I've seen is the amount of practical thinking being put forth, which I appreciate.

Practical thinking, and cm's knowledge of the technologies.

But yes, I think I just repeated what you said in different words. It's hard to imagine this dance ending in our lifetimes. One would think that at some point desktop computers should get more powerful than we can use for musical applications, but that seems pretty far off.

Hi Jerome and Nick.

@Another User said:

What makes you sure of such a statement?

Cause I network myself. (No choice...)

More PCs I have, less music I do... Installations, activations, to power and open every thing, to create, name, open, close, save, backup each files, to route midi and audio over the network, to switch monitors, to face hardware and OS maintenances (bugs, ghosts or defrag) etc... etc...
(Plus, networking always adds some latency...)

Well, if you have an assistant, no problem. You’ll make only music...

But, as probably all composers who don’t have any assistant, I’m really more productive with a single PC than with several. For me it’s night and day.



The theory that softwares always need more than hardwares can afford, is not so clear to me.

It is, clearly, for processing. (I don’t care myself for symphonic music, I always process the reverb off-line.) (MIR will be off line anyway...)

But not for recording / playback, specially with sampling library.

Look what appended last years : samplers’ polyphony were restricted, then recently unlimited.

So what was the consequences ? Take Gigastudio, for example. With the best actual PCs, you’ll need five or six of them with GS2.5, instead of one or two with GS3 to get the same result in terms of polyphony.

So, some times limitations are really in the software, not in the hardware.


There is also some limits for us, users. Polyphony again. We can’t ear an unlimited number of voices, with unlimited number of bits and samples rates formats, for physical reasons, simply.
And I think actual DAWs, (except maybe for extreme projects) provide more polyphony than we need to run a decent full orchestra (without processing, of course).

Today, for my use, it means VSL only (+ piano) with off-line processing, the only limitation (to get a full orchestra) that forces me to network is... RAM.

I just did a little orchestral demo with a ‘prehistorical’ PC, it’s a quite light and simple orchestration, so it’s OK like this, but, if I would add more polyphony, no problem, if I would add some instruments, impossible : my RAM is already full...

With a P4 2.6...


For me, the RAM limitation (and its ‘non-productive-at-all’ loading time, not to forget...) is really specific.

Crystal, I totally agree with you re. the RAM problem and the fact that it would of course be better having only one workstation without any slaves.

However, you were making a general statement which I (still) don't think was totally accurate. Once setup properly, you can be very productive with a few slaves, especially if you set them up in a way in which you do not have to access them (they just run in the background with all your samples pre-loaded).

Of course, there's always troubleshooting (especially on a PC running Windows, and even more with Gigastudio), which I agree is a pain and would be simpler if we were only talking about one computer.

However, I know that before I started using Gigastudio in my home studio, I would constantly battle against performances issues on my main sequencer. In that way, Gigastudio saved me a *lot* of time. This also include a very easy and efficient way to work within Logic.

Now, in regard to Gigastudio 2.5 and 3 and their limits in polyphony, sometines the developpers willingly choose to limit themselves because of performances consideration. What's the point of add 64-voice polyphony to your product if - at the time of purchase - most computer won't be able to run it properly?

Plus, it's always better to come up with a new version a few years later, especially when you look at the commercial side of things... [:)]

Jerome

@Another User said:

Plus, it's always better to come up with a new version a few years later, especially when you look at the commercial side of things...

Good point. But I can’t imagine that VSL practices this kind of policy... (Let’s keep this company special...)