I want to know if will have x64 dirivers only (working in 32bit emulation) or if will have a x64 native software (not 32bit emulated) or if simply not support x64.
32 bit OSs can use only 3.2GB (a portion of 4GB is used to index hardware), today this is a big limit:
Try to start SFP, CubaseSX 3, virtual instruments (as Toontrack Drum kit from hell)... The system will be unstable and probably will slow down for swapping memory to disks!!!
Thanks,
Andrea
Does version 5 will support x64 computing?
-
chriskorff
- Posts: 371
- Joined: Sun Nov 26, 2006 4:09 am
32 bit is a limitation ? 3 GB isn't enough ? it's unstable ?
LMAO
I've seen high class publications done on 33 MHZ machines in 32 Mbyte of Ram - and you bet it was not that slow as the figures might suggest
And what do I see today ?
Pictures arriving via Digicam as 100 MByte raw data each, but some cannot even compete with a 16 MByte high class Scan that's 10 years old.
In other words 80% of the data are effectively noise
Do you really think it's different in music production ?
The machine I mentioned above has been used to house Protools 10 years ago - it had a meager 1% of the resources a regular DAW has today.
Now you're up to show me the better results - or the dramatic loss in quality of old Protools recordings made on a 16(audio)bit system.
it's a matter of fact that there's such an amount of bugs in operating systems that they cannot be fixed anymore - they are only shifted around from one release to the other.
it's also a matter of fact that for cost efficiency the major part of (large scale) developement today is fully automated - resulting in slow and inefficient code.
it's no wonder at all that things are kind of 'unstable', often because programs allocate mad amounts of memory for the 'convenience' of the programmer - or his lazyness, or stupidity... whatever you prefer.
I've recently seen an office word processor (on Mac OSX) working on a 50 pages table that ran in 200-300MB of physical memory allocating 3 GB of virtual stuff - and all of this for 800 kilo-Bytes of data.
64bit adressing is perfectly ok for internet and database servers - for user processing it's plain bullshit.
cheers, Tom
LMAO
I've seen high class publications done on 33 MHZ machines in 32 Mbyte of Ram - and you bet it was not that slow as the figures might suggest
And what do I see today ?
Pictures arriving via Digicam as 100 MByte raw data each, but some cannot even compete with a 16 MByte high class Scan that's 10 years old.
In other words 80% of the data are effectively noise
Do you really think it's different in music production ?
The machine I mentioned above has been used to house Protools 10 years ago - it had a meager 1% of the resources a regular DAW has today.
Now you're up to show me the better results - or the dramatic loss in quality of old Protools recordings made on a 16(audio)bit system.
it's a matter of fact that there's such an amount of bugs in operating systems that they cannot be fixed anymore - they are only shifted around from one release to the other.
it's also a matter of fact that for cost efficiency the major part of (large scale) developement today is fully automated - resulting in slow and inefficient code.
it's no wonder at all that things are kind of 'unstable', often because programs allocate mad amounts of memory for the 'convenience' of the programmer - or his lazyness, or stupidity... whatever you prefer.
I've recently seen an office word processor (on Mac OSX) working on a 50 pages table that ran in 200-300MB of physical memory allocating 3 GB of virtual stuff - and all of this for 800 kilo-Bytes of data.
64bit adressing is perfectly ok for internet and database servers - for user processing it's plain bullshit.
cheers, Tom
-
chriskorff
- Posts: 371
- Joined: Sun Nov 26, 2006 4:09 am
Hmm... 64bit computing (and, more importantly, the inherent ability to access larger amounts of RAM) has its place in DAWs — sample libraries aren't getting any smaller! For large-scale productions with über-libraries at high sample rates & bit depths (eg. film scoring), the ability to access more than 3gb RAM will become more and more important, I think.
IMHO
Cheers,
Chris
IMHO
Cheers,
Chris
Yet we have managed to cope very well so far using limited 32bit machines. I have to agree with Astroman on this one. People have and some are still making music with Atari's. Most studios, ie commercial music production, pop chart music... they have no need to have such an over powered super computer to make music. I even doubt, although this is just a blind thought of mine, that film production studios dont have need either for example a 16 gig octocore machines. Hard disk's are fast enough to meet your demmand.chriskorff wrote:Hmm... 64bit computing (and, more importantly, the inherent ability to access larger amounts of RAM) has its place in DAWs — sample libraries aren't getting any smaller! For large-scale productions with über-libraries at high sample rates & bit depths (eg. film scoring), the ability to access more than 3gb RAM will become more and more important, I think.
IMHO
Cheers,
Chris
What I have learned is to take a more pragmatic approach to music, more does not mean better, specially with something that you must be creative in.
cheers.
not at all - I've already posted my opinion about this specific topic.
There's no need to have all sounds loaded in Ram - dunno if the Gigasampler patent may stand in the way for others...
Giga has the attack phase of a sound in Ram and can fetch the 'decay' even over a network(!) - I've tried that with Gigapiano on a 300MHZ Celeron years ago and it worked perfectly. The machine had 64MB of ram...
The technology is patented, now hold by Tascam, and the original program was coded in X86 assembly. Maybe they are not as smart today...
cheers, Tom
There's no need to have all sounds loaded in Ram - dunno if the Gigasampler patent may stand in the way for others...
Giga has the attack phase of a sound in Ram and can fetch the 'decay' even over a network(!) - I've tried that with Gigapiano on a 300MHZ Celeron years ago and it worked perfectly. The machine had 64MB of ram...
The technology is patented, now hold by Tascam, and the original program was coded in X86 assembly. Maybe they are not as smart today...
cheers, Tom
ok, this is personal opinions.
But in some cases large amounts of RAM is really important.
One of this cases could it be film scoring, another could be video promotion productions, another could be fast-productions related to broadcasting (where people works a lot with sample libraries)....
Keep in mind that RAM acces is 0.3ms, HD access is 15ms...
The world is not the same for all!!
Bye, and let me hope for 64bit support
Andrea
But in some cases large amounts of RAM is really important.
One of this cases could it be film scoring, another could be video promotion productions, another could be fast-productions related to broadcasting (where people works a lot with sample libraries)....
Keep in mind that RAM acces is 0.3ms, HD access is 15ms...
The world is not the same for all!!
Bye, and let me hope for 64bit support
Andrea
I read this on microsoft web site.
"On a 64-bit platform, for optimal performance, all PCI adapters (including 32-bit PCI adapters) must be able to address the full physical address space. For 32-bit PCI adapters, this means that they must be able to support the Dual Address Cycle (DAC) command to permit them to transfer 64-bit addresses to the adapter or device (that is, addresses above the 4 GB address space). Adapters that cannot provide this support cannot directly access the full address space on a 64-bit platform."
in that case i would think that if the old cards do not have this "dac" then it will never work. but who knows!
also remember that creative had a hell of a time getting stable drivers for xp64, and they have a few more people working there.
it may even mean that using windows 2003 with PAE might not work either.
i would just get an RME or something and run sequencer/sampler stuff on a seperate XP 64 bit machine connected to scope machine with MIDI and ADAT or something..
"On a 64-bit platform, for optimal performance, all PCI adapters (including 32-bit PCI adapters) must be able to address the full physical address space. For 32-bit PCI adapters, this means that they must be able to support the Dual Address Cycle (DAC) command to permit them to transfer 64-bit addresses to the adapter or device (that is, addresses above the 4 GB address space). Adapters that cannot provide this support cannot directly access the full address space on a 64-bit platform."
in that case i would think that if the old cards do not have this "dac" then it will never work. but who knows!
also remember that creative had a hell of a time getting stable drivers for xp64, and they have a few more people working there.
it may even mean that using windows 2003 with PAE might not work either.
i would just get an RME or something and run sequencer/sampler stuff on a seperate XP 64 bit machine connected to scope machine with MIDI and ADAT or something..
-
chriskorff
- Posts: 371
- Joined: Sun Nov 26, 2006 4:09 am
Interesting...
Limited knowledge of PCI busses notwithstanding:
The processing power on a DSP card will never be improved by the operating system it's being used on, so that doesn't seem to mean anything — what would be most important, I'm guessing, is for SFP to be able to work on a 64bit computer (even in 32bit emulation), while other programs (like Gigastudio) can run alongside it, and be able to access the oodles of RAM that your 64bit OS can access... is that right?
I'm prepared to be corrected here...
Cheers!
Chris
Limited knowledge of PCI busses notwithstanding:
The processing power on a DSP card will never be improved by the operating system it's being used on, so that doesn't seem to mean anything — what would be most important, I'm guessing, is for SFP to be able to work on a 64bit computer (even in 32bit emulation), while other programs (like Gigastudio) can run alongside it, and be able to access the oodles of RAM that your 64bit OS can access... is that right?
I'm prepared to be corrected here...
Cheers!
Chris
-
chriskorff
- Posts: 371
- Joined: Sun Nov 26, 2006 4:09 am
How much RAM do Scope cards need anyway? I think the most important thing is that SFP can run on an OS that allows OTHER software to access the full address space... no?Neutron wrote:IAdapters that cannot provide this support cannot directly access the full address space on a 64-bit platform."
Scope routing & DSP effects on a computer that can load ridiculous amounts of sample libraries in RAM is the main thing, I'd have thought.
Does the paragraph above suggest that 32bit cards CANNOT run on Vista 64, or just that they won't get direct access to 'the full address space' (which I take to mean 128gb RAM, or whatever the theoretical limit is)?
since that seems to be adressed at my statement above:burdello wrote:ok, this is personal opinions.
no, it isn't - it is experience
I've sold, serviced and supported such machines for over 20 years.
Some of our customers were industrial market leaders in their domain and one is among the largest suppliers of print services in Europe.
I make a significant part of my living from developing custom database applications and I have worked for years with Oracle RDBMS
didn't I write that my tiny Windoze box 5 or 7 years ago with just 10% of the clockrate of todays systems could handle a 1 GB piano in (say) 24MB of memory ?...Keep in mind that RAM acces is 0.3ms, HD access is 15ms...
The world is not the same for all!!
(the whole machine had only 64 MB and the sample was streamed from a network disk via 10 MBit network, using a hardisk spinning at 4200 rpm at best)
The ability to use network servers for huge centralized sample libraries was one of the core aruments for Gigasampler and I verified their statement with the most humble system available.
your just waffling about numbers with not the slightest idea about data organisation and access strategies.
I have seen databases in RAM and in RAM Disks that were outperfomed by an Oracle system running from a traditional harddisk - for the simple fact that the latter was programmed classes smarter.
I'm not talking about an opinion - I report as a witness.
you're welcome to tell us a bit about your experiences with huge film scores and the respective problems.
You've probably considered the flipside of the story that in 64bit adressing anything below 4 GB takes twice the amount of resources and time to adjust numers, didn't you ?
cheers, Tom
Hi,
You need XP SP1 and more than 4Gb
To understand, XP addressable space is 4Gb and memory mapped I/O fall in that range.
When you enable PAE, it allow to reallocate this address space outside of the first 4 Gb memory address space in order to have a single block of free memory which can be addressed by window.
For the PCI card to access these extended address (36bits), they need the DAC thingy
cheers
Michel
You need XP SP1 and more than 4Gb
To understand, XP addressable space is 4Gb and memory mapped I/O fall in that range.
When you enable PAE, it allow to reallocate this address space outside of the first 4 Gb memory address space in order to have a single block of free memory which can be addressed by window.
For the PCI card to access these extended address (36bits), they need the DAC thingy
cheers
Michel
I simply asked: Does version 5 will support x64 computing?astroman wrote:since that seems to be adressed at my statement above:burdello wrote:ok, this is personal opinions.
no, it isn't - it is experience
I've sold, serviced and supported such machines for over 20 years.
Some of our customers were industrial market leaders in their domain and one is among the largest suppliers of print services in Europe.
I make a significant part of my living from developing custom database applications and I have worked for years with Oracle RDBMS
didn't I write that my tiny Windoze box 5 or 7 years ago with just 10% of the clockrate of todays systems could handle a 1 GB piano in (say) 24MB of memory ?...Keep in mind that RAM acces is 0.3ms, HD access is 15ms...
The world is not the same for all!!
(the whole machine had only 64 MB and the sample was streamed from a network disk via 10 MBit network, using a hardisk spinning at 4200 rpm at best)
The ability to use network servers for huge centralized sample libraries was one of the core aruments for Gigasampler and I verified their statement with the most humble system available.
your just waffling about numbers with not the slightest idea about data organisation and access strategies.
I have seen databases in RAM and in RAM Disks that were outperfomed by an Oracle system running from a traditional harddisk - for the simple fact that the latter was programmed classes smarter.
I'm not talking about an opinion - I report as a witness.
you're welcome to tell us a bit about your experiences with huge film scores and the respective problems.
You've probably considered the flipside of the story that in 64bit adressing anything below 4 GB takes twice the amount of resources and time to adjust numers, didn't you ?
cheers, Tom
You say "I don't need it and I don't want it" and I say "I need it, I hope for it"
Where is the problem?
Ok, I understand you, don't take so personal
Sincerely, this thread is becoming a flame but this is not in my intentions.
Sorry Tom,
bye
sorry, you did not simply ask
your reasoning ...32 bit OSs can use only 3.2GB (a portion of 4GB is used to index hardware), today this is a big limit: ... is just plain wrong
I only commented it because it's also the argumenent of (a specfic part of) the industry - spread by marketing departments in the first place.
There is already more than enough hype confusing people every day.
You might have written '... for me its a limitation because I do this and that..., but instead you even assume that ...a limited adress space is the source of instability... which again is just wrong.
Instability is the consequence of bad programming and nothing but bad programming.
Again I have seen the contrary of your 'hypothesis' myself when running programms that were larger than available memory on dataset that also exceeded the amount of free memory - according to your wisdom a 'double no-go'... maybe not the fastest thing on earth, but it really worked.
This is not 'personal' or flaming in any way at all - yet a few sarcastic sidenotes are appropriate due to the annoyance level those marketing bubbles bring with them.
At least I can provide (realworld) examples I've experienced myself for all of my arguments - where are yours ?
If you want to buy into 64 bit for your own entertainment - just do it
if you prefer a truck to fetch a sixpack - do it
but don't suggest that 64bit is improving or speeding up anything below 4 GB, which 99.5 % of all user projects are
cheers, Tom
your reasoning ...32 bit OSs can use only 3.2GB (a portion of 4GB is used to index hardware), today this is a big limit: ... is just plain wrong
I only commented it because it's also the argumenent of (a specfic part of) the industry - spread by marketing departments in the first place.
There is already more than enough hype confusing people every day.
You might have written '... for me its a limitation because I do this and that..., but instead you even assume that ...a limited adress space is the source of instability... which again is just wrong.
Instability is the consequence of bad programming and nothing but bad programming.
Again I have seen the contrary of your 'hypothesis' myself when running programms that were larger than available memory on dataset that also exceeded the amount of free memory - according to your wisdom a 'double no-go'... maybe not the fastest thing on earth, but it really worked.
This is not 'personal' or flaming in any way at all - yet a few sarcastic sidenotes are appropriate due to the annoyance level those marketing bubbles bring with them.
At least I can provide (realworld) examples I've experienced myself for all of my arguments - where are yours ?
If you want to buy into 64 bit for your own entertainment - just do it
if you prefer a truck to fetch a sixpack - do it
but don't suggest that 64bit is improving or speeding up anything below 4 GB, which 99.5 % of all user projects are
cheers, Tom
I installed a new board with my cards today and when i chacked how much memory was left of the 4 GB it said it had PAE on, even though i did not set it?MD69 wrote:Hi,
This Dac can be easily checked!
Find a XP SP1 (SP2 locked the option for PAE as it was the main source of problems with extended address space) and enable /3Gb and /PAE if it works and it relocate the IO address space, then your cards are compatibles.
cheers
Michel
Re: Does version 5 will support x64 computing?
wish i could have a faster pc that save timebordello wrote: 32 bit Ross can use only 3.2GB (a portion of 4GB is used to index hardware); today this is a big limit:
Try to start SFP, Ceases 3, virtual instruments (as Too track Drum kit from hell)... The system will be unstable and probably will slow down for swapping memory to disks!!!
Thanks,
Andrea
one of my friend work with 3d, animation etc and they switch to xp64 and saying they found 4GB RAM working... i was thinking if xp64 could help me.... with such a ram I’ll become lazy
perhaps jimmy could join this topic and share how he's working with 4gb ram in live.
thanks
Ok, I agreee with you, but now, loading big sample VSTi instruments in Cubase or Nuendo mean 2GB maximum because they run all in a single application, So, to enable 3GB for Cubase you can insert the "/3GB" in the boot.ini file. In this case, when Cubase exceed the memory used by the OS (which can allocate only 3.2 GB) XP goes in blue screen or Cubase crash. Perhaps the real limit I mean Isn't directly 3.2 GB but 2GB.astroman wrote:sorry, you did not simply ask
your reasoning ...32 bit OSs can use only 3.2GB (a portion of 4GB is used to index hardware), today this is a big limit: ... is just plain wrong
I only commented it because it's also the argumenent of (a specfic part of) the industry - spread by marketing departments in the first place.
There is already more than enough hype confusing people every day.
You might have written '... for me its a limitation because I do this and that..., but instead you even assume that ...a limited adress space is the source of instability... which again is just wrong.
Instability is the consequence of bad programming and nothing but bad programming.
Again I have seen the contrary of your 'hypothesis' myself when running programms that were larger than available memory on dataset that also exceeded the amount of free memory - according to your wisdom a 'double no-go'... maybe not the fastest thing on earth, but it really worked.
This is not 'personal' or flaming in any way at all - yet a few sarcastic sidenotes are appropriate due to the annoyance level those marketing bubbles bring with them.
At least I can provide (realworld) examples I've experienced myself for all of my arguments - where are yours ?
If you want to buy into 64 bit for your own entertainment - just do it
if you prefer a truck to fetch a sixpack - do it
but don't suggest that 64bit is improving or speeding up anything below 4 GB, which 99.5 % of all user projects are
cheers, Tom
If you install Cubase SX 3.1 in XP x64, Cubase take advantage of bigger memory allocation. My friend, wich have a Motu Ultralite on XP x64, can play live on an electronic drum kit the VSTi "Drumkit From Hell" in full setup, loading all samples in memory (more than 2GB!!) for a near to zero latency.
This is one case, but there are many more, in example for video and music editing, I don't dubt about your experience Tom, sorry for misundersandings, good day and thanks for your replies.