Home |
Search |
Today's Posts |
#41
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Laurence Payne" lpayne1NOSPAM@dslDOTpipexDOTcom wrote in message
... On Tue, 20 Mar 2007 20:58:08 +0200, "DeeAa" wrote: You get the picture why I'm constantly busy getting to the next task at hand ;-) ? Well, I get a fair picture of a butterfly mind who obsesses on saving a few seconds here and there but will happily waste a whole evening rebuilding his computer system. (His? Or am I talking to a woman? This smells rather of that female "Look at me! I'm SOOOO busy!" thing :-) I'm a dude...and I'm NOT happy about rebuilding my system once again as it works fine. But what can I do...this is an AGP based system and I've used all options for upgrading it. I can swap the innards for (only) slightly better but Pci-e system for free, as my friend is upgrading his machine and he'll get the same money whether he gives my stuff in trade or his own - and if I swap parts with him, I can again upgrade with the newest displaycards and memory etc. and not have to swap the whole system at once the next time I need to do that. Believe me, I've been trying to avoid it but it's a sweet deal. I get a better platform for free, and can finally get a decent displaycard for games without having to spend $1000 at once. BUT I just made a test...in fact the programs are loaded much faster all at once rather than separately. I opened all the everyday-use programs I have on the QuickLaunch bar as quickly as I could click 'em - 20 of them including Cubase and Wavelab - and it took 1 minute 26 seconds till the disk light went out again and all were running. Even before that I would have been able to write at least one or two quick email replies or start d/l:lling headers etc. on the programs that had completed loading already. I then tried to give the programs time to load properly, but also tried to launch the next the same instant the previous one finished loading, and I could not get them all launched under 2 minutes even though they had just been launched and the all-at once test was just after a cold boot, even if I hastened it a bit and launched the next one a bit early. So it does work faster to open programs simultaneously :-) Probably would not on a single-drive system but I have 3 separate zero-raid setups on this... What soundcard? Is "Release driver in background" selected in Cubase? Yep, but it doesn't help. If I try to click anything outside Cubase, the playback freezes right away. If I proceed to load another audio program it complains about the driver OR says nothing, but won't give any audio. It did that both with Echo Layla and now MAudio Firewire 410. |
#42
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Laurence Payne wrote:
On Tue, 20 Mar 2007 10:24:24 -0400, Jim Gilliland wrote: I'm not complaining. This system is fast enough. g But if I wanted to make it faster, I'd have to speed up the disk - probably adding another drive or two to the RAID. Adding more memory or a faster CPU wouldn't give me much benefit. We'll have to agree to disagree about that :-) Really? Sorry, but I can't even imagine what the disagreement would be about. Are you saying that you doubt my word on this? If so, well, don't! vbg You can tell me all you want about how your system behaves, but it wouldn't make much sense for you to try to tell me about mine. Not when I'm here observing it firsthand! Perhaps I'm misinterpreting your remark - if so, I apologize. |
#43
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Tobiah wrote:
The bottom line is that more memory will always help to some degree, because it reduces the amount of redundant disk reads. I don't know how Windows handles this, but if it is not in a similar way, then it is quite lacking in this regard. Perhaps, but that kind of caching provides almost zero benefit in audio processing. Sequential reads of long files don't really take much advantage of a cache. Especially when you're processing a dozen or two of them concurrently. |
#44
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Laurence Payne wrote:
I thought you said a faster processor WOULDN'T make any great difference? Laurence, I think you're getting your correspondents mixed up. g |
#45
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Laurence Payne" lpayne1NOSPAM@dslDOTpipexDOTcom wrote in
message On Tue, 20 Mar 2007 10:24:24 -0400, Jim Gilliland wrote: Sure it does. That two hour concert that I told you about takes around 30 minutes to render into a stereo mix. On last year's machine, it would have taken about 60. And the difference between last year's machine and this year's is solely a faster disk system? :-) No, I updated most of the system. But I/O was the limiting factor on both. I'm not complaining. This system is fast enough. g But if I wanted to make it faster, I'd have to speed up the disk - probably adding another drive or two to the RAID. Adding more memory or a faster CPU wouldn't give me much benefit. We'll have to agree to disagree about that :-) Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. |
#46
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Jim Gilliland" wrote in
message Tobiah wrote: The bottom line is that more memory will always help to some degree, because it reduces the amount of redundant disk reads. I don't know how Windows handles this, but if it is not in a similar way, then it is quite lacking in this regard. Perhaps, but that kind of caching provides almost zero benefit in audio processing. Sequential reads of long files don't really take much advantage of a cache. Especially when you're processing a dozen or two of them concurrently. When files are small enough to completely fit in RAM, and the RAM doesn't get stolen for other uses for long enough, then recurrent operations upon the same file can take place primarily in RAM. Changes will be written out to disk as possible. This is much more likely to happen with smaller files, but if you have enough RAM, a whole track or even a collection of tracks can start fitting into the "smaller" category. |
#47
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Arny Krueger wrote:
Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. True, but it's not likely to have the same kind of impact as adding another drive to a RAID array. Again, assuming large files accessed sequentially. |
#48
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Jim Gilliland" wrote in
message Arny Krueger wrote: Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. True, but it's not likely to have the same kind of impact as adding another drive to a RAID array. I don't know about that. As I've said before, when you're doing sequential copies, doing them drive-to-drive as opposed to on the same drive (even a drive that is implemented as an array) is one of the cheapest thrills around. Once upon a time I built the same machine twice for running Audition/CEP. (1) 2 120 GB drives striped, with all files on the array (2) 2 120 GB drives, temp files on one drive, permanent files on the other (2) blew (1) away through a normal record/save/edit cycle involving 30 minute 16 track recording/editing sessions. |
#49
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
On Mar 20, 4:29 am, Laurence Payne lpayne1NOSPAM@dslDOTpipexDOTcom
wrote: On Tue, 20 Mar 2007 01:31:54 -0800, spud wrote: Is there a specific problem you are looking to solve? With 16 main tracks and just a few clips on maybe 10 others, after I add fx the machine runs at roughly 70% and feels glitchy and sometimes shows funky screen redraws. I thought adding memory might solve that. Also, I meant gig of memory not meg, sorry. Maybe the key is in that throwaway line "...once I add fx...." :-) Manufacturers delight in making e.g. high-quality reverb plugins that are very resource-hungry. It's no great trick to max out ANY hardware by adding instances of these. Tell us about your FX structure. Are you using a single Send FX for several tracks where possible? You could maybe spend a lot of money and buy a new machine that allowed 50% or even 100% more plugins. Or you could maybe rationalise your use of FX and require enormously less resources. Anyway, tell us how your tracks and FX are set up? Yes, I agree--could very well be the FX stealing CPU cycles, especially if they are running on multiple tracks. Always best to bounce multiple tracks to a single track before applying plugin if possible. |
#50
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Arny Krueger" wrote in message
... "Jim Gilliland" wrote in message Arny Krueger wrote: Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. True, but it's not likely to have the same kind of impact as adding another drive to a RAID array. I don't know about that. As I've said before, when you're doing sequential copies, doing them drive-to-drive as opposed to on the same drive (even a drive that is implemented as an array) is one of the cheapest thrills around. Once upon a time I built the same machine twice for running Audition/CEP. (1) 2 120 GB drives striped, with all files on the array (2) 2 120 GB drives, temp files on one drive, permanent files on the other (2) blew (1) away through a normal record/save/edit cycle involving 30 minute 16 track recording/editing sessions. Doesn't surprise me at all. But then again, I use r zero-raid arrays; one for system, one for data and one for temps. Works pretty well...get the best of both worlds. The basic zero-raid ain't that superfast or anything on the level of 'whoaa' when it comes to performance. But it does do away with the constant 'bzzz bzzz' of the hardrives and lets them run real cool, because they share both the load and the seek tasks. That's the best part of raid arrays in my book. |
#51
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Arny Krueger wrote:
"Jim Gilliland" wrote in message Arny Krueger wrote: Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. True, but it's not likely to have the same kind of impact as adding another drive to a RAID array. I don't know about that. As I've said before, when you're doing sequential copies, doing them drive-to-drive as opposed to on the same drive (even a drive that is implemented as an array) is one of the cheapest thrills around. Once upon a time I built the same machine twice for running Audition/CEP. (1) 2 120 GB drives striped, with all files on the array (2) 2 120 GB drives, temp files on one drive, permanent files on the other (2) blew (1) away through a normal record/save/edit cycle involving 30 minute 16 track recording/editing sessions. Yes, I'd expect that to make a big difference. Audition uses its temp files fairly extensively. They strongly recommend providing temp space on multiple drives and keeping it away from the permanent files that you are processing. The only work that I do where duration is much of an issue is rendering concert mixes in Vegas. Vegas doesn't really use temp files, it just cranks right through from the multiple source files to a single output file. I have found that keeping the output file on a separate drive will make things about 5% faster, but it's not a huge improvement. The bulk of the I/O is in reading the source files, not writing the target. So I'm really just looking at my own circumstances. With other software and other processes, the scenario may be different. |
#52
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
"Jim Gilliland" wrote in
message Arny Krueger wrote: "Jim Gilliland" wrote in message Arny Krueger wrote: Memory can indirectly speed up disk operations by allowing more RAM for disk caching. Dual channel RAM as opposed to single channel RAM has the potential to indirectly speed up disk operations by allowing the RAM buffers to work faster. True, but it's not likely to have the same kind of impact as adding another drive to a RAID array. I don't know about that. As I've said before, when you're doing sequential copies, doing them drive-to-drive as opposed to on the same drive (even a drive that is implemented as an array) is one of the cheapest thrills around. Once upon a time I built the same machine twice for running Audition/CEP. (1) 2 120 GB drives striped, with all files on the array (2) 2 120 GB drives, temp files on one drive, permanent files on the other (2) blew (1) away through a normal record/save/edit cycle involving 30 minute 16 track recording/editing sessions. Yes, I'd expect that to make a big difference. Audition uses its temp files fairly extensively. They strongly recommend providing temp space on multiple drives and keeping it away from the permanent files that you are processing. The only work that I do where duration is much of an issue is rendering concert mixes in Vegas. Vegas doesn't really use temp files, it just cranks right through from the multiple source files to a single output file. I have found that keeping the output file on a separate drive will make things about 5% faster, but it's not a huge improvement. The bulk of the I/O is in reading the source files, not writing the target. When you add the second temp drive, Audition/CEP splits the temp files over two drives, which is very similar to striping, whithout making a permanent commitment to striping and its unfavorable reliability impact. |
#53
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Julien BH wrote:
Three choices the Memory Hard Drive CPU Normally the slowest of the three is the Hard drive. If you want more speed be prepared for a salty bill. SCSI is the fastest solution, but SATA Raid 0 is a good solution also. Provided the right stripe size is used for the intended application. Also, some 'fake' (i.e. pure software) raid controllers may not yield the promised performance gains. It's worth remembering that one failing drive in a raid 0 easily may render all data unrescueable. -- Kind regards, Mogens V. |
#54
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
Tobiah wrote:
Now I'm not saying that having 2gb is not necessary. In fact if you load up VERY LARGE libraries it could be used. But most of the time it'll stay idle and wait for some data. On a Linux box, all available memory will quickly be consumed for caching hard drive reads. It doesn't matter how large the memory space is. If your memory size equals that of the disk, you could conceivably cache the entire disk, and not have to read from the disk again until reboot. Not fully true. Data is sync'ed to disk all the time, so real file changes needs to be rehashed. The same goes for writes, although the OS usually syncs writes to the disk now and then to prevent loss of data through loss Not usually; always, and not now and then, all the time. of power. The memory is not 'full', in the sense that the OS will simply drop cached disk reads when the memory is needed by a process. The bottom line is that more memory will always help to some degree, because it reduces the amount of redundant disk reads. I don't know how Windows handles this, but if it is not in a similar way, then it is quite lacking in this regard. In this respect, M$ lacks, but it's a bit irrelevant as Linux isn't yet useful as base for a DAW. -- Kind regards, Mogens V. |
#55
Posted to rec.audio.pro
|
|||
|
|||
XP based DAW: More than one meg of memory?
On Mar 19, 4:08 pm, "Geoff" wrote:
I doubled my memory and disc size. My CPU is 100% faster. How come my music isn't twice as good ? You have to pay Bill Gates for the secret update. |
Reply |
|
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Sampler memory | Pro Audio | |||
Memory distortion? | Pro Audio | |||
Memory jog? | Audio Opinions | |||
Flash memory recorder without built-in memory. Is there one? | General | |||
Flash memory recorder without built-in memory. Is there one? | Tech |