We're now using more RAM for the O/S than hard drive space was available 15 short years ago.jupiter8 wrote: Ok let me put it this way:i am convinced i am going to see it at one point (hey a bigger number is always better) but will have no use for it in my lifetime.
Or simply: i doubt i'll ever see a mobo with support for over 4 petabytes of RAM.
Multi processor support differences in various hosts
-
- KVRAF
- 1869 posts since 15 Sep, 2003 from Land of Crazies, USA
- KVRAF
- 9457 posts since 17 Sep, 2002 from Gothenburg Sweden
Yeah i know and mobile phones today are about as powerful as the computer i bought 10 years ago. Moores law is mindboggling.Dominus wrote:We're now using more RAM for the O/S than hard drive space was available 15 short years ago.jupiter8 wrote: Ok let me put it this way:i am convinced i am going to see it at one point (hey a bigger number is always better) but will have no use for it in my lifetime.
Or simply: i doubt i'll ever see a mobo with support for over 4 petabytes of RAM.
I haven't actually calculated how long it would take to break the 64 bit barrier but i have feeling it'll be a long long time.
-
- KVRAF
- 1888 posts since 13 Aug, 2011 from Berlin
OK, thanx for the details on this about Reaper. How is FL doing it? I read at the beginning of this(?) thread that it is doing rather well, too?tony tony chopper wrote:So when you're comparing stuff, know what you're comparing - no one can do magic with things they can't control (third party plugins), there are always drawbacks.
-
tony tony chopper tony tony chopper https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=3103
- KVRAF
- 3561 posts since 20 Jun, 2002
It has triple buffering in the last release, effectively processing one buffer ahead, it appeared to be more efficient depending on the drivers.How is FL doing it?
As for multithreading, same as everyone else, parallelizing what can be, works very well in benchmarks, not so much with real projects.
DOLPH WILL PWNZ0R J00r LAWZ!!!!
- KVRAF
- 25053 posts since 20 Oct, 2007 from gonesville
well, here is my observation of how my preferred instruments and FX host is working on this project:
about 2/3rds of the CPU is reported as "idle".
both the 32 bit and 64 bit 'servers' are instantiated; they are using, combined, around 450"% CPU", which to me seems to indicate 4.5 threads out of 16 logical threads (the 1/3rd of the CPU reportedly used).
so, my systems's performance at 64 samples buffer in a mixing scenario is not a problem.
about 2/3rds of the CPU is reported as "idle".
both the 32 bit and 64 bit 'servers' are instantiated; they are using, combined, around 450"% CPU", which to me seems to indicate 4.5 threads out of 16 logical threads (the 1/3rd of the CPU reportedly used).
so, my systems's performance at 64 samples buffer in a mixing scenario is not a problem.
-
- KVRAF
- 42529 posts since 21 Dec, 2005
How can it be the plugin when you launch an empty version of S2 and it takes up 5-10% cpu?Benjaminjo wrote: Richard from Orion said it is the plugin's fault, but how come in Reaper the same plug behaves correctly? How's that possible?
-
- KVRist
- 327 posts since 13 Nov, 2002 from Germany, Darmstadt
This is a claim that came up in 2009 which has since been demented by Microsoft.UltraJv wrote:jupiter8 wrote:I doubt i'll see that in my lifetime actually and i do plan to live another 30-40 years or so.UltraJv wrote:128 bit CPUs and OS will soon be here and the whole shooting match kicks off again.
Windows 8 is designed as 128 bit:
"Robert Morgan, Senior Research & Development at Microsoft, is already testing Windows 8 and Windows 8 128-bit support"
http://news.softpedia.com/news/128-bit- ... 3691.shtml
Nope, the address space of Bulldozer is still 64 bit. With 64 bit you can already address 16 Exabyte. Nobody is currently near that and nobody will for many years.UltraJv wrote:AMD have been working on 128 bit Bulldozer CPU since 2009.
Maybe you confused something with the SIMD registers which are 128 bit wide and for SIMD floating point instructions even increase to 256 bit because of AVX support?
-
- KVRAF
- 6323 posts since 30 Dec, 2004 from London uk
Truth is that neither will give roadmaps for products like that until existing ones have been sold. Lets see what happens in the next few years, multi core isnt a wow factor anymore, 128 bit will be.helium wrote:This is a claim that came up in 2009 which has since been demented by Microsoft.UltraJv wrote:jupiter8 wrote:I doubt i'll see that in my lifetime actually and i do plan to live another 30-40 years or so.UltraJv wrote:128 bit CPUs and OS will soon be here and the whole shooting match kicks off again.
Windows 8 is designed as 128 bit:
"Robert Morgan, Senior Research & Development at Microsoft, is already testing Windows 8 and Windows 8 128-bit support"
http://news.softpedia.com/news/128-bit- ... 3691.shtml
Nope, the address space of Bulldozer is still 64 bit. With 64 bit you can already address 16 Exabyte. Nobody is currently near that and nobody will for many years.UltraJv wrote:AMD have been working on 128 bit Bulldozer CPU since 2009.
Maybe you confused something with the SIMD registers which are 128 bit wide and for SIMD floating point instructions even increase to 256 bit because of AVX support?
-
tony tony chopper tony tony chopper https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=3103
- KVRAF
- 3561 posts since 20 Jun, 2002
Sadly I'm pretty sure that marketing can sell 128bit. The question is, is it only marketing that drives technology now, that is, would serious people at Intel start to work on something as silly just because it can be sold? Since they fail to bring anything new.. possibly.Truth is that neither will give roadmaps for products like that until existing ones have been sold. Lets see what happens in the next few years, multi core isnt a wow factor anymore, 128 bit will be.
But PC users are more serious, and there are serious blogs that wouldn't hesitate to shoot "128bit" down if anyone was announcing that.
Besides, marketing can use 128bit without it really being 128bit addressing. So if I was into marketing I'd rather use 256bit, because those VEX instructions are.
DOLPH WILL PWNZ0R J00r LAWZ!!!!
-
- KVRAF
- 1888 posts since 13 Aug, 2011 from Berlin
Cool, thanx for that. And how is the performance of FL compared to Reaper (which I'm using right now)?tony tony chopper wrote:It has triple buffering in the last release, effectively processing one buffer ahead, it appeared to be more efficient depending on the drivers.How is FL doing it?
As for multithreading, same as everyone else, parallelizing what can be, works very well in benchmarks, not so much with real projects.
-
tony tony chopper tony tony chopper https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=3103
- KVRAF
- 3561 posts since 20 Jun, 2002
You shoudln't be trusting any answer of mine & do a benchmark, as long as you're sure of the proper settings in both.Cool, thanx for that. And how is the performance of FL compared to Reaper (which I'm using right now)?
Scott made a related video recently btw
http://www.youtube.com/watch?v=DgOV21oi06Q
because another problem when testing CPU usage these days, is today's energy saving features.
DOLPH WILL PWNZ0R J00r LAWZ!!!!
-
- KVRAF
- 42529 posts since 21 Dec, 2005
Informative how? I found it insulting, then again, no surprise considering.
Studio One has a little thing you can get which disables core parking, speed stepping, all that. On my system it did nothing
However, I can run gobs of high cpu instruments is REAPER and there is never any kind of issue whatsoever.
Studio One has a little thing you can get which disables core parking, speed stepping, all that. On my system it did nothing
However, I can run gobs of high cpu instruments is REAPER and there is never any kind of issue whatsoever.