|
Prev: lame-3.98b7.tar.tar (gz)
Next: DAW and PowerPoint
From: kitekrazy on 12 Apr 2008 16:09 Glennbo wrote: > In news:On4Mj.8373$2g1.4905(a)nlpi068.nbdc.sbc.com the killer robot kitekrazy > <kitekrazy(a)sbcglobal.net> grabbed the controls of the spaceship > cakewalk.audio and pressed these buttons... > >> Anyone problems running certain software with a Quad? >> I know Sonar will be no problem. > > There's a thread in the weeb forum about how Sonar doesn't allocate it's > tasks very well on multi-core machines. If you run ultra low latency and > use a lot of plugins (which is how I run), the problem is magnified. > > http://forum.cakewalk.com/tm.asp?m=1356871 > Very interesting. I'm surprised the OP didn't need a heavily armored flame suit. Maybe they'll fix it in Sonar 8. Interesting comment: "Obviously we will investigate improving performance scaling across cores in the future but this is very unlikely to impact most users running multi-core in a very significant way. Remember that this is a benchmark designed as a stress test - great for debugging but very different from most real world projects. So getting worried about just this one test is a bit extreme. IMHO Scott hit the nail on the head in his reply when when he said "this does not effect most users to where its a deal breaker"."
From: John Braner on 13 Apr 2008 07:36 Glennbo wrote: > In news:_F8Mj.1227$7Z2.710(a)newssvr12.news.prodigy.net the killer robot > kitekrazy <kitekrazy(a)sbcglobal.net> grabbed the controls of the > spaceship cakewalk.audio and pressed these buttons... > >>>> Anyone problems running certain software with a Quad? >>>> I know Sonar will be no problem. >>> There's a thread in the weeb forum about how Sonar doesn't allocate >>> it's tasks very well on multi-core machines. If you run ultra low >>> latency and use a lot of plugins (which is how I run), the problem is >>> magnified. >>> >>> http://forum.cakewalk.com/tm.asp?m=1356871 >>> >> Very interesting. I'm surprised the OP didn't need a heavily armored >> flame suit. >> >> Maybe they'll fix it in Sonar 8. >> >> Interesting comment: >> >> "Obviously we will investigate improving performance scaling across >> cores in the future but this is very unlikely to impact most users >> running multi-core in a very significant way. Remember that this is a >> benchmark designed as a stress test - great for debugging but very >> different from most real world projects. >> So getting worried about just this one test is a bit extreme. IMHO >> Scott hit the nail on the head in his reply when when he said "this >> does not effect most users to where its a deal breaker"." > > Noel also just said that plugin processing is done serially for all plugins > in a particular FX bin, which means if you have several CPU hungry plugs in > a bin together, they will all be assigned to the same core, rather than > their load being distributed across the available cores. > Q6600 quad core works fine here. OK, I have the same comp/eq plug on all tracks and then just some other stuff on one or two tracks (or busses) - but everything is pretty enely split and I'm running a *hell* of a lot more than I could on a single core 3.2 GHz P4. -- =========== John Braner jbraner(a)NOblueyonderSPAM.co.uk http://www.soundclick.com/johnbraner_music.htm (that's an underscore) ** Posted from http://www.teranews.com **
From: malachi on 13 Apr 2008 11:07 "Glennbo" <vdrumsYourHeadFromYourAss(a)cox.net> wrote in message news:Xns9A7F66463D96ABrownShoesDontMakeIt(a)69.28.186.120... > In news:8b3d7$4801f03c$19774(a)news.teranews.com the killer robot John > Braner <me(a)myhouse.com> grabbed the controls of the spaceship > cakewalk.audio and pressed these buttons... > .... I'm only running a Core 2 Duo > 2.4Ghz machine, but the same kind of project in Reaper purrs like a > kitten. > Which has let me to question whether the trouble of upgrading my Core 2 Duo to a quad is worth it at least for the near term. malachi
From: Sue Morton on 13 Apr 2008 16:40 I haven't found latency to be an issue except when recording live tracking to playback. Or the case where latency was actually too high and a synth couldn't handle it. It was Sonar 3 (I think), and the Roland VSC synth (I think). I had raised the latency for a complex mixing job, and had forgotten to lower it back down. Midi was a mess with that synth, and it took me awhile to realize it was the buffering. Dropping latency back down to 'normal' solved the problem with the synth. I generally run everything at 11ms latency unless tracking to playback. I can run with less, and often do because I forget to raise it after tracking :-) 11ms seems to keep my Lynx One happy. -- Sue Morton Organfreak wrote: > Glennbo <vdrumsYourHeadFromYourAss(a)cox.net> blatted: >> In news:5fk404tjhrs8adhv5ehdeej2r8n3lfvfg1(a)4ax.com the killer robot >> Organfreak <plonk(a)plinkety.plunk> grabbed the controls of the >> spaceship cakewalk.audio and pressed these buttons... >> >>>> I would expect that you should be able run all day long at 1.5ms >>>> just like I do. All but a handful of my projects in Sonar play >>>> without looking stressed at 1.5ms. That's actually 3.0ms round >>>> trip through the input and output buffers and running 44.1Khz >>>> 16bit. In Reaper, I'm running 44.1Khz 32bit using ASIO which is >>>> set at 1ms, which is 2ms round trip through the input and ouput >>>> buffers. It displays in the upper right corner of Reaper. >>>> >>>> http://members.cox.net/vdrums/Troublemaker/ >>>> >>>> I hadn't tried to drop it to 16bit so far, but I can run complex >>>> projects with tons of effects all day long set like this, and it >>>> never breaks a sweat. >>> >>> Forgot to specify that I am, of course, using ASIO. I can't see >>> where I could adjust it downwards; the slider is full-left at 5.8 >>> ms. >>> >>> My buffers are both set at only 32, with no read or write ahead >>> eneabled. >> >> In the Delta control panel on the tab labeled "hardware settings" a >> DMA buffer size of 64 samples should get 1ms @ 44.1Khz. I just set >> Sonar 5 to ASIO set like that, played my song El-Six-Es and it had >> pops and clicks several times through the playback. I can run that >> same latency using WDM in Sonar, but not ASIO. In Reaper, I can run >> ASIO at 64 samples and never get a pop or click. I haven't bothered >> trying using WDM in Reaper, since ASIO is working so well. > > Thanks. What a dumbass I am; I had forgotten all about that! I had > the MF set on 256, no wonder! 128 gets me to 2ms, haven't actually > tried it out yet.
From: kitekrazy on 13 Apr 2008 17:41
Organfreak wrote: > Glennbo <vdrumsYourHeadFromYourAss(a)cox.net> blatted: >> In news:8b3d7$4801f03c$19774(a)news.teranews.com the killer robot John >> Braner <me(a)myhouse.com> grabbed the controls of the spaceship >> cakewalk.audio and pressed these buttons... >> >>>> Noel also just said that plugin processing is done serially for all >>>> plugins in a particular FX bin, which means if you have several CPU >>>> hungry plugs in a bin together, they will all be assigned to the same >>>> core, rather than their load being distributed across the available >>>> cores. >>>> >>> Q6600 quad core works fine here. OK, I have the same comp/eq plug on >>> all tracks and then just some other stuff on one or two tracks (or >>> busses) - but everything is pretty enely split and I'm running a >>> *hell* of a lot more than I could on a single core 3.2 GHz P4. >> I've got some songs that have so many CPU hungry plugs that I can see Sonar >> breaking a sweat trying to play them back at 1.5ms latency. I still have >> S5, so I don't have the per core meters, but when the counters and VU >> meters start becoming visibly laggy, and the CPU meter is dancing into the >> upper limits, I can tell it's at the edge. I'm only running a Core 2 Duo >> 2.4Ghz machine, but the same kind of project in Reaper purrs like a kitten. > > <hijack> > > I wonder why my latency is 5.8? Isn't that right on the edge of too much? (Not > doing any MIDI in the current projects.) > > Delta 1010LT > Dual AMD 5400+ X2 > 2 GB RAM > > </hijack> > No. It must not bother you. I run most of my cards in ASIO where the latency is 5. Sure I could run them at 2 ms but I don't notice the difference. |