I'm pretty sure we are.

yorolpal


I'm pretty sure we are.

 
It should be pretty simple really 'Olpal - one John is a loyal over-zealous fanboi who lurves everything there is about Cakewalk and won't have a word said against their products, and the other John..................... oh ...... er ..........yeah ...
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Only kidding you John
 
 

I've been studiously ignoring this thread based initially on its title, and later on the number of posts (almost always a bad sign ), but now that I've glanced through it...

I'm with the camp that says the performance difference between x86 and x64 is not "night and day". That's pure hyperbole. If you don't currently have or need the extra RAM that x64 supports, there is virtually no point in going there right now.
 
I did mainly because the "builder" copy of Win7 that shipped with my Quad processor upgrade did not include the x86 version, and partly because I thought the promise of improved performance, no matter how modest, would be nice. I only have 4GB of RAM, and do not currently own any synths or create projects that need more, so that was not a consideration.
 
Through much playing around, and some deliberate testing, I have been able to discern exactly zero performance improvement, and there have been a number of costs:

1. Loss of x86 DXis.
2. Loss of Rewire support.
3. The Bitbridge timing bug that adds an empty buffer of audio to some synth outputs (Dropzone to name one that is included with SONAR).
4. Other Bitbridge awkwardnesses and instability related to its running as a separate application process.
5. Buggy x64 drivers for my audio interface.

Some of this can be addressed by running x86 SONAR, but I prefer not to have to maintain two installs, and have been willing to put up with the compromises.

The Bitbridge problems were addressed by running most x86 plug-ins under Jbridge.

The x64 E-MU driver problems continue to plague me, and will necessitate a pricy interface upgrade and associated learning curve.

All things considered, I probably would not have made the move had I anticipated all the consequences, and if having my DAW work as well as possible were my only goal. I blame only myself for this choice, though the few datapoints I could find on how the E-MU would work under x64 were misleadingly positive.

But I have to say, also, that I like to stay somewhat current on technology (and need to for my work). And for that reason, I haven't been massively disappointed in the "upgrade". It's been a valuable learning experience.

Here I have a few links to articles and threads that discuss this.

Noal on 64 bits

Article on 64 bits

Another from MS

A thread some my find useful

brundlefly



1. Loss of x86 DXis.
2. Loss of Rewire support.
3. The Bitbridge timing bug that adds an empty buffer of audio to some synth outputs (Dropzone to name one that is included with SONAR).
4. Other Bitbridge awkwardnesses and instability related to its running as a separate application process.
5. Buggy x64 drivers for my audio interface.

Some of this can be addressed by running x86 SONAR, but I prefer not to have to maintain two installs, and have been willing to put up with the compromises.

The x64 E-MU driver problems continue to plague me, and will necessitate a pricy interface upgrade and associated learning curve.

I do maintain the two install precisely because of the caveats you cite there, Rewire for example I couldn't live without, and having the 32 bit version to hand is certainly much less effort than correcting some of the weirdness apparent running 32 plugs via a bridge.  The only consideration is keeping the VST folders synced up, the rest of it they share the same data and content.

As for the E-MU driver issue... sheesh over a year in beta with the same old version and they still see fit to bring out 'new' interfaces meantime.

The key to the transitional phase we are still in WRT to 64 bit despite it being not that new is user base, and for me it is the entry level users machines being equipped with a 64 bit and a massive price drop in some real hefty processors that herald the real change which undoubtedly isn't going to be backwards.

I'm sorry Brundlefly, ol pal but when I say that "for me" the difference is night and day I am not being hyperbolic.  Granted there is little (although I still maintain there is some) "performance" gain but, as I stated earlier, there is tremendous "headroom" gain providing you load your DAW up with memory (I run 12 gigs).  On my system there are no more crashes, hangups and such than I had under 32bit and most all my plugs (and I have way too many) run just fine under bitbridge.  I'm quite sure this is not the case for many others but I believe most of the differing viewpoints on this issue come down to system specs and exactly how you work and what VST instruments and effects you happen to use.  Again, try running six concurrent instances of Omnisphere and three or four concurrent instaces of Amplitube 3 plus FX plugs galore on all tracks on your 32 bit system WITHOUT freezing anything and tell me how you faired.  I can do such now and more without even straining my system.  There is no question that 64bit systems afford more "headroom" and that equals "performance" in my book.  YMMV.

One thing I have noticed is the 4 gb memory that those that seem not to see much difference between 32 bits and 64 bits have. I wonder if that could be a reason. I have 8 gb and I notice a big difference.

yorolpal

... I believe most of the differing viewpoints on this issue come down to system specs and exactly how you work and what VST instruments and effects you happen to use.

I don't even think it comes down to that (referring to specs such as RAM, CPU etc etc). I think it comes down to the fact that you have a professionally built DAW from someone who knows what components to use and which ones not to.

Not all RAM is equal, even though spec's tell you they are. Not all CPU's are the same. Even ones that are made by the same manufacturer with the same identical model number on them are sometimes not the same. Couple all of that with constantly changing motherboard bios updates, motherboard chipset updates, Windows updates, video card driver updates, sound card driver updates, X1 a-b-c patches and quick fixes, it's really amazing all this stuff works half as well as it does.

John

One thing I have noticed is the 4 gb memory that those that seem not to see much difference between 32 bits and 64 bits have. I wonder if that could be a reason. I have 8 gb and I notice a big difference.

Hi John,

I swear I'm not saying this just to go against you, but I have to say I had 8GB of RAM in my DAW and saw absolutely no difference in Windows 7 x64 or Sonar x64 compared to when I switched back to 4GB. I stated earlier that I had 8GB of RAM but downgraded.

Now ... the difference may be your memory requirements are different than mine. It could be you go over the 4GB limit. I never have. I just checked, and my largest project consumes only 48% of my total system RAM (*), assuming the RAM indicator in X1 is accurate.

If you've always had 8GB, how do you know there's a difference or not between 4GB and 8GB? You know what I'm saying?

I haven't seen any specs come from you, if I missed them my apologies.

I'm running an i5 750 Quad with 4GB RAM on a Gigabyte Mobo.

What are you running?

(*) UPDATE: 90% of the samples I use stream off the disk. I'm sure my memory would be filled to the max if they were loaded in to RAM.

yorolpalI'm sorry Brundlefly, ol pal but when I say that "for me" the difference is night and day I am not being hyperbolic. 

But you got that Studiocat I7 machine with 12GB at the same time you migrated to x64, right? You didn't just upgrade the O/S on an existing machine with no other changes like did.


I don't know what you were running before, but I imagine the overall difference in performance load capability between your I7 and my Q9550 is pretty significant, even though they run the same O/S.


All I know is that all the benchmark projects I tested before and after the upgrade yielded exactly the same CPU usage numbers at any given buffer setting.