http://www.gearslutz.com/board/music-computers/887142-dawbench-sonar-x3c-vs-reaper-4-57-a.html

can anyone reproduce?

Kind of pointless to my mind. Who cares how many instances of a single VST will can run in a project? That might be a reasonable way to benchmark machine performance as you upgrade or make tweaks, but it's a lousy way to compare different DAW apps.
 
And a discrepancy that big is probably due to a config issue or something peculiar to that VST.

I've always been a little leery of benchmarks that load up one plugin over and over until dropouts occur. It seems like a clunky and imprecise method that might not even produce repeatable results, much less translate into real-world benefits.
 
Even when testing things with far fewer variables, such as disk performance, I can run the same benchmark with the same software on the same machine and get different results on different days.
 
One would reasonably expect Reaper to perform well, just because it's not been around as long as SONAR and is not burdened by legacy code (can Reaper load a 10-year-old project? Nobody knows...there aren't any 10-year-old Reaper projects.) But I'd be very surprised if the performance gap was a) large and b) across the board for all activities.

Very interesting, though I agree with Dave none the less I was a little surprised.  

I ran my own version of a "test" just two days ago. Instead of loading twenty hundred copies of a compressor vst, I tried loading around forty synth instances (various brands), applied midi tracks, sang a little, added some acoustic guitar (call me a purist), and put a song together. Test results confirm, X3 works as advertised for me. Liked the song. Think I'll keep using Sonar on the next song. End of test.
 

bitflipper
 One would reasonably expect Reaper to perform well, just because it's not been around as long as SONAR and is not burdened by legacy code (can Reaper load a 10-year-old project? Nobody knows...there aren't any 10-year-old Reaper projects.) But I'd be very surprised if the performance gap was a) large and b) across the board for all activities.

burdened ? I doubt.. it's a type of structure.. so the like a Header/Lib in C your more interested than the function over all , opposed to the content within ... that gets from A-B . Even if the code itself does have varients of the same code within, no doubt the header of the file gives revision / version no. anyway.
 
all the DAW makers run benchmarks against their rivals.. and no doubt part of the dev cycle is streamlining certain existing code.
 
it's all abstract.. or a degree of separation, much like anyone using the SDK for say asio etc.

People always want what they use to be the best, and they'll go to great lengths to make it so. It's all bollocks.

That guy seemed like he wants Reaper, or at least expects Reaper to work better, which is not even close to his findings thus far.
 
Interesting indeed.
 
Lance

czyky
I ran my own version of a "test" just two days ago. Instead of loading twenty hundred copies of a compressor vst, I tried loading around forty synth instances (various brands), applied midi tracks, sang a little, added some acoustic guitar (call me a purist), and put a song together. Test results confirm, X3 works as advertised for me. Liked the song. Think I'll keep using Sonar on the next song. End of test.

Sounds like we use a remarkably similar procedure for testing software. Except I also do the "how fast can I get narration recorded when I have to make the Fed Ex 6 PM dropoff?" test.
 
Yes, it's nice Sonar came out "on top"...but on top of what? A flawed test that has little, if anything, to do with the intended function of either piece of software?
 
I can just see this mentality applied to testing acoustic guitars:
 
"I used D'Addario EXL140 strings on both a Gibson and Taylor acoustic guitar. With the Gibson, all strings had to be tuned up a third above the standard EADGBE tuning, and then the 1st and 2nd strings tuned up until they snapped, for the string relief to exceed 1/8th inch above the 12th fret at the time just before the strings snapped. With the Taylor, it was necessary to tune the strings the same as the Gibson, but also tune the sixth string up an additional semitone and the G string until it snapped, for the string relief to exceed 1/8th inch just before the point where the G string snapped.
 
"From this scholarly and carefully controlled scientific test, I have concluded that both instruments are, in fact, guitars."