| View previous topic :: View next topic |
| Author |
Message |
jcschild
Senior Member
Joined: 06 Jan 2005
Posts: 1245
Location: Kentucky
|
 Posted: Wed Jun 15, 2005 3:14 pm Post subject: |
 |
|
Hey VIn,
i just noticed something about your latency....
the echo reads differantly that others (i forgot about that)
@256 it shows 7.4 when it should read 5.8. this would skew the test but not by as much as i am behind.
Hey Guy, what audio card are you testing with?
_________________
Scott
ADK
Home of the "Kentucky Fried DAW" |
|
| Back to top |
|
 |
Beat Productions
Member
Joined: 18 Jan 2005
Posts: 766
|
| Posted: Wed Jun 15, 2005 4:26 pm Post subject: |
|
|
Please specify if all your tests are being done with N2 or N3 (big difference).
_________________
Comp1: Windows 7 64bit, Gigabyte GA-EX58-UD3R, Intel I7 920, 6 gig Patriot DDR3 ram, RME Digiface, Geoforce 66000 PCIe, , Cubase 5.5
Comp 2:Windows 7 64bit, Asus P5W DH Deluxe, Q6600, 8 gig Corsair XMS ram, , MR816 X, Saffire Pro 40, VE Pro |
|
| Back to top |
|
|
SONICA-X
Member
Joined: 28 Jan 2005
Posts: 169
Location: BOSTON, MA USA
|
| Posted: Wed Jun 15, 2005 5:47 pm Post subject: |
|
|
| jcschild wrote: | Hey VIn,
i just noticed something about your latency....
the echo reads differantly that others (i forgot about that)
@256 it shows 7.4 when it should read 5.8. this would skew the test but not by as much as i am behind.
Hey Guy, what audio card are you testing with? |
Scott,
Fireface 800 and Presonus Firepod.
Hey those X2 benchmarks sure look good, in some the 4200+ will beat the D840 extreme at 1/2 the cost.
I can't wait anymore I am going to put one in N4 and run Thonex.
Guy. |
|
| Back to top |
|
|
jcschild
Senior Member
Joined: 06 Jan 2005
Posts: 1245
Location: Kentucky
|
| Posted: Wed Jun 15, 2005 7:32 pm Post subject: |
|
|
Hey Guy,
"fireface" yeah i thought so! dang it i am out!
i have not bothered to try the N4 boards as they suck so bad with a 3500 standard.
that at best i see them with a x2 being on par with a standard in a N3.
i just got the asus a8v in today so i will have answers shortly...
_________________
Scott
ADK
Home of the "Kentucky Fried DAW" |
|
| Back to top |
|
|
jcschild
Senior Member
Joined: 06 Jan 2005
Posts: 1245
Location: Kentucky
|
| Posted: Wed Jun 15, 2005 7:33 pm Post subject: |
|
|
all my tests are done with N3 all updates.
_________________
Scott
ADK
Home of the "Kentucky Fried DAW" |
|
| Back to top |
|
|
TAFKAT
Member
Joined: 05 Apr 2005
Posts: 997
Location: Melbourne: Australia
|
| Posted: Wed Jun 15, 2005 11:33 pm Post subject: |
|
|
| jcschild wrote: | Hey VIn,
i just noticed something about your latency....
the echo reads differantly that others (i forgot about that)
@256 it shows 7.4 when it should read 5.8. this would skew the test but not by as much as i am behind.
Hey Guy, what audio card are you testing with? |
Hey Scott,
128 Samples or 256 Samples is the constant, not what the ASIO driver is actually reporting back, I have noticed Nuendo and SX both sometimes display weird arse latency figures depending on the hardware installed.
I have had close enough to identical figures with a Fireface and a Lynx 22 , so the reported latecy isn't effecting the result.
BTW: All of my results have been on N2 to maintain consistancy with my earlier figures...
@ Beat Productions,
Big Difference between N2 and N3 ??? Really ??
Not according to Steinberg :-) BTW: From memory , some posters were reporting slightly better figures with N3 .. :-)
Again, if there is any doubt , then maybe Scott can run the benches on N2 and see if there is any discernable difference, if there is, what is it saying about N3 ?
Either way, I am not approaching this as a pissing competition, the figures/screenshots are posted as an indication, I'll let them speak for themselves. :-)
Peace: |
|
| Back to top |
|
|
Thonex
Senior Member
Joined: 06 Jan 2005
Posts: 2049
Location: Los Angeles, CA
|
| Posted: Thu Jun 16, 2005 12:01 am Post subject: |
|
|
Hey Tafkat,
I noticed on your screen shots that you have direct monitoring engaged. Have you tried to see if there is a performance hit when it is DIS-engaged?
THanks,
T
_________________
Nuendo 5, Mac Pro 8 Core 2.33 Mhz, OS 10.6, 16 gigs of Ram, MOL , RME 9652, 2xSamsung 21.3" LCDs (1600x1200 DVI), 2 X-Keys. |
|
| Back to top |
|
|
TAFKAT
Member
Joined: 05 Apr 2005
Posts: 997
Location: Melbourne: Australia
|
| Posted: Thu Jun 16, 2005 12:51 am Post subject: |
|
|
| Thonex wrote: | Hey Tafkat,
I noticed on your screen shots that you have direct monitoring engaged. Have you tried to see if there is a performance hit when it is DIS-engaged?
THanks,
T |
Hey T,
I just re-ran the tests with direct monitoring off - No change :-)
btw: I usually do have it off with any cards with hardware monitoring ..
Peace: |
|
| Back to top |
|
|
franYo
Member
Joined: 06 Jan 2005
Posts: 369
Location: Croatia
|
| Posted: Thu Jun 16, 2005 8:46 am Post subject: |
|
|
| TAFKAT wrote: | @ Beat Productions,
Big Difference between N2 and N3 ??? Really ??
Not according to Steinberg :-) BTW: From memory , some posters were reporting slightly better figures with N3 .. :-)
Again, if there is any doubt , then maybe Scott can run the benches on N2 and see if there is any discernable difference, if there is, what is it saying about N3 ?
Either way, I am not approaching this as a pissing competition, the figures/screenshots are posted as an indication, I'll let them speak for themselves. :-)
Peace: |
Hey Vin,
The performance difference between N2 and N3 in this particular test originates from the Magneto plug. N3 has a newer version of it which seems to use considerably more CPU cycles. Some people, however, report they are not affected by this. But the difference in performance can be as much as 10%, so it's not insignificant.
Franjo
_________________
ASUS A8N-E / AMD Athlon 64 X2 4200+ / SuperTalent 2GB PC3200/ nVIDIA GeForce 6600 / PreSonus FireBOX on PCI-e FW card/ 2xUAD-1 /
Windows XP Home SP2
Cubase 4.0.1.2074 |
|
| Back to top |
|
|
TAFKAT
Member
Joined: 05 Apr 2005
Posts: 997
Location: Melbourne: Australia
|
| Posted: Thu Jun 16, 2005 12:48 pm Post subject: |
|
|
| franYo wrote: |
Hey Vin,
The performance difference between N2 and N3 in this particular test originates from the Magneto plug. N3 has a newer version of it which seems to use considerably more CPU cycles. Some people, however, report they are not affected by this. But the difference in performance can be as much as 10%, so it's not insignificant.
Franjo |
Hey Franjo,
O.K , you sparked my curiousity so I ran the benches on N3.0.2.23, and you were right , there is a marked difference.
I always maintained using N2 to have some corrilation with all of the earlier data I had compiled for my own in house records.
Also with some reports from BrianT and Bullmoon that they had slightly better performance with N3, I let it slide..
My apologies to Beat Production for going off half cocked :-)
Now the grid that you prepared comparing all of the posted results will have to be double checked with what Nuendo version they were running as well, but from memory, a lot of the figures were on N2, except perhaps the 2895 results, which will hold them in good stead , as there was some discussion that they were a little below par compared to the 2885 numbers that were posted earlier.
I also ran the benches in N3 with the older version of Magneto as well , ahhh don't you love it how Steini are going out arse backwards with the code optimisations lately... N3 feels like absolute treacle compared to N2 , I don't care what any of the apologists have to say..lets hope that 3.1 lifts the veil ... I won't even go into some of the Video related bullsh*te I have been trying to sort for some Post Clients.. :-)
I digress, here are the results..
PentiumD 830 :
1024 : N2 : 52 % N3 : 62 % N3 with N2 Magneto : 54%
512 : N2 : 54 % N3 : 64 % N3 with N2 Magneto : 56%
256 : N2 : 58 % N3 : 68 % N3 with N2 Magneto : 60%
128 : N2 : 65 % N3 : 80 % N3 with N2 Magneto : 70%
64 : N2 : 80+% Clicks-Pops N3 & N3 with N2 Magneto : No-Go
So there you have it , at the lower latencies N2 is far more efficient than N3 , even with the older Magneto .
The figures are still substantially better than Scotts 820 results, mind you I am a clock grade up.. !
I wonder what Version Guy was using ?
I can just picture that Hill Billy cackling like a MadMan, now that his X2 figures have just got a further polish... LOL !!
Out of interest we should get Scott to Post some X2 - N2 results.., they should be thru the roof !!
BTW: Everyone should also start posting the CitriQ results as well , just as a further reference point, this issue just goes to prove how variable these benchmark results can become when we are using results across Software Versions. In theory it shouldn't have made that much difference, but of course in practice, with the help of some sloppy code , it can add up very quickly. Going back and re-reading the earlier thread , it was made very clear that there was a problem with the current Magneto build , you would have thought some alarm bells would go off at HQ .., how hard would it have been to return to the previous build ?? I've gotta wonder just how many of the other PlugIn builds are dogs as well, luckily I don't have a use for the vast majority of the Nuendo plugs..
Either way, I'll let you do all of the analysing, re- analysing, grids, post mortems, clinical examinations, etc...
Peace: |
|
| Back to top |
|
|
franYo
Member
Joined: 06 Jan 2005
Posts: 369
Location: Croatia
|
| Posted: Thu Jun 16, 2005 1:53 pm Post subject: |
|
|
Hey Vin,
Thanks for the updated report.
| Quote: | | Now the grid that you prepared comparing all of the posted results will have to be double checked with what Nuendo version they were running as well, but from memory, a lot of the figures were on N2, except perhaps the 2895 results, which will hold them in good stead , as there was some discussion that they were a little below par compared to the 2885 numbers that were posted earlier. |
You're absolutely right. And to think I was blaming NF4 for the slightly worse results. LOL! It didn't even occur to me that most people are using N3 now. It would of course be good if everyone could re-run their tests with N2, but of course some people might have uninstalled it.
| Quote: | | Going back and re-reading the earlier thread , it was made very clear that there was a problem with the current Magneto build , you would have thought some alarm bells would go off at HQ .., how hard would it have been to return to the previous build ?? I've gotta wonder just how many of the other PlugIn builds are dogs as well, luckily I don't have a use for the vast majority of the Nuendo plugs.. |
According to your latest results, it's not just Magneto... I wonder if the performance difference between the N2 and N3 with N2 Magneto is due to other plug-ins putting higher load on CPUs than it's N2 equivalents or is it due to changes in Nuendo engine...
Franjo
_________________
ASUS A8N-E / AMD Athlon 64 X2 4200+ / SuperTalent 2GB PC3200/ nVIDIA GeForce 6600 / PreSonus FireBOX on PCI-e FW card/ 2xUAD-1 /
Windows XP Home SP2
Cubase 4.0.1.2074 |
|
| Back to top |
|
|
jcschild
Senior Member
Joined: 06 Jan 2005
Posts: 1245
Location: Kentucky
|
| Posted: Thu Jun 16, 2005 3:28 pm Post subject: |
|
|
| Quote: | I can just picture that Hill Billy cackling like a MadMan, now that his X2 figures have just got a further polish... LOL !!
|
Cackle, cackle, cakle LOL lmao too dang funny man, thanks for the morning giggle.
on the other hand i am itiching to go test now and i am only 1/2 awake.
nothing worse than a 1/2 awake hill billy for skewed test results.
_________________
Scott
ADK
Home of the "Kentucky Fried DAW" |
|
| Back to top |
|
|
Beat Productions
Member
Joined: 18 Jan 2005
Posts: 766
|
| Posted: Thu Jun 16, 2005 4:36 pm Post subject: |
|
|
I copied the N3 vstplugins folder to another location, then deleted it and copied the N2 vstplugins folder into N3 and ended up with very similar benchmarks between N3 and N2, so, I guess the performance drain is largely due to the newer vst plugin versions.
_________________
Comp1: Windows 7 64bit, Gigabyte GA-EX58-UD3R, Intel I7 920, 6 gig Patriot DDR3 ram, RME Digiface, Geoforce 66000 PCIe, , Cubase 5.5
Comp 2:Windows 7 64bit, Asus P5W DH Deluxe, Q6600, 8 gig Corsair XMS ram, , MR816 X, Saffire Pro 40, VE Pro |
|
| Back to top |
|
|
SONICA-X
Member
Joined: 28 Jan 2005
Posts: 169
Location: BOSTON, MA USA
|
| Posted: Thu Jun 16, 2005 6:50 pm Post subject: |
|
|
| TAFKAT wrote: | | franYo wrote: |
Hey Vin,
The performance difference between N2 and N3 in this particular test originates from the Magneto plug. N3 has a newer version of it which seems to use considerably more CPU cycles. Some people, however, report they are not affected by this. But the difference in performance can be as much as 10%, so it's not insignificant.
Franjo |
Hey Franjo,
O.K , you sparked my curiousity so I ran the benches on N3.0.2.23, and you were right , there is a marked difference.
I always maintained using N2 to have some corrilation with all of the earlier data I had compiled for my own in house records.
Also with some reports from BrianT and Bullmoon that they had slightly better performance with N3, I let it slide..
My apologies to Beat Production for going off half cocked :-)
Now the grid that you prepared comparing all of the posted results will have to be double checked with what Nuendo version they were running as well, but from memory, a lot of the figures were on N2, except perhaps the 2895 results, which will hold them in good stead , as there was some discussion that they were a little below par compared to the 2885 numbers that were posted earlier.
I also ran the benches in N3 with the older version of Magneto as well , ahhh don't you love it how Steini are going out arse backwards with the code optimisations lately... N3 feels like absolute treacle compared to N2 , I don't care what any of the apologists have to say..lets hope that 3.1 lifts the veil ... I won't even go into some of the Video related bullsh*te I have been trying to sort for some Post Clients.. :-)
I digress, here are the results..
PentiumD 830 :
1024 : N2 : 52 % N3 : 62 % N3 with N2 Magneto : 54%
512 : N2 : 54 % N3 : 64 % N3 with N2 Magneto : 56%
256 : N2 : 58 % N3 : 68 % N3 with N2 Magneto : 60%
128 : N2 : 65 % N3 : 80 % N3 with N2 Magneto : 70%
64 : N2 : 80+% Clicks-Pops N3 & N3 with N2 Magneto : No-Go
So there you have it , at the lower latencies N2 is far more efficient than N3 , even with the older Magneto .
The figures are still substantially better than Scotts 820 results, mind you I am a clock grade up.. !
I wonder what Version Guy was using ?
I can just picture that Hill Billy cackling like a MadMan, now that his X2 figures have just got a further polish... LOL !!
Out of interest we should get Scott to Post some X2 - N2 results.., they should be thru the roof !!
BTW: Everyone should also start posting the CitriQ results as well , just as a further reference point, this issue just goes to prove how variable these benchmark results can become when we are using results across Software Versions. In theory it shouldn't have made that much difference, but of course in practice, with the help of some sloppy code , it can add up very quickly. Going back and re-reading the earlier thread , it was made very clear that there was a problem with the current Magneto build , you would have thought some alarm bells would go off at HQ .., how hard would it have been to return to the previous build ?? I've gotta wonder just how many of the other PlugIn builds are dogs as well, luckily I don't have a use for the vast majority of the Nuendo plugs..
Either way, I'll let you do all of the analysing, re- analysing, grids, post mortems, clinical examinations, etc...
Peace: |
Vin,
now it all makes sense. My first D830 reports where about 10% higher than yours with N3 at home. The next day at the office I got much better results and that was N2. Now I think that the Matrox PCIe video card or the memory had nothing to do with the better benchmark.
I will retest with N3 and both video cards this weekend. I think my extreme will be back today and I have a 4800+ here tomorrow.
Guy. |
|
| Back to top |
|
|
jcschild
Senior Member
Joined: 06 Jan 2005
Posts: 1245
Location: Kentucky
|
| Posted: Thu Jun 16, 2005 11:47 pm Post subject: |
|
|
ok update with a differant card
firewire 410 N3 ...... from before Madi .......... madi 6800 video
1024 65% ...... 1024 70% ....... identical numbers as X300.
512 69% ...... 512 75%
256 78% ...... 256 80%
128 87% ...... 128 90%
this would now seem to line up alot better with Vin and Guy.
i am using the 820, they the 830.
i feel much better now... thought i was loosing my mind or my computer was broke.
it might also be that since my number from the 6800 and X300 are indentical the PCIe issue is not an issue (with Intel 955).
hmmm time to change the staement on my website it would seem
_________________
Scott
ADK
Home of the "Kentucky Fried DAW" |
|
| Back to top |
|
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
Powered by phpBB © 2001, 2005 phpBB Group
|
Product Page 
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
