Xeon Nehalem benchies

[Jcschild]

finally

c4 32 bit 4.52
Vista 64 biz SP1
12 gig ram
2x 5550 2.66mhz HT on


rme raydat PCI-e drv 3.077 (and UADs installed)

this is the blowfelds C4 test not the universal with

512 -248

256- 237

128 - 227

64 - 180

32 - 147

compared to single i7 940 (2.66GHz) HT enabled (fireface 800)

512 - 148
256 - 139
128 - 132
64 - 128
32 - 105

will try to get the newer test.

Scott
ADK

[Andrew J]

:eusa_whistle:

Thanks Scott. How much of the difference can you attribute to RayDat PCI-E vs FF800?

[TAFKAT]

Thanks Mate..

I'll have a dig thru these numbers a little later...

Quick edit : Good point made by Andrew, missed that on the first browse.

Once you run up DAWbench DSP Universal on the RYADAT, we will get a better idea, as I have the i7 / RME numbers there.

:wink:

[Jcschild]

:eusa_whistle:

Thanks Scott. How much of the difference can you attribute to RayDat PCI-E vs FF800?

very little based on past experiance.

[ROCKINROG]

Have either of you tried any tests with Windows 7? I haven't got around to testing the betas with anything yet. I only checked it out on one of my main computers just to see what's different then re-imaged back. I did notice it was a bit of a resource hog (over 500Mb) and alot of services and name changes of things that confused the hell outta me. :eusa_think: Hopefully I'll get around to checking it out further.

[Jcschild]

just for gigles

HT OFF.....

c4 32 bit 4.52
Vista 64
12 gig ram
x 5550 2.66mhz


rme raydat PCI-e drv 3.077


512 -206

256- 208

128 - 190

64 - 156

32 - 134

[Jcschild]

Vin,
please update your graph and send me the file so i can post (with your permission of course on my site as well)
rather than me use all the data i compiled using the old test...


c4 32 bit 4.52
Vista 64
12 gig ram
x 5550 2.66mhz
DSP universal this time

rme raydat drv 3.077


HT on

512 - 265

256- 248

128 - 231

64 - 214

32 - 172



HT Off

512 - 211

256- 197

128 - 190

64 - 171

32 - 147

[TAFKAT]

Hey Scott,

Awesome, thanks for that.. :-)

I'll update the graph and send you a copy, no problem.. :-)

Just make a note that all of the other results are on XP , but no matter, still shows a significant improvement in scaling over the last gen.

For a bit of fun, if you have a 2.66GHZ - Nehalem MacPro to run some number against, that would be interesting..

[Jcschild]

did you want that with XP bootcamp or OSX?

[TAFKAT]

Hey Scott,

Heres the updated graph.., I have sent you the Excel file with the raw data so you can knock up your own in a different format if you prefer..

http://www.dawbench.com/images/dawbenchdsp-rxc-may09-2.jpg

Wow its a pity Cubendo "officially" doesn't work on Dual Quad or with Hyperthreading , otherwise these numbers would be something to be singing about, we can only hope that Steini can resolve what ever problem they are still " looking into "..... :eusa_whistle:

[TAFKAT]

did you want that with XP bootcamp or OSX?

Hey Scott,

XP/Vista bootcamped would be interesting, as we have data to directly compare with DAWbench DSP-RXC, with OSX you will need to use either Blofelds DSP II , or the MD5 version of the Universal Test.

What ever is easiest for you.

You already have the Blofelds II results in Vista 64, so for the OSX comparison on MacPro in OSX that would probably be easier for you..

Its starting to feel like the old days, except this time we are not on opposing sides of the fence..

So wheres BT , we need some AMDian superiority frothing... LOL

[Jcschild]

i was thinking more along the lines of our other friend
you know 16 cores and all that...:rotfl::rotfl::rotfl:

[TAFKAT]

... you know 16 cores and all that...

..not to mention the Top Secret ingredient.. :eusa_whistle:

[Jcschild]

no no the special case

[TAFKAT]

.. thats what I meant by the "ingredient", I was trying to be subtle... :wink:

May has well be special sauce for all the difference it will make..:eusa_whistle:

[Softy]

.. thats what I meant by the "ingredient", I was trying to be subtle... :wink:

May has well be special sauce for all the difference it will make..:eusa_whistle:

I'm trying to not let it drive me crazy that I don't know who it is that you're talking about here.

[TAFKAT]

LOL,

Its an inside joke.., I'll send you a PM... :D

[Jcschild]

might be awhile in those Apple benchies.. just shipped several out the door
the next 2 are sorta rushes and my demo unit is still the old one...

the options for demo this go round is
1) single quad (cheap)
2) dual 2.93GHz etc etc i dont need a 6k demo box....

so i will need to grab numbers from a sold system which is kinda of a pain...

[TAFKAT]

Hey Scott,

Don't stress it .., when ever you get a spare moment on a system , without it being a PITA.. :D

[Jcschild]

not having a demo Nehalem is also why i got caught with my pants down on Nuendo...

[TAFKAT]

Hey Scott,

I hear you, its a tough one to have a 6K piece of kit around to do some qualifying , especially if they don't turn over fast enough..

I got caught with Dual Quad which I ended up losing many thousands on in the end , and with prospects being pretty thin due to Steini's "official" insistence that it is still broken makes it even less viable to invest the time/dev at the moment..

Don't give the benches another thought Mate, in this climate, better to keep to whats more important..

:009:

[Jcschild]

good news...

i forgot i loaned Cineform an Apple Nehalem system for NAB and then they needed to borrow it after the show for a few weeks to test/dev their software on it.

on its way back to me today should have it in a week or less.

benchies forthcoming...

[TAFKAT]

Cool,

Thanks Scott..

:D

[Jcschild]

sorry for the delay. we have been in laptop hell for weeks.

so here is both

Intel mac Pro E2270 2.27ghz
12 gig ram
vista 64 sp1
rme fireface 800 drv 2.95
cubase 5.01 32 bit

blofeld mk 2 C4 test

512 - 216
256 -200
128 168
64 142
48 -120


mac pro
osx 10.5.7
rme fireface 800 drv 2.95
cubase 5.01 32 bit

512 - 146
256 - 138
128 - 98
64 - 78
48- 56

[TAFKAT]

Thanks Scott,

Just a quick note for clarification..

The Xeon CPU's you listed, did you mean E5520 ?

The numbers are confirming that Cubendo still sucks on OSX compared to Windows , so I can't see the geniuses at Steinberg moving off their official stance of it being broken across the board... :icon_no:

I'll knock up a graph when I get a chance and drop that into the cross platform Multiprocessor X-scaling summary I posted Here (http://cubendo.com/showthread.php?t=241)

[Vinark]

:pop_corn:
(:No irony intended:)

[TAFKAT]

Here are the results in graphical form..

http://www.dawbench.com/images/blofeldsdsp2-xp-v-osx-0609.JPG