preliminary Nehalem numbers

[jcschild]

Hey Guys,

well finally got a nehalem up and running. normally we get beta versions much sooner than this. (months sooner)
first board Intel sent us was DOA. 2 weeks later we got another so here they are.

Cubendo coming next

Mikes Sonar6 test on Sonar7PE (x32) (64 sucks, wont play back 128)
X58
DDR3 1600 (6GB, They are now put in in 3's)
i7 3.2GHz
RME Fireface
Vista 64 (dont ask) but bear in mind this will skew previous #s
oh wait i found some internal #s, for Vista 64 i dont think we posted...

1024 - 11%
512 - 11%
256 - 13%
128 - 17%
64 - 28%
48 - 32%

All Clean


previous numbers (from Jan)

VISTA 64 NUMBERS using SONAR 7.02
QX 9770, 3.4Ghz, 4GB, X38
(Scott's machine)

SONAR 32 BIT
1024 - 18%
768 - 16%
512 - 18%
256 - 22%
128 - 29%
64 - 42%
48 - 52% pops

SONAR 64 BIT
1024 - 21%
768 - 25%
512 - 26%
256 - 37%
128 - 56% - light pops
64 - 94% - pops, glitches, bad
48 - no way

Sonar6 Test (ver 6.21) XP Pro
QX 9770, 3.4Ghz, 4GB, X38
(Scott's machine)

1024 - 16%
768 - 16%
512 - 20%
256 - 20%
128 - 24%
96 - 29%
64 - 38%
48 - 43% (very slight pops in audio)


Sonar6 Test, Sonar 6.21, RME MADI XP Pro
Xeon 5440 @ 2.83 Ghz, 4GB RAM

1024 - 14%
512 - 15%
256 - 16%
128 - 19%
64 - 31%
32 - 50%*** ***Clean Audio, but audio engine drops out after a few seconds. Very unbalanced CPU load***



so let pull out the important #s

Nehalem 3.2GHz 64 28%

Q9770 3.4GHz 64 42% (vista) 38% XP

8 Coe Xeon 2.83GHz 64 31% (XP)

so thats a large improvement.

in testing with Sisoft Sandra the numbers were off the chart.
i cant wait until Mike finishes the Komplete test for samples

oh wait here he is now... i will go and update the VSTi thread
it did 64 Buffer clean. and again bear in mind this is Vista (the HOG)

Scott
ADK

[radiant]

Thanks for the info Scott!