Post 1 of 4
Hope im putting this in the right forum but i've just upgraded my intel core 2 duo E6600 to a core 2 quad E6600 and thaught i'd do a little benchmark test to see how much faster the new processor does it's job with a visionlab test clip, if anyone is interested in the results then read on.
My basic system setup is as follows:
EVGA 680i mobo,
2Gig OCZ SLI ready DDR2 running @ 1066mhz,
core 2 quad E6600 running 2 cores or 4 cores respectivly @ 1066mhz FSB,
250GB seagate barracuda SATA II drive,
750W silverstone SLI power supply,
Nvidia geforce 8800GTX.
Software setup as follows:
Windows XP (32 bit) fully updated,
Visionlab Studio (latest Build)
Media & effects used in test:
x1 clip @ 720 x 576, 16:9, interlaced, captured to disk as a .AVI using vegas and it's standard DV Widescreen codec from an old Sony HDR-HC3 filmed in standard DV wide.
x3 Single core lightning bolts customised from a preset & animated.
x1 muzzle flash preset with 'front projection' set to off, side projection set to on with 5 arms and flare also set to on.
The in-out markers were set at zero and 00:00:07:10 (7secs 10 frames or 185 total frames) respectivly for all tests, purpose of the tests were to see the difference in performance between 2 cores and 4 cores running the same task. To start each test, visionlab was shut down & started fresh so each test would theoreticly start in the same state, ie empty cache etc. Test one tests the time it takes to cache the full 185 frames into memory using the 'RAM' button (Preview Render Timeline Selection button), test two tests how long it takes to do a final render of the 185 frame region. No other applications were running during the tests.
Canvas display quality - set to Full
cache memory % (In preferences) - set to 60%
No of frames to render / cache - 185
Output Type - Single File
Render to - Directshow (avi)
Codec used @ render time - DV Video Encoder
Retain Sound and Alpha - Both set to Off / disabled
Background plate Scan method - Deinterlace(odd)
Background plate UV blur - set to zero
Running 2 cores only:
'RAM' cache - 1 min 39 sec's
Final Render - 2 min 08 sec's
Running all 4 cores:
'RAM' cache - 1 min 22 sec's
Final Render - 1 min 37 sec's
Im quite impressed with the final render 4 core result, thats 31 seconds faster than with only 2 cores running, I dont think thats too bad at all, any comments ???
Best wishes all