new HD x264 video encoding benchmark

[The Nemesis]

First run @378x9 DDR2-757 was 3-3-3-10 1:1. The run @ 400x9 DDR2-1200 is 5-5-5-15 2:3 divider.

[graysky]

@KTE - cool man, XP 32 or 64?

[KTE]

Sorry, 32-bit for both.

Added above for clarity.

[CyberDruid]

---------- RUN1PASS1.LOG
encoded 1442 frames, 51.74 fps, 3904.94 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 51.29 fps, 3904.94 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 32.94 fps, 3904.94 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 51.03 fps, 3904.94 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 13.12 fps, 3974.08 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 13.07 fps, 3974.08 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 13.14 fps, 3974.08 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 13.07 fps, 3974.08 kb/s

I did the run while Folding two SMP...

I'll see if it makes any difference with Folding closed.

2 x E5420 stock on D5400XS 8GB FB DIMM DDR2 800 4 Raptor RAID0

Same settings with Folding closed

---------- RUN1PASS1.LOG
encoded 1442 frames, 57.91 fps, 3904.94 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 58.18 fps, 3904.94 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 58.17 fps, 3904.94 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 58.11 fps, 3904.94 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 28.09 fps, 3974.08 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 28.13 fps, 3974.08 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 28.22 fps, 3974.08 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 28.15 fps, 3974.08 kb/s

[graysky]

@CyberDruid - nice system dude. FSB = 333 and RAM = 800? So 5:6 divider then? Also, which O/S and 32 or 64 bit?

[CyberDruid]

Sorry forgot to specify. This board is weird the RAM always runs at DDR2 800...no way I can see to OC it...raising the FSB does not seem to change the RAM speed. Not sure how that works to be honest...I've got very limited OC-ability on the Xeons...I can jack them to 7.5 x 389.

They are running at 7.5 x 333 and I am using Vista Ultimate 64 bit.

[CyberDruid]

---------- RUN1PASS1.LOG
encoded 1442 frames, 68.82 fps, 3904.94 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 69.70 fps, 3904.94 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 69.49 fps, 3904.94 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 69.18 fps, 3904.94 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 33.26 fps, 3974.08 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 33.30 fps, 3974.08 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 33.34 fps, 3974.08 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 33.38 fps, 3974.08 kb/s

That's with the Xeons clocked to 7.5 x 388 CPU-Z does not read the RAM.

[graysky]

@cyber - is it still 5:6 or since you said it runs @ 800, then 1:1?

[G H Z]

Here's one of my rigs.

[dkx64]

Q6600
Vista x64
x38 chipset
5-5-5-15 SPD

@3.33GHz

---------- RUN1PASS1.LOG
encoded 1442 frames, 66.50 fps, 3904.67 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 66.19 fps, 3904.67 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 67.14 fps, 3904.67 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 67.18 fps, 3904.67 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 18.55 fps, 3952.97 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 18.46 fps, 3952.97 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 18.47 fps, 3952.97 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 18.80 fps, 3952.97 kb/s

@3.60GHz

---------- RUN1PASS1.LOG
encoded 1442 frames, 71.49 fps, 3905.42 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 72.03 fps, 3905.42 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 71.66 fps, 3905.42 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 56.30 fps, 3905.42 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 20.22 fps, 3952.85 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 20.27 fps, 3952.85 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 20.30 fps, 3952.85 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 20.25 fps, 3952.85 kb/s

I was loading another program during run4pass1, thats why it dropped to 56
It looks like Vista is indeed slower than XP reading these results.

[nosboost300]

question... how do you aquire a skulltrail mobo.... i'd love to get a hold of one w/ two xeons

[Toysoldier]

---------- RUN1PASS1.LOG
encoded 1442 frames, 73.65 fps, 3905.42 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 74.55 fps, 3905.42 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 74.43 fps, 3905.42 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 74.55 fps, 3905.42 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 20.35 fps, 3952.85 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 20.38 fps, 3952.85 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 20.40 fps, 3952.85 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 20.42 fps, 3952.85 kb/s

[Gen]

It looks like Vista is indeed slower than XP reading these results.

+1, by looking at all the Q6600's @3.6 XP looks to be 3-4fps faster.

i rebenched with firewall, a/v, some processes and everest all disabled and still, with Vista x64, got the same fps.

[graysky]

@G H Z - thanks for the results... what chipset is your board and what o/s (also 32 or 64 bit)?
@dkx64 - first run is 9x370 and 2nd is 9x400? Also, what is your mem speeds for each run (not timings but speed)?

[graysky]

@toysoldier - got it thanks!

[graysky]

+1, by looking at all the Q6600's @3.6 XP looks to be 3-4fps faster.

i rebenched with firewall, a/v, some processes and everest all disabled and still, with Vista x64, got the same fps.

Yeah, vista is pretty, but man is it bloaty too. My XP system boots up with about 200 megs of the SF used. The same system under Vista boots up with 800 megs!

[dkx64]

@dkx64 - first run is 9x370 and 2nd is 9x400? Also, what is your mem speeds for each run (not timings but speed)?

I remember 3.60 was 9x400 @ 1066mhz ram.

I forget what the ram speed was at 3.33, but i've ran it again at 3.42. Use these results instead:

Q6600 @3.42 9x380 - 1013mhz RAM

---------- RUN1PASS1.LOG
encoded 1442 frames, 69.82 fps, 3905.42 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 68.52 fps, 3905.42 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 69.19 fps, 3905.42 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 69.29 fps, 3905.42 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 19.30 fps, 3952.85 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 19.33 fps, 3952.85 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 19.33 fps, 3952.85 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 19.32 fps, 3952.85 kb/s

[graysky]

got it man, thanks!

[Gen]

i like this benchmark, really goes to show how little significance ram clocking over 1:1 has.

looking at the Q6600 9x400 results having DDR2-800 44412 ram vs. DDR2-1200 55515 ram only equals 1.x fps increase.

this benchmark along with a few others in this section really convinced me ram o/c past 1:1, with Conroe, is fairly useless in real world apps.

[graysky]

@Gen - yeah, that's certainly the conclusion I drew from the data. Glad to see people are playing with it.

[The Nemesis]

i like this benchmark, really goes to show how little significance ram clocking over 1:1 has.

looking at the Q6600 9x400 results having DDR2-800 44412 ram vs. DDR2-1200 55515 ram only equals 1.x fps increase.

this benchmark along with a few others in this section really convinced me ram o/c past 1:1, with Conroe, is fairly useless in real world apps.

Yeah I been looking at that myself. I submited the 9X400 5-5-5-15 run. I'm gonna do a 9x400 4-4-4-10 & 3-3-3-10 run. It will be useful to see how the aged P965 compares to the P35, X38, Nvidia offerings.

[Erdemliol]

Test System->>

E2160 M0@3.4 ghz
2x1 GB Twinmos Ddr2 800@1020 @5 5 5 15


---------- RUN1PASS1.LOG
encoded 1442 frames, 36.72 fps, 3905.08 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 36.75 fps, 3905.08 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 36.75 fps, 3905.08 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 36.84 fps, 3905.08 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 9.15 fps, 3942.92 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 9.22 fps, 3942.92 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 9.19 fps, 3942.92 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 9.14 fps, 3942.92 kb/s

[CyberDruid]

@cyber - is it still 5:6 or since you said it runs @ 800, then 1:1?

My guess is that at FSB 333 I am DDR2 667 and as I get closer to FSB 400 I am finally at DDR2 800

I did manage to get the board to run at 400FSB for a while but it would occasionally fritz so I backed down again. I'll post a run at 7.5 * 400

[Gen]

the ellusive 4ghz run evaded me this weekend but 3900mhz was pretty stable.

---------- RUN1PASS1.LOG
encoded 1442 frames, 73.89 fps, 3905.42 kb/s

---------- RUN2PASS1.LOG
encoded 1442 frames, 75.40 fps, 3905.42 kb/s

---------- RUN3PASS1.LOG
encoded 1442 frames, 74.67 fps, 3905.42 kb/s

---------- RUN4PASS1.LOG
encoded 1442 frames, 75.09 fps, 3905.42 kb/s

---------- RUN1PASS2.LOG
encoded 1442 frames, 21.71 fps, 3952.85 kb/s

---------- RUN2PASS2.LOG
encoded 1442 frames, 21.71 fps, 3952.85 kb/s

---------- RUN3PASS2.LOG
encoded 1442 frames, 21.73 fps, 3952.85 kb/s

---------- RUN4PASS2.LOG
encoded 1442 frames, 21.70 fps, 3952.85 kb/s

[graysky]

@Erdemliol - thanks for the data... what is your multiplier and FSB. Also, what O/S and is it 32 or 64 bit?
@Gen - got your new results, thanks... if your first post you said you were running 9x450=3600... is that a typeo? Was it really 9x400=3600 or 9x450=4050?