TG2bench - Terragen 2 Rendering Benchmark

[Precog]

Hi,

A new TG2 benchmark...

   http://tg2bench.kk3d.de/

Simply follow the link, it's self-explaining. :-)

- Karsten

[3DGuy]

Hmm seems my single core AMD4000+ beats a Core Duo2 E6600 and AMd64 4200 X2. Bench took 3:52s on my system.

[Njen]

Cool. And I am very happy with my laptop, clocking in at 4:20. Just 11 seconds slower than the Core Duo2 E6600.

Of course I am sure these times will have to be redone once multi threading is activated in TG2

[king_tiger_666]

my 2.8ghz p4 took 6mins 50sec no wonder for my slow renders... also seems to run slow on my cpu looking at the list

[king_tiger_666]

seems like TG2 performs better on amd over all a friends score Opteron 165 2.7GHz 1GB DDR1-500
benchie 3 min 37 seconds

[Ricowan]

Hmm seems my single core AMD4000+ beats a Core Duo2 E6600 and AMd64 4200 X2. Bench took 3:52s on my system.

Are you overclocking that CPU?  My X2 4400+ is only reporting 2.21 GHz, so I can't figure out how your 4000+ is 2.4 GHz. 

By the way, I got 4:25 while downloading a 6gb file from fileplanet and two other files with BitTorrent.  I'll try again when I can turn off all of the extra activity.

Rich

[3DGuy]

No, not overclocking. Nothing is overclocked in my system. These day I prefer stability over a few percent performance gain The 2.4Ghz is standard for this 4000+ Single core (San Diego). Here's a screenshot of an diagnostic program (Everest home edition). This CPU is multiplier locked and as you can see it's running at the stock 200Mhz FSB.


The 4400+ is a perfomance rating. Your CPU is built differently than mine and should theoretically perform as a classic AMD Thunderbird running at 4400MHz. But as you see from this benchmark, that's not completely true, at least for single-threaded applications. In that case the single core CPU usually 'wins'.

[Moose]

I've rendered the test a number of occasions all with different times. Is this due to my system setup or something to do with TG2 (GI ?). My times ranged from 4.04 - 4.09 on an Athlon 64 3800+. I'm reluctant to submit my result however (which one would I use anyway - averaged I suppose), before knowing exactly what's causing the disparity.

Would it be possible for you guys to try the benchmark again to see if your times remain the same.

Thanks

[MeltingIce]

Hmm seems my single core AMD4000+ beats a Core Duo2 E6600 and AMd64 4200 X2. Bench took 3:52s on my system.

That makes perfect sense.  TG2 isn't optmized for dual core processors yet, so it's only using 1 of your cores in your X2 processor.  The single core AMD is faster because TG2 is utilizing the whole processor and since its single core, it has a higher clock speed than each core in your X2.

[3DGuy]

Yeah I know that It's just that it's an older processor with a lower speedrating. I would like to see the results when T2 gets multicore support. And I don't have an X2, I got that from the benchmarkpage.

[MeltingIce]

Yeah I know that It's just that it's an older processor with a lower speedrating. I would like to see the results when T2 gets multicore support. And I don't have an X2, I got that from the benchmarkpage.

Ohh gotcha, my bad.  Yea I can't wait until TG2 actually uses both cores, then maybe my laptop will beat my desktop 

[king_tiger_666]

strange my 2.8ghz 1gb ram is slower than my sempron 3000+ with 256mb 2700 ram both xp home. anyone had a try without hyperthreading enabled?

[Oshyan]

Variations of a few seconds are not unusual on this kind of test. That's why it's important that the test be reasonably long to reduce the significance of these errors (if the error is 5 seconds and total time is only 10 seconds, that's a 50% margin of error!). I would submit the lowest time you achieved, unless you care to do the average. Most are probably using the lowest though (I know I will ).

It's not surprising that a Sempron 3000+ would beat a P4 2.8. Terragen 2 isn't yet optimized for SSE/2/3 and it relies heavily on traditional floating point math, which the Athlon core handled much better than the P4. Hyperthreading isn't likely to make a positive difference since TG2 doesn't yet support multithreading in the renderer.

- Oshyan

[xanga433]

Does Windows XP 64-bit edition have a big effect on rendering time?? Because...I get render times ranging from 0:03:51 to 0:03:55 and my specs are somewhat worse than a few people I see on the list.   

[3DGuy]

There's many factors that determine the render speed. CPU is ofcourse a main factor, but memory speed is also important.If you have very fast memory (i.e. 800Mhz or 1000Mhz) that will also help the time because of the large amounts of data being generated.