![]() |
|
#1
|
||||
|
||||
![]() Quote:
Hope the kids don't wear you out! ![]() Flyby out btw, how a sim runs with either Crossfire or SLi implementation is not a function of the sim, but rather of the GPU's technology. There have been some interesting articles around talking about a new technology that allows multiple GPUs to scale nearly 100%. It's called Hydra and it's been tested, but is still in development. Neither ATi nor Nvidia seem inclined to fix their multiple-GPU tech. It's really sad too because people spend their hard-earned bucks to buy two GPUs because of the implied promise that two work twice as fast as one (universally) and that's just not true. Which is also why some sims run better in SLi (or Crossfire) than others. IMO, Hydra may represent what ATi and Nvidia should have been about all along:true multi-GPU scaling no matter what the sim.
__________________
the warrior creed: crap happens to the other guy! Last edited by Flyby; 12-04-2008 at 01:56 PM. |
#2
|
|||
|
|||
![]() Quote:
There is some amazing technology being released and I guess it's a bit frustrating when it doesn't realise its promised potential or is so new that there are no applications that use it fully. Cheers and all the best! |
#3
|
||||
|
||||
![]()
the best back atcha!
![]() Flyby out
__________________
the warrior creed: crap happens to the other guy! |
#4
|
|||
|
|||
![]()
For those with interest:
Gigabyte EP35-DS4, bios ver F3 E8400 (2x 3.0G) , 2 Gig HyperX DDR2 Ram 1333, BFG 8800 GT OC 512 Meg, 625/900/1560 default (this default GT OC is already approx 5 %) VSYNC on at 60 (Max) Using Gigabyte Easy Tune Pro 5.0 to perform the overclocks, and confirmed with GPU-Z. Il-2 2.04 Video Mode: ALL HIGHEST OPTIONS Video Option: 1280 x 960 x 32 A couple FPS change is really meaningless as I did not run each 10 times and take an average. With multiple runs of same settings one might see two FPS change between the runs. RESULTS FOR BLACK DEATH ( min:sec FPS ) 1. ALL STOCK CPU and Vid Card 1:07 56 1:45 49 2. CPU @ 3.3 G: 1:07 54 1:45 49 3. CPU @ 3.3 + VID CARD 630 / 930 1:07 54 1:45 49 4. CPU @4.0 G + VID CARD 640 /950 1:07 52 1:45 49 CONCLUSION: OC CPU 17% combined with total 7% OC of Vid Card (over non OC version) seems to have no benefit. I used GPU-Z to confirm the OC values. Undoubtedly there will be significant increases in heat buildup + reduced max theoretical life expectancy of the chip + increased energy consumption. Interestingly from the above metrics my FPS appears to drop when OC'ing.<< I certainly was not expecting that, was expecting either no change or a few frames of trending upward. Note that I have another sys (old specs) which was AMD (one core) 2.2 G + 6800 Ultra. Stepping up to my current PC specs was significant in terms of visuals / smoothness / etc. THe MINIMUM FPS in IL-2 were significantly improved (perhaps 2x if my memory serves). After reading the data on many of the new tech CPUs/VIDCARDS/MEMORY it is not of use nor value for me to spend any more money to improve my gaming. I am perfectly satisfied with my current rig gaming capacity. My min FPS are sufficiently close to 60 (LCD MAX CAPACITY). All the I7 MOBOS have had significant problems as shipped to consumers. This new platform will need at least a year to mature to the point of being a viable / reliable / stable option. The technology on paper of the I7 shows a significant potential step up but in my opinion it is not ready for market nor my $$$ (this is pretty typical). My testing for your benefit, no cost to you. I also play LOCKON and BF2. Note that LOCKON is also a huge resource hog as is IL-2. I have OC'd before and recall similiar non benefits for LOCKON as were presented above for IL-2. Now I will take the CPU and VID CARD back to their native settings. If I can just get my buddy SimDee to recover from his latest flight experiences with me on his 6 perhaps I will continue to home my pitiful flight skills. Peace Last edited by SPUDLEY1977; 12-07-2008 at 06:17 PM. |
#5
|
||||
|
||||
![]()
then there's SoW_BoB which may be coded to work with multi-cored cpus. But perhaps your GPU was becoming a bottleneck during your overclocking, or maybe the ram? I don't know, but it looks like the cpu was waiting more and more when it was overclocking. Too bad I'm not an expert on such things. So it's just my take on your test data. Or maybe it is the PCI-e 1.0 P35 chipset with limited PCI-e bandwidth compared to say a P45 which supports PCI-e 2.0 with greater bandwidth pass-through? Maybe I'm an optimist?
![]() At any rate, are those the average FPS for the BD track? Pretty damn solild if they are, if you ask me. Well done! ![]() Flyby out
__________________
the warrior creed: crap happens to the other guy! |
#6
|
|||
|
|||
![]()
At 1min 7 sec and 1 min 45 seconds into the track I see my lowest FPS in the Death Track with the default views and the settings as indicated in the post. So those are the lowest FPS at the two respective times. Note that VSYNC caps at 60.
Unknown which if any one particular component is the sole bottleneck. When I build em I do my research to ensure I don't make the mistake of not properly matching the capacity of the critical components. Otherwise the upgrade/build is a bit pointless. ![]() |
#7
|
||||
|
||||
![]()
rgrt, Spudley,
I apperciate your work and the fact that you aren't charging for it! ![]() ![]() Flyby out
__________________
the warrior creed: crap happens to the other guy! |
![]() |
|
|