Heavy Multitasking: Our typical workday at home.

The vast majority of our benchmarks are single task events that utilize anywhere from 23MB up to 1.4GB of memory space during the course of the benchmark. Obviously, this is not enough to fully stress test our 4GB memory configuration. We devised a benchmark that would simulate a typical home situation and consume as much memory without crashing the machine.

We start by opening two instances of Internet Explorer 8.0 each with six tabs opened to flash intensive websites followed by Adobe Reader 9.1 with a rather large PDF document open, and a nice game of video poker banished to the taskbar. We then open Bibble 5 with our standard test setup, and CyberLink Espresso with the YouTube HD conversion file, Microsoft Excel and Word 2007 with large documents, Hulu TV, and finally Photoshop CS4 with our test image.

We wait two minutes for all activities to cease and then start playing Legend of the Seeker via Hulu HD TV at 1280x720, start the photo conversion in Bibble, and then the HD transcode in Espresso. Our maximum memory usage during the benchmark is 3.37GB with 100% CPU utilization across the two threads.

Application Performance - MultiTask Test - Total Score

Application Performance - MultiTask Test - Bibble 5.0

Application Performance - MultiTask Test - Espresso

Thanks to the hardware decoding offload in Espresso, our AMD 785G DDR3 platform finishes our two tasks 40 seconds before the Intel G41 configuration or for our percentage trackers, AMD is over 9% quicker. Looking at the individual scores, the Intel system walks away in the Bibble test but falls short in the Espresso application.

We did experience a few stutters with the Intel G41 system during heavy action sequences in our HD video feed that were not present on the 785G even though our Espresso test sequence was off loaded to the GPU. Throughout all of our testing, it was this one test that impressed us the most with the 785G platform, especially its video playback capabilities.

Application Performance - MultiTask Test

While the Intel platform was 9% slower, it also consumed almost 8% less energy.

Video Content Creation Performance IG Gaming Performance
Comments Locked

43 Comments

View All Comments

  • MrCommunistGen - Tuesday, August 4, 2009 - link

    I was really curious about that section so I'm glad that I can actually view it. Thanks!
  • fic2 - Tuesday, August 4, 2009 - link

    I find the choice of a 10k RPM VelociRaptor odd for either HTPC or integrated graphic system. I have my doubts that it would be the HD choice of either user profile.
  • b15h09 - Friday, August 7, 2009 - link

    VelociRaptor because it eliminates a potential bottleneck. This isn't a real world system test.
  • Taft12 - Tuesday, August 4, 2009 - link

    Yes, I would expect a WD Greenpower drive or one of those new slow-spinning Seagates.
  • Fox5 - Tuesday, August 4, 2009 - link

    I remember roadmaps putting the 880G launching this month, but we're just getting the 785G. The 880G seems to be the chipset worth waiting for. Virtualized 3d hardware? Yes please.
  • fzkl - Tuesday, August 4, 2009 - link

    A lot of us know that when it comes to an HTPC, Intel is the worst option of the lot. To make things fair, why isn't this a 3-way roundup with the Nvidia GeForce 9300?
  • rtallmansu - Tuesday, August 4, 2009 - link

    Even more odd to me is why you would compare an Intel G41 with a ICH7 and not a G45 chipset with the newer ICH10. G41 buyers are not interested in any of these performance metrics, were as someone might want to know how the G45 compares for HTPC duties in HD playback.
  • Shaffan - Tuesday, August 4, 2009 - link

    Such a pity you did not test the RAID performances : RAID5 in particular. I heard the integrated RAID5 of Intel chipset is much better than the one of AMD, but I can't find a decent comparative test about this !
  • flipmode - Tuesday, August 4, 2009 - link

    SB710 = no RAID 5
  • mybook4 - Tuesday, August 4, 2009 - link

    Gary, you put in a lot of effort into this article, thank you! However, I found one part misleading.

    Most of the application benchmarks are dependent on only cpu, not gpu. It confused me that these benchmarks were in an article that compared 780G vs G41. I understand if you were trying to compare the platforms as a whole, but wouldn't that also constitute a component price match (price an amd 780g system with an equal priced G41 system then compare).

    PS. I am a little sick of people saying that anything under 60fps is unplayable. Most people that frequent this site play Crysis at under 40fps.

Log in

Don't have an account? Sign up now