Our Take

Our first benchmarking tests indicate that buyers have some reason to be excited about NVIDIA's new 6100 Integrated Graphics chipset family. Despite the fact that serious gamers would find any integrated graphics board far too limiting, the great majority of boards that are used in systems are integrated graphics boards. In our head-to-head competition with the ATI RS480, the current integrated performance king, the bottom-of - the-line 6100 outperformed the ATI in almost every benchmark. That certainly means that the 6150, clocked at 475MHz instead of the 6100's 425, should perform even better.

There are always some exceptions. Not surprisingly, ATI integrated graphics is still best in games optimized for ATI, like Half Life 2. However, the 6100 is close enough to the ATI that the 6150 may even obliterate that advantage. Far Cry, optimized for NVIDIA, performs significantly better on NVIDIA 6100 than ATI RS480. NVIDIA performs best in every other game that we tested, and it was the top performer in both 3D benchmarks and the General Performance PCMark2005.

It should also be pointed out that NVIDIA is just introducing AMD integrated graphics, while ATI has integrated graphics solutions for both AMD and Intel platforms. The great majority of integrated graphics boards are now based on the poorer-performing Intel platform, so that fact alone will keep ATI's market share of integrated graphics high for the time being.

While we are excited about the improved integrated graphics performance within the NVIDIA 6100 family, this is not to say integrated graphics have completely arrived. Who would really want to play Doom 3 at 24FPS at 800x600 - and that's with no eye candy. However, by lowering detail and resolution, you should be able to find a playable 640x480 or 800x600 in most games. However, if you want better detail or higher resolution, you need to use a better video card. There is even good news here as NVIDIA mirrors ATI in now giving the user the option to run integrated graphics and a PCIe video card at the same time.

The Biostar TForce 6100-939 may not use the top-line 6150/430 combo, but it does extremely well with the 6100/410. The overclocking and memory tweaking options were surprisingly good. The Biostar has enough flexibility to satisfy many users, and it even has the voltage adjustment options that seem to be the last thing to appear on value boards. The TForce 6100 was fast, stable, trouble-free in our brief testing, and extremely flexible - particularly for a micro ATX integrated graphics motherboard. This Biostar would make a great foundation for a cheap system with decent performance, though it is missing the desired options that would make it a good multimedia box.

All-in-all, the NVIDIA 6100 is a decent integrated graphics solution and the new performance leader in AMD integrated graphics. It would have been even better if NVIDIA had made it 4 pixel pipelines instead of two, but the performance and options at the higher end does make the NVIDIA first choice among AMD integrated graphics solutions - at least until the next round from ATI.

Gaming Performance
Comments Locked

43 Comments

View All Comments

  • reload1992 - Wednesday, February 27, 2008 - link

    I am building a pc for the first time Is this a good motherboard for gaming?
  • Calin - Monday, October 10, 2005 - link

    I would like to change my mainboard/processor (as they are still in the 600MHz range), and I really would like one of those things. However, not wanting to pay the whole extra $100 for a socket 939 processor, I would like to know the performance of the Socket 754 board (coupled with an Sempron processor).
    In case your plans does not accout for making tests with the dual channel memory board, could you please test it with a single DIMM (in order to simulate a 754 board on the hardware you have)?

    Thank you very much
    Calin
  • varundubey - Tuesday, September 27, 2005 - link

    Hi, in your article here:
    http://www.anandtech.com/mb/showdoc.aspx?i=2539">http://www.anandtech.com/mb/showdoc.aspx?i=2539

    you say that NVIDIA 6100/430 and NVIDIA 6100/410 support pure video but Nvidia disagrees here:
    http://www.nvidia.com/page/gpu_mobo.html">http://www.nvidia.com/page/gpu_mobo.html

    should be a tad more careful no?
  • glennpratt - Wednesday, October 5, 2005 - link

    NVIDIA has corrected the link.
  • Wesley Fink - Thursday, September 29, 2005 - link

    The chart showing Pure Video support for the 6100/410 is cut and pasted from nVidia launch literature for the 6100. If it's wrong or in conflict with other nVidia publications please talk with nVidia for clarification.
  • Phiro - Tuesday, September 27, 2005 - link

    Varun, they've abandoned us to fend for ourselves.
  • LoneWolf15 - Monday, September 26, 2005 - link

    My best use for this would be an nForce 430/GF6150 mATX board for a HTPC setup. I'll be eagerly waiting to see how this does. It'd be great to use an entry-level Sempron or A64 as part of a small, quiet rig for this sort of application, and the onboard HD audio, gig Ethernet (for transferring files from a media server across the intranet), and HD over TV-out are all big plusses.

    I'll be waiting to see this tested when it becomes available, as I'm once-bitten-twice-shy on the nVidia PureVideo thing, but if it can meet or beat the on-paper specs, it looks to be a winner.
  • Phiro - Friday, September 23, 2005 - link

    Question - was the HL2 benchmark run under the DX8 or DX9 codepath?

    I ask because the numbers to me seemed to compare the 6100 to the Geforce 5700. You show the 6100 getting 55fps at 800x600 in HL2 with normal/low settings. The 5900XT gets 104fps with highest settings w/o AA in a similar setup, until I read closely and saw that was with the DX8 codepath. Forcing the 5900XT to use the DX9 codepath knocks it down to ~30fps if I recall.

    If that's true, and the 6100 was run under the DX9 codepath and it got 55fps with lowered settings, that still puts this video card a big chunk above a 5700 IMO.
  • Phiro - Monday, September 26, 2005 - link

    Wesley, you can't quit reading the feedback from an article you write less than 24 hours after you post said article.

    Come on, fess up w/the information!
  • Wesley Fink - Thursday, September 29, 2005 - link

    I'm not the graphics guru, but I thought the DX9/DX8 codepaths were an early HL2 concern long since resolved. At any rate we run HL2 thorugh Steam and update before benching. The system is XP SP2 9.0c. The bench utility is the Guru3D HL2 benchmarking tool running Guru3D Demo5. The HL2 update from the last few days has corrupted all of of HL2 benchmarks - including the Guru3D tool - but that's another story.

Log in

Don't have an account? Sign up now