Reply 160 of 430, by nd22
And lastly, the KD7A with VIA kT400A chipset at the 3 resolutions
And lastly, the KD7A with VIA kT400A chipset at the 3 resolutions
AT7 with KT333 makes a comeback and wins at all 3 resolutions tested! I retested 4 times and the results stay the same. Maybe KT333 is better suited for old benchmarks? We shall now test that assumption with the next benchmark: 3d mark 2001 which uses DirectX 8! Also we see clearly in this particular test that all 3 systems are strongly bottlenecked by the CPU!
First in 3d mark 2001 is the AT7 based system. Again tested at 1024*768/1280*1024/1600*1200
Next is the At7-MAX 2 bases system; again 3 resolutions tested!
And finally the system with the KD7A board:
This time there is no winner; all 3 chipsets produced results that are really close! 3d mark 2001 is considered one of the best system benchmarks because every setting/component influence the result such as loose memory timings, incorrect settings in the BIOS and so on. You would expect that a difference of at least 5% should exist but there no such thing! Remember, the memory controller is inside the northbridge and results back in the day indicated jumps of around 3-4% when testing KT400 and again 5% when testing KT400A.
I come back with a correction! I could not understand why KD7A has such a large advantage when testing with sound against the other 2 boards so I went down in storage and looked up the boards when I saw right next to KD7A the next one: KV7 which represents KT600 chipset and it suddenly hit me! They share the same PCB, only the Northbridge and Southbridge are different! VIA did something really good when designing chipsets: they are pin compatible – you can drop KT400/KT400A/KT600 in the same board without redesign; you can do the same with VT8235 and VT8237 Southbridge! And because Abit made the KD7A and KV7 at the same time they share most of the components including the audio codec! Which is not a Realtek one but from VIA: VT1616. Still far from Creative but better than Realtek. It is a shame that VIA did not choose to go on because this codec is really good! It also shows why it is better to make screenshots every time; if I had an Everest print screen I could easily see why KD7A jumps forward in the sound tests!
We now go to the next benchmark: 3d mark 2003 - DirectX 9!
As usual the first is AT7 with KT333 at 1024, 1280, 1600:
Now is the turn of AT7-MAX 2 with KT400:
And finally KD7a with KT400A:
At7-MAX 2 comes up first but the differences are small! Less than 5%!
Next is the last 3d mark: 2005. Based on DirectX 9, this a very demanding test that requires a strong video card!
AT7 - KT333
Followed by AT7-MAX 2 - KT400
And lastly KD7A with KT400A:
3d mark benchmarks are over , now is the turn of PC MARK: 2002, 2004 and 2005.
As usual, AT7 with KT333
Second is AT7-MAX 2 with KT400
And finally KD7A with VIA KT400A
There is no mistake! I have run all the tests 4 times on each system and, yes, geforce 7800gs performs best on AT7 with KT333 chipset at AGP 4X! I think by now is obvious that AGP 4x is not limiting any video card to show its true potential!
Super PI 1M is next; usual order: first AT7, next AT7-MAX 2 and last KD7A:
There is no mistake here!! KD7A despite being the newest board and using what should be theoretically the best chipset is the slowest of the bunch! This is the effect of what I said in the beginning: MAX is not just a moniker, a publicity stunt done by Abit. AT7/AT7-MAX2 have 3 options in the BIOS that are having a cumulative effect, a positive one, on the results!
Those 3 options are:
1. enhance for benchmark
2. enhance DRAM performance
3. enhance AGP performance
The final synthetic test is CINE BENCH 2003. I choose this version specically because it is the oldest I could find and is perfectly suited for old systems with single core CPU.
Again AT7 is first, followed by AT7-MAX2 and last one is the system with KD7A: