DFI AK76-SN: DDR-SDRAM at a reasonable price

His publicity in CsH

Let's be clear: the principal thing with a motherboard is not to do thousands of tests, but to verify that it works correctly, and the AK76-SN expires with this requisite. In all the tests he has behaved of stable form and his yield has been the awaited one: more rapid than a badge of high scale with SDRAM, but neither something incredible, because the DDR-SDRAM is an evolution, not a revolution.

For some of the tests, we have been slightly cruel: we have used an ABIT KT7A (chipset ROUTE KT133A) with memory PC133 CAS2 and the last available BIOS, with all the yield options to the maximum (Interleave 4 way, etc.); the best of his class, against the almost such DFI which comes from factory.

Ah, the mike was an Athlon to 1 GHz with 266 MHz bus (he knows already, 133 "physical" MHz with DDR) and the memory DDR-SDRAM was the PC2100 of Crucial, of excelentísima value for money. As for the graphic card, it was a GeForce2 MX400 (200/183 MHz).

Here we see the big jump given from a not ancient at all chipset ROUTE KT133 with memory SDRAM PC133 and bus of the mike to 200 MHz, up to the DDR-SDRAM PC2100 and bus to 266 MHz of our DFI AK76-SN: more of 30 % additional of band, which is said soon. The Pentium 4 stays much superficially, but eye: in the real life the Intel mike does not make use of this theoretical advantage...

Sandra 2001 is a "synthetic" test; how is this band reflected in the real life? Since according to the test of simulation of games 3DMark2001 of MadOnion.com, we obtain 2 % of progress. An increase of small yield, but bear in mind that in this case the opponent is our system "SDRAM with steroids", with quite expensive motherboard, bus to 266MHz, memory CAS2, optimized BIOS and all the advances. I dress like that, it is not badly...

... And it is better if we change the memory PC133 CAS2 into memory PC133 CAS3 "generic": the difference grows up to 4 %, which is not great but yes it begins being significant.

• www.dfi.com