Ook bij AnandTech, een review van de uitgestelde i820 chipset van Intel. Hier heb je vast een gedeelte, alle info vind je in de review.
One thing is for sure, the i820 platform won’t be cheap and the cost is a very important factor to keep in mind. If a Coppermine + 820 platform ends up costing more than an Athlon system then Intel better hope that their Coppermine outperforms the Athlon by a good percentage. It looks like AMD may have won this battle, although a Coppermine 700 with 64KB L1 and 256KB integrated L2 cache could give the Athlon a run for its money.
So what about the weakness of the i820 chipset? RDRAM?
The theory behind RDRAM is that the only time faster memory transfer rates are necessary is when you are running an application that can’t fit entirely within a chip’s cache. In those applications, the theory holds true as we’ve proved in the benchmarks from the Photoshop and 3D Studio MAX scores in particular. When the Coppermine comes around with its 256KB of on-die L2 cache, it should be interesting to see how the matchup of the i820 and the Coppermine turns out, in terms of performance.
For overclocking, with the inclusion of the ˝ AGP clock divider, a number of overclocking options once confined to our dreams have now been resurrected. Hitting the 150MHz FSB frequency shouldn’t be that big of a problem, but the main thing to remember is that the limiting factor here will again be the AGP frequency, so reaching the 160MHz FSB frequency will probably not be too successful.