[title]AMD Athlon 64 & Sempron Comparison Pt. II[/title]
After posting my AMD Athlon64 vs. Sempron comparison last week I was still puzzled to why the overclocked Sempron 3100+ was able to easily out perform the Athlon64 3200+ in all the synthetic tests but not the real world tests. A large number of readers also thought this was quite strange, so I spent most of this week trying to find out why. The problem seemed to be the Gigabyte K8NSNXP motherboard which featured an out-dated BIOS version. While the board did detect the Sempron 3100+ processor, the board's 3D performance when using the Sempron was for some unknown reason quite poor.The previous results showed a 300MB/s performance discrepancy in SiSoft2005 between the two processors which favored the Athlon64. Both the tested processors utilize the same 200MHz FSB with the only real difference being the much smaller L2 Cache of the Sempron. Of course the Sempron processor does not support 64-bit instructions though this is irrelevant for such testing. Bottom-line is I cannot explain why the Sempron memory bandwidth is 300MB/s slower in SiSoft2005. The only possible limiting factor could be the Gigabyte K8NSNXP motherboard, unfortunately I do not have another 754-pin board handy at the moment.
Nevertheless, when using the Radeon X800 Pro the Sempron delivered gaming performance that were much the same as the Athlon64. Given the Athlon64 3200+ is clocked 200MHz higher, this is quite impressive. The original article titled "Athon64 vs. Sempron" was not about discovering which was the fastest processor. Clearly the Athlon64 is going to win this competition, given I used the Athlon64 3200+ which is clocked at 2.0GHz to compare against the Sempron which is clocked 1.8GHz.
This entire article came about because a number of readers had asked what would be a better purchase 搕he Sempron 3100+ or the Athlon64 3200+? These readers wanted to know if the $50 US saving made the Sempron 3100+ a better option. I admit the title was quite misleading and I should have explained this more in the article. I also admit a GeForce 6800 Ultra was not the best graphics card for the article. For these reasons I have re-published the article using different hardware. This second article will also include an Athlon64 2800+ system which of course operates 1.8GHz. This article will now not only offer a fair comparison GHz wise but will also compare two processors that share a similar AMD rating (3100+/3200+).
