I have a question about the performance difference between an Athlon 64 and a similarly-numbered Athlon 64 X2.

For instance, I have an Athlon 64 3700+. Its clock speed is 2.4GHz. An Athlon 64 X2 3800+, the lowest-end of the X2 series, is clocked at 2.0GHz.

The numbers (3700+ and 3800+) suggest that the X2 3800+ is faster than the 3700+, but the clock speeds don't seem to agree. I realize that the X2 CPU is dual-core, but is that enough to make a 2.0GHz chip faster than a 2.4GHz one on all occasions?

Any feedback would be appreciated.

TIA,
Dai

Hello!

The X2 3800+ is not one 2 Ghz chip, it is basically two 2 Ghz cores (on the same chip)!

No, the X2 will not be faster in all occasions. I bought the x2 3800 chip a few weeks ago and I am happy with the performance, but depending on what you do a single core chip with the same spped rating can also be faster. The x2 is basically 2 2ghz chips working together, and if the program/OS/whatever cannot utilize symmetrical multiprosessors (SMP), then you are only using 1 2Ghz core, and in that case a single 2,4Ghz core is faster.

With 2 cores the load spreads between the cores, and you can work on several things at once.

An example (unfortunately I cannot remember where I read this):
A test on X2 3800+ where the performance in a game (one of the measurements) and the performance for encoding mp3-files (another measurement) was quite OK in relation to the other processors in the test. When they then tried doing both these things at once the performance did not suffer very much because each task was only use one core.

Linux has good support for SMP, so you can use the benefits 2 cores offer.

Razze

I would add it is probably a no-brainer to overclock the X2 to at least the 2400 speed probably 2500 without taking the HT bus and ram above its rated speeds that is if the BIOS on your motherboard supports this.

Thanks guys. Yes, I realize that dual-core means two CPU cores in one chip. Some applications take advantage of SMP while others don't; which led me to ask if the number rating (3800+ vs 3700+) can be misleading. One can easily tell that a dual-processor P4 2.6GHz computer may not always be faster than a single P4 3.0GHz box because you compare raw clock speeds, but the AMD naming convention seems to suggest that an X2 3800+ will always be faster than a 3700+. I just wanted to make that clear - which you guys certainly did for me.

Which core do you have? My San Diego is only stock at 2.2GHz, 2.4Ghz is the 4000+...

Ok that P4 example is really not a good one though they used the Hyper-Threading which was just software on the chip to emulate a second processor plus they had to share the cache X2 as you know have two cores separate caches. The rating system AMD uses is kind of strange sometimes it is supposed to be a speed relative to IIRC an Athlon (XP?) processor when I look at them I just think of the actual mhz they run at and the rated speed on the chip is just a flat out lie to scheme money from people. I am convinced that any of the 64s will run at a higher speed than they lock them at, especially the Venice and later cores my 3000+ has a 1.8 rated runs perfectly fine at 2.4ghz would probably go higher if I had higher speed ram.

Mine's CrawHammer (for socket 754). Looks like San Diego is more efficient if it achieves the same performance with a lower clock speed.