AMDBoard Review of X2 3800+

[fbrdphreak]

http://www.amdboard.com/athlon_64_3800_review_2.html

Cachemem memory and cache access latencies with the memory bus running synchronously at 400 MHz (solid blocks) or else at 433 MHz (transparent blocks) -- lower is better. All latency settings were identical. One issue we ran into originally was that upon increasing the memory frequency to 433, the CPU HT frequency increased (without any manual interference from 200 to 216 MHz) meaning that the CPU was running faster, apparently using the memory clock (433 MHz) as reference. We manually adjusted the clock back to 200 MHz and checked the performance using entirely CPU-centric benchmarks such as Caligari TrueSpace as well as by comparing the cache access latencies in cachemem as reference. The graph clearly shows a significant performance degradation resulting from running the memory bus faster - in asynchronous mode.

Well, first off the writing in this article is just BAD, so damn confusing half the time. That may be part of my confusion.

But is this guy saying that running the mem clock @ 216 with same CPU clock actually yields lower performance @ same timings vs mem clock @ 200? "Asynchronous mode?" I thought since there wasn't an FSB anymore that memory wouldn't have an advantage being "synchoronous" with HTT.....

 

[slash196]

I haven't heard this issue voiced by anyone else. It could be a measurement error.

 

[fbrdphreak]

The article is so poorly written I personally can't put any faith in those results. Poor writing doesn't mean their methods are off, but it sure gives me reason to doubt their competency.

For a website called "AMDBoard," it is quite sad to see such poor writing.

 

[DAPUNISHER]

I think I get what they are trying to impart, poorly articulated/written as it is. He seems to be exploring this

* the revised memory controller allows to run the memory in pseudosynchronous DDR433 / 466 mode

So he efforted to set the ram@that speed while keeping the CPU@the default 200mhz. The Asus board used was automatically increasing it in relation to the ram speed so he had to manually tweak it.

The point of his testing seems to be that without the on-dice memory controller coming along for the ride, raising the ram frequency alone actually had a negative impact on performance. As you are aware, Without uber-ram, most usually use a divider where the memory controller gets pumped up while reducing the ram frequency in asynch mode. This is the opposite, increasing ram frequency while leaving the mem controller@default.

The only point I can see in his exploring this, is to evaluate what would happen if OEMs decide to ship systems with those specs in order to promote it having faster DDR? Marketing decision? Maybe completely irrelevant? I don't know, it is confusing, why he did this otherwise

 

[fbrdphreak]

That makes a little more sense, thanks bud :beer:

 

[DAPUNISHER]

Originally posted by: fbrdphreak
That makes a little more sense, thanks bud :beer:

I had quite a few of those last night, subsequently my mental processes are impaired Hopefully another member will have more illuminating remarks to be made. You could e-mail the reviewer and quiz him on it. I'm certain you have had a few readers do so to you by now?

 

[fbrdphreak]

Originally posted by: DAPUNISHER
You could e-mail the reviewer and quiz him on it. I'm certain you have had a few readers do so to you by now?

Oh yes! We had quite a flame war going on with the Turion vs Pentium M article Luckily I didn't write that one!

 

[DAPUNISHER]

Originally posted by: fbrdphreak

Originally posted by: DAPUNISHER
You could e-mail the reviewer and quiz him on it. I'm certain you have had a few readers do so to you by now?

Oh yes! We had quite a flame war going on with the Turion vs Pentium M article Luckily I didn't write that one!

LOL. I haven't been keeping up with the mobile scene and I based a few comments on that review, which I should have read stone cold sober to begin with Yep, been getting in the bag too much of late because I celebrate the holidays rather exuberantly I got owned for it :laugh:

Accord99 near the bottom of the page and my evident lack of reading comprehension pointed out by dexvx. Text