Originally posted by: davidos
Yeah, I'll probably get the Epox KT333 board or the MSI basic KT333.
I keep getting IRQ_NOT_LESS_THAN_EQUAL errors with my Shuttle AK31 v3.1. Is that a motherboard thing or a Windows XP issue? and are there any solutions?
Highly off-topic, but after reading the rest of this thread, well... Anyway, someone else might have this problem.
I had that issue on my 8K3A+. Took me half an hour of rebooting, playing MOH (Games are great stress testers.
), and changing settings to figure it out. At first, I thought it was the CPU, since I'd underclocked and undervolted it due to bad A/C in the room. Then, I changed memory settings. The one thing I changed that seems to have cured the probelm was the queue depth. 8k3a+ has a setting for 8, which seemed high at the time, but I had borrowed some good RAM (Corsair PC3000 C2) and thought it would work. Dropping that to 4 solved the problem.
Steve: You're a stupid fool.
Gary: You're an arrogant ass.
and me: I'm a natural-born tightrope walker. (see my signature)
Price has a factor in quality. Higher-quality = higher price. However, this only applies to the same process. Different processes have different price/quality trade-offs. You would think the Intel vs AMD wars would've taught us something.
I am inclined to believe the wives tales and surrealist legends about 1600+ AGOIA. My 1700+ AGOIA will post at 162 default voltage and a HSF that barely kept my previous chip (AXIA 1.2) barely below 60C under load. It won't post a second time, though. Gee, wonder why.
I could always use a new HSF, but since this room doesn't have good A/C (barely keeps one guy comfy during a heat wave) I see no point. Besides, it's now underclocked to 1100MHz 1.4V (Thank you, Epox) and does about 39C full load with the same heat sink with 28-30C ambient. MBM picks it up as 92C; guessing it's the VCore.
Overclocking is an exercise in patience and statistics. The majority of each stepping will reach the same ceiling. i.e. 1700+ MHz for AGOIA, something like 1450+MHz for the first steppings, maybe 1200+MHz for the engineering samples. Depending on your multiplier, your FSB will be different. The really lucky chips of each stepping will, of course, hit higher speeds. Due to their lower probability, supply is low, thus the higher prices for the upper end. And then, there's the really unlucky chips that fail and the not-so-unlucky ones that can barely run at the cutoff. You do hear of the latter ones occasionaly when you see a post by some poor guy who's CPU crashes after bumping up the FSB 2MHz.
It's no secret that AMD and Intel survive by selling cheap CPUS since the high revenue chips that cost an arm and a leg don't sell as well. That's why occasionally, chips that can run faster get binned lower. Much better to sell a processor for 15% profit in two weeks than to have it take up storage space for two months to get the same profit margin due to falling prices.
So, in conclusion, my undervalued opinion is that this WAS a good deal. Up until the thread crapping and Newegg to raising prices again. My supporting statements are above, and this circus act is over. *dismount and bow*
Facebook
Twitter