![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Let me start off by saying that I know how memory frequency calculation works with Athlon 64 -- the formula RAM frequency = CPU speed / ( CPU multiplier / RAM divider ). The divisor in this formula is the next whole number up from the actual multiplier:divider ratio.
Of course, CPU-Z is handy for detecting the RAM speed and divisor for me, but my recent overclocking work has gotten me skeptical of its readings. I have my CPU at 3060 MHz (360 HTT x 8.5)right now, and in order to keep my system from crashing, I'm using the 3:5 divider from my DFI nF4 board. I chose this divider in order to keep the memory close to 200 MHz (DDR400) because my system has problems when the RAM gets close to 300 MHz, even though I have OCZ's limited run PC5000/DDR625 RAM, at stock timings and voltage.
Here's the weird part: When I boot Windows and run CPU-Z, my RAM is detected as running at 255.0 MHz (CPU/12). Note the '12' divisor. The speed should be 3060 / [ 8.5 / ( 3 / 5 ) ] = 3060 / 15 = 204.
When I try the 2:3 divider, CPU-Z gives me an odd divisor (detected as 13) that brings RAM speed to about 249 MHz. The next divider on the list (7:10) gives me a real 13 divisor for a speed of about 235 MHz. The other dividers seem to be correct: The highest ones make the system refuse to boot (since they would give me more than 300 MHz), and the lowest one, 1:2, makes CPU-Z detect 180 MHz RAM frequency. I'm most concerned about the 3:5 divider since it supposedly gives me the highest working RAM speed. I'd like to know if the ram really is at DDR510. I'm doing a Prime95 stability test right now, but when I'm finished I'll go to Memtest86+ for a second opinion on the RAM speed. Until then, does anyone here have any ideas as to how I could be getting such an anomalous frequency/divider?
Of course, CPU-Z is handy for detecting the RAM speed and divisor for me, but my recent overclocking work has gotten me skeptical of its readings. I have my CPU at 3060 MHz (360 HTT x 8.5)right now, and in order to keep my system from crashing, I'm using the 3:5 divider from my DFI nF4 board. I chose this divider in order to keep the memory close to 200 MHz (DDR400) because my system has problems when the RAM gets close to 300 MHz, even though I have OCZ's limited run PC5000/DDR625 RAM, at stock timings and voltage.
Here's the weird part: When I boot Windows and run CPU-Z, my RAM is detected as running at 255.0 MHz (CPU/12). Note the '12' divisor. The speed should be 3060 / [ 8.5 / ( 3 / 5 ) ] = 3060 / 15 = 204.
When I try the 2:3 divider, CPU-Z gives me an odd divisor (detected as 13) that brings RAM speed to about 249 MHz. The next divider on the list (7:10) gives me a real 13 divisor for a speed of about 235 MHz. The other dividers seem to be correct: The highest ones make the system refuse to boot (since they would give me more than 300 MHz), and the lowest one, 1:2, makes CPU-Z detect 180 MHz RAM frequency. I'm most concerned about the 3:5 divider since it supposedly gives me the highest working RAM speed. I'd like to know if the ram really is at DDR510. I'm doing a Prime95 stability test right now, but when I'm finished I'll go to Memtest86+ for a second opinion on the RAM speed. Until then, does anyone here have any ideas as to how I could be getting such an anomalous frequency/divider?