HTT & Mem. speed for same CPU speed

[proxopspete]

Ok...
I got two setups that I can get my board to run at...

245x10 w/ 166 mem. divider (204 MHz mem)

280x9 w/ 133 mem. divider (187 MHz mem) <-- tried this after reading Zebo's sticky

They are almost identical in speed (2.45 vs 2.52) but I was wondering which has more impact in EVERYDAY performance (esp. in games). Anyone know of a thread here or experience with that? Thanks...

 

[monster64]

Yes, the higher htt will give you about an avg of 5 more fps in games. My venice 3500+ can run at 250x11 1:1 on the memory, or I can run 305x9 with a 333 divider (may be a different divider for you) and luckily it gives me a the exact same memory speed (250 mhz) as with 250x11. As for you, the higher htt is better, because if anything lower the ram timings, since you're running under ddr400, and your performance will be about the same ram wise, but better cpu wise. Remember, CPU mhz>Ram mhz.

 

[proxopspete]

Well, I found this article from the 939 memory matrix sticky.. It's an article from Neoseeker and decent one that compares things from HTT to memory...

In the end, it seems the clock speed is the biggest factor and then HTT and then memory... Hmm...

 

[coomar]

the 70 mhz more would probably be faster, run superpi and compare the 2 settings

 

[lifeguard1999]

Speed is life. coomar is right. The 2.52 GHz will be faster.

 

[proxopspete]

Yeah... speed is faster. 😉
However, CS: Source wasn't stable at that setting of 280x9. 🙁
So I am back at 2.45. Oh well...

 

[theMan]

why dont you try 278x9? or lower? or, even go down to 8.5 you should try to find out what the highest stable htt your board can put out. then find how fast your cpu will go. then use gogar's calculator: http://math.gogar.com/athlon64.cgi