Thank you! I have been hollering about this since the inception of the platform - and, I have pretty much given up trying to help with any i5/i7 issues, as nobody wants to hear,
or admit, that they've been 'hosed' by the memory manufacturers...
Fact 1 - i5/i7 supports
three memory speeds, 800, 1066, and, on some, 1333 - that's
IT!
Fact 2 - memory speed does
not 'scale' to better performance on these platforms - the
only increase in
anything that you'll
ever likely see is in '
synthetic' memory benchmarks - and that's
WHY they're
called SYNTHETIC!
Fact 3 - decreases in latency DO pay off in the real world; you are never, ever, ever doing large contiguous reads from/writes to memory; the vast majority of memory accesses are for a page to cache, and, pretty much every time, you are 'waiting on' one or the other of the latency periods to pass...
Fact 4 - because of the memory manufacturer's shenanigans, it is nearly impossible for the majority of potential purchasers to compare latencies at varying clocks, as most are either 'innumerate', or 'math-phobic' (q.v.!)!!
Fact 5 - this idiocy is starting to 'leak down' into AMD and 775 buyers, because they are purchasing RAM sold purporting 'raw speed' (and designed for the 1156/1366 platforms) that
will not likely work for them
at all, unless they are astute enough to figure out actual workable timings...
Fact 6 - the latencies are physical periods of time that, in reality, have nothing to do with memory clock counts - they are in nanoseconds and picoseconds, and these do not change for different clock speeds - the BIOS simply
'counts' them in clocks; if a column address strobe takes eight nanoseconds, it's nine counts at 1333, or seven at 1066, but it's still eight nanoseconds!
from an earlier post:
... there appears to be no real advantage to running the memory faster:
http://www.tomshardware.com/reviews/memory-scaling-i7,2325-5.html
it just:
uses more power
makes your memory hotter and more 'stressed'
makes your CPU hotter and more 'stressed'
makes it harder to achieve long-term stability...
If you are interested in the Intel docs, I've zipped 'em up here:
http://www.mediafire.com/?yzj5ggmyt4g
The only thing that appears to matter to these platforms is
latency - and you
pay for that, buying the RAM; here's a little tool that I've been working on:
It 'looks at' the listed latency at whatever claimed speed, converts it to 'likely latency' at 1066 in the 'adjusted' columns, and then, in the LV (latency value) column, 'weights' the average latency for 'bang for the buck'; notice that the entire 'spread' of average latency is 4.2 counts, and the median value, 8.1 is already reached by the fourth item, the $140 F3-12800CL8T-6GBRM - at that point you are paying $15 more (12%) to get to half the available range, but to get to the other half, the price differential rises to more than double the cost of the base entry (a whopping 108% increase!)
If you'd like the Excel file itself, to 'play' with, I've put the current copy here:
http://www.mediafire.com/?yizmaza3kaz
note: the little thingie in the lower right corner is just for my convenience in 'averaging' NewEgg review listings; I don't place too much credence in them, as the vast majority of review posters seem to be dumber than driveway gravel...
This stuff is months old, so the shown prices are likely no langer valid, but the idea remains the same: pay for low latency, which will do you some good; or, pay for high clock speed, which will do you
no good, but
will fatten the coffers of the memory manufacturers -
precisely as they intend!