Which Is Faster 825 tight timeing or 1100 with default

[john19055]

I was just seeing if it really made that big of a difference with that higher price ram,so I ran 3Dmark06 at 3.7g at 1100 at 5-5-5-15-2T and 4-3-3-9-2T
at 825 I lost 1 point with 825mhz at 4-3-3-9-2T compared to 1100 at 5-5-5-15 2T ,but there were other things to at 825 I could ru 1:1 ratio and at 1100 I think it was 5:4 ratio,So IMO I think you should just save your money and buy some good PC2 6400C4
I was using PC2 8500C5D dominator with a fan.

[OrangeBlast]

so your saying smaller mhz on the ram and tighter timings and keeping a 1:1 ratio is best?

[john19055]

In the test I have done yes.

[Dazz]

You really need to run 3D Mark 2001 too see the diffrence as it's CPU/system memory limited.

[john19055]

I ran all the test and there is not much of a difference.

[ViN86]

i would guess 1100 with default is faster, saying how c2d isnt as timing dependent as the athlon64 is.

[HeavyH20]

Well, just for comparison, I ran some tests, as well.

FSB 1066 with DDR2 at 853, 4-4-4-12, 1T, linked, 5:4 - 15627

FSB 1066 with DDR2 at 1066, 5-5-5-15, 2T, linked, 1:1 - 15523

FSB 1066 with DDR2 at 900, 5-5-5-15, 1T, unlinked - 15611

So, looks like high DDR2 speeds are not much of a gain over lower and tigher.

[Roadhog]

Originally Posted by HeavyH20

Well, just for comparison, I ran some tests, as well.

FSB 1066 with DDR2 at 853, 4-4-4-12, 1T, linked, 5:4 - 15627

FSB 1066 with DDR2 at 1066, 5-5-5-15, 2T, linked, 1:1 - 15523

FSB 1066 with DDR2 at 900, 5-5-5-15, 1T, unlinked - 15611

So, looks like high DDR2 speeds are not much of a gain over lower and tigher.

This is what I get in everest.

1000Mhz 5-5-5-15 @ 2.7Ghz
Read: 7715 MB/s
Write: 5319 MB/s
Copy: 6028 MB/s
Latency: 58.5 ns

889Mhz: 5-5-5-15
Read: 6843 MB/s
Write: 4720 MB/s
Copy: 5341 MB/s
Latency: 65.9 ns

800Mhz: 4-4-4-12
Read: 6641 MB/s
Write: 4718 MB/s
Copy: 5344 MB/s
Latency: 69.2 ns

533Mhz: 4-4-4-12
Read: 5479 MB/s
Write: 4731 MB/s
Copy: 4845 MB/s
Latency: 80.7 ns

[HeavyH20]

You never mention whether you are running 1T or 2T, but since it is an Intel chipset board, it is 2T. Also, since your better run was at elevated FSB rates on the CPU, it cannot be considered a valid comparison. Run the RAM at 1000 Mhz with the FSB on the CPU at 1066 and you will see no gain.

You can make your own conclusions, but I think this shows that there is little to no gain with 1000 plus MHz RAM at stock FSB. Now, if you increase the FSB of the CPU, then the extra bandwidth will make a difference.

What I run, 4-4-4-12, 1T, 885 MHz.




Here is 4-4-4-12, 1T at 853 MHz.



And, 4-4-4-12, 2T at 853 MHz.



Here is 5-5-5-15, 2T at 1000 MHz.



And, as you did on your tests, which is not a fair comparison of the relative timings, here is 4-4-4-12, 1T at 800 with the CPU using 1200 FSB.