I just have a couple questions. I've been reading as much as I can on overclocking CPU, GPU, and RAM. I've already got my CPU and GPU at decent speeds that I'm happy with, and I've started to tweak the RAM.
First thing is, in all the literature I've found they've been outdated and talking about DDR2 RAM (even all the memory OC links in the sticky here). In all their examples they talk about mobos with FSB of 300, 400, even 500. So, have FSB speeds been seriously gimped these days or what? Is there a reason? I'm running an i5 750 at 3.6ghz (200bclk x 18), on a MSI P55GD65 mobo, The base bclk is only 133.
Next question is regarding RAM timings. Everything I've read talks strictly about the first few timings but not all the other ones I can change in my BIOS. For example, my memory is Patriot Viper II Sector 5, manufacturer recommended 8-8-8-24 2T 100 @ 1.65V. This is what I initially set it to, although I found I could run it at 8-8-7-23 2T 80 @1.65V, and according to the XMP profile (viewed in CPUZ) I could run it at 7-7-7-21 2T 100 @ 1.7V. So I get that is faster because it's less latency, but then there's all this stuff (copied from wikipedia):
tCL: CAS Latency
tCR: Command Rate
tPTP: Precharge To Precharge Delay
tRAS: RAS Active Time
tRCD: RAS To CAS Delay
tREF: Refresh Period
tRFC: Row Refresh Cycle Time
tRP: RAS Precharge
tRRD: RAS To RAS Delay
tRTP: Read To Precharge Delay
tRTR: Read To Read Delay
tRTW: Read To Write Delay
tWR: Write Recovery Time
tWTP: Write To Precharge Delay
tWTR: Write To Read Delay
tWTW: Write To Write Delay
I found that I actually had to change a couple of the settings lower down on this list to exactly match the recommended specs, but I have no idea how those work or affect the mem speeds. Is there much point in changing them? Does changing from 2T to 1T greatly increase speed or is it impossible if the RAM's not built for it?
Any help in understanding this stuff would be appreciated.
You may also find the CPUTweaker utility to be handy. It's similar to the MemSet program you'll see referenced (and from the same author) but for chips with integrated memory controllers: http://www.tweakers.fr/cputweaker.html
Ok thanks. I'll check it out. Just a note, I decided to give it a try and now I'm smoothly running a Prime95 blend test with 8-8-7-23-1T 80 @1.65V. I thought going from 2T to 1T would crash it but nope! It's fine... hopefully this read will explain why lol.
I ran a few benchmarking tests on the memory using different settings. I found the fastest was 1T 8-8-8-24 100, compared to 2T 8-8-8-24 100, 1T 8-8-7-23 100, 1T 8-8-8-24 80, 1T 8-8-7-23 80. I found that rather interesting as in theory the last one I listed should have been fastest. The benchmark I used was just a trial version of, I think it was called, Performance Test 7. Fastest results scored about 1445 (composite score) and the next was about 1430, all the other timings were under that. I did 3 tests per set. I also tried setting the RAM to the higher timings as per XMP, 2T 7-7-7-21 100 1.7V but it was stuck in a POST loop. Probably needed more voltage, but meh I'm happy enough as it is.