3

I tried to read about this, like what is the actual difference, but I could not find any benchmarks or helpful posts. Most of them are simple placebo effects, like "oh my god the 1600mhz module made my office load faster". Sure, bro.

I'm curious about the actual difference.
All of my modules are XMP compatible, but for some reason the new 8GB pack does not like the XMP profile at all. For now, they are running just fine on 1066mhz. Triple channel, that is.

Is there really a performance loss like this?
Like if I compile a lot (as a programmer, I often max out the CPU full), or for gaming, or anything?
(I'm not an avid gamer, but games are so taxing that my PC struggled to get 60FPS in most modern games anyway. With the 1600Mhz XMP speed, that is.)

Best would be, to run my own tests. But again, I cannot really do that. If I take out the new modules, it's not triple-channel anymore. If I put them in and switch on XMP, BSOD.

Biggest anomaly is that memtest works and passes 100% each and every time.
But once I fire up prime95 with "Blend" profile, I get a BSOD pretty soon. (with XMP).
First I did not even stress the modules, just started working. Then I met the BSOD pretty soon.

Specs
CPU: i7 950
RAM: 4x Kingston 2048MB 1600Mhz XMP + 2x Kingston 4096MB 1600Mhz XMP
MOBO: Gigabyte X58A-UD3R

Update - 2015.12.23
I have not noticed any performance difference. There is none.

Apache
  • 16,299

1 Answers1

1

You will generally not notice a distinct difference from faster ram. Faster ram does mean every operation is a little faster, but it does not directly address a bottleneck except for a small class of applications that maintain millions of small objects which are each accessed every cycle. The only app I can think of that meets this definition is Dwarf Fortress.

Frank Thomas
  • 37,476