HR Forums

Full Version: Experimenting with Intel Extreme Tuning Utility
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
The summary: First row is default CAS11, second is CAS7...
[Image: attachment.php?aid=186]

RED is 7-9-8-18 timings.
GREEN is default 11-11-11-28 timing. Blue=Green in the first screenshot.
This was not very scientific. I did not reboot before the 7-9-8-18 run. Virus scanner was running
[Image: attachment.php?aid=178]
[Image: attachment.php?aid=179]
[Image: attachment.php?aid=180]
[Image: attachment.php?aid=181]
[Image: attachment.php?aid=182]

note: my SSD is temporarily MIA supporting other work, but I have not updated my sig as I expect it will find its way back to this system, hopefully soon.
[Image: attachment.php?aid=183][Image: attachment.php?aid=184]
The yellow text indicates all of the changes made for the tighter timings.
[Image: attachment.php?aid=185]
[Image: attachment.php?aid=187]

(10-08-2014, 02:30 PM)Rapajez Wrote: [ -> ]How does it play out in Benchmarks? Specifically gaming. Have you ran some before/after in Furmark, 3D Mark, etc?

I have not had time to do much bench marking. The above was just a quick run for comparison. note his was done with an fairly old 1600x1200 monitor (dell ultrasharp 2007FP) so I'm not sure the age, and potentially slow data interface affects the numbers. It was the same monitor for both tests, and connected using DVI-D single link.

I might have some time this weekend to test more.

I'd like to re-run these with a 1920x1080 monitor so the benchmarks are more easily compared with everyone else out there. That might happen this weekend too, but it may not.

Anyway, I think this turned out well, and in the end, maybe I did get something for nothing Smile
Cool. I'd be interested to see how that relates to gaming. I might play around with that tool this weekend. Hopefully I don't blow anything up.
Honestly, I'm not sure how to measure FPS to compare a before and after while playing a game. I downloaded fraps, and I can see numbers, but I'm not sure how to do something repeatable and measure it.

Games seemed a little smoother but it is hard to quantify. Maybe I'll try an A-B test with my son playing. He can push games a little harder than I can, and then I can watch the visuals while he plays and see if I see a difference.

as far as blowing things up, I am a little concerned about the on-chip items that are getting slightly over volted and if this bad or not. I might start dialing the system voltage offset 10 mVolts at a time and see if I can get closer to +100 mV.
You can run Heaven Benchmark before/after. 3DMark is also really popular, but usually criticized for not correlated to real-world gaming. I think it's good enough to quickly test whether a change made things better or worse, and it's definitely the prettiest.

Some games have benchmarks built in, that will give you average/minimum frames. Usually hiding in the settings menu.
Heaven benchmark at 1600x1200 Quality == Ultra, Tessellation = Extreme, AA = 8x was nearly identical. GPU was temp limited at 80C and the CPU was barely working.
ive recommended XTU on here for beginning overclockers and its overclocking algorithms are really good. overclocking memory alone your probably not going to see a big difference in real world gaming even at higher speeds even with haswell at 1080p, unless your running at very low resolution. i think SC2 and Thief is about the only games ive seen that benefits from higher memory clocks. although benchmarking and video editing(compression, encoding, transcoding, etc.) also sees a marked improvement.

nice work on the timings though.
Pages: 1 2