Westmere-EP - X5660 Full Review
X58 DDR3- 1600Mhz vs 1900Mhz vs 2000Mhz Performance % Comparisons
I’ve dedicated a large portion of my weekend to these benches so please bear with me. I've done as much as I could. Also big thanks to kpforce1 for the help with these benchmarks
It’s been proven that faster RAM [1600+] can lead to better scores in several programs. This is more likely on newer architectures [Post-X58]. It’s possible on the X58 platform, but you’ll nearly destroy the CPU, MB or the RAM attempting to emulate the scores you see on X79. Or you could potentially do some damage and cause some major degrading. I tried to be as conservative as I could with the voltages and focused around 1900Mhz - 2000Mhz. I probably could have and probable should have used more voltage; I’m not risking my build for faster RAM frequency. That's to big of a risk for me. My board is only rated at 1866Mhz.
However, I’ve always read different views on the benefits for gaming & CPU benchmarks. So I’ve taken many hours to compare RAM frequencies. We all know that the GPU and CPU matters, but how much does RAM matter on the X58 platform. The fastest frequency I could achieve was 2200Mhz. The problem is that it required a lot of voltage to POST and run. The other problem is that these voltages affect the CPU and the RAM the most. Therefore I’m running only 1600Mhz – 2000Mhz. I’m focusing mostly on the gaming side. Remember that my board is only rated for 1866Mhz so hopefully this barrier didn’t influence my scores that much. I also attempted to show as many scores for drivers “331.82” & “337.50” with 1600Mhz, 1900Mhz and 2000Mhz in the chart. Below I'm going to compare the core speed+RAM frequency performance percentage.
Here is the list of the benchmarks that I’ve ran:
Here are the breakdowns from my 1600Mhz, 1900Mhz and 2000Mhz scores [all graphic and physx drivers.] I will be taking the highest scores and comparing them from the results [increase or decrease%].
3DMark 11 – Performance % Results:
4Ghz-1900Mhz = +0.72%
4Ghz - 1600Mhz = 0.0%'
4.6Ghz - 2000Mhz =+ 1.9%
4.6Ghz - 1600Mhz = 0.0%
4.8Ghz - 2000Mhz = +0.93%
4.8Ghz - 1600Mhz = 0.0%
--
3DMark 11 - Extreme % Results:
4Ghz-1900Mhz = +0.59%
4Ghz - 1600Mhz = 0.0%
4.8Ghz-1900Mhz = +0.29%
4.8Ghz - 1600Mhz = 0.0%
--
3D Mark Fire Strike – Performance:
4Ghz-1900Mhz = +0.29%
4Ghz - 1600Mhz = 0.0%
4.6Ghz - 2000Mhz = +0.94%
4.6Ghz- 1900Mhz = +0.62%
4.6Ghz - 1600Mhz = 0.0%
4.8Ghz - 2000Mhz = +0.18%
4.8Ghz - 1600Mhz = 0.0%
--
3D Mark Ice Storm - Performance:
4.6Ghz - 1600Mhz = +1.07%
4.6Ghz - 2000Mhz = 0.0%
4.8Ghz - 2000Mhz = +3.30%
4.8Ghz -1600Mhz = 0.0%
--
Tomb Raider 100% Maxed 1080p - Avg. FPS
4.8Ghz - 2000Mhz = +1.88%
4.8Ghz - 1600Mhz = 0.0%
--
Tomb Raider 100% Maxed 1600p - Avg. FPS
4.8Ghz - 2000Mhz = +0.88%
4.8Ghz - 1600Mhz = 0.0%
--
Battlefield 4 100% Maxed 1600p – Avg. FPS
4.6Ghz - 1600Mhz = +7.5%
4.8Ghz - 2000Mhz = 0.0%
--
Crysis 3
4.6Ghz - 1600Mhz = +5.87%
4.8Ghz - 2000Mhz = 0.0%
Here are some high results I didn't think I'd ever get
So are RAM modules running Frequencies higher than 1600Mhz worth it?
The answer is yes and no depending on a few things. For the X58 platform 1600Mhz – 1866Mhz should be more than enough and has been for several years now. Now if you see 1600Mhz & 2000Mhz kit around the same price and your motherboard supports the frequency; then it might not be a bad decision to go with the higher frequency. If the 2000Mhz support a CAS of 7 or 8, I’d go with the 2000Mhz modules instead of the 1600Mhz personally. RAM timings are another thing to consider. Personally I would go with 1600Mhz RAM due to the fact that you can tighten the timings much better and usually get a lower CAS.
As far as overclocking goes 1600Mhz gives me a bit more flexibility when overclocking. My board [or probably the X58 platform] is limiting me what I can achieve with 2000Mhz RAM. I believe I could get more out of 2000Mhz RAM, but the potential consequences aren’t worth it. This is my only build and I don’t know what I’d do without my 1st Generation Beast. I actually broke the some of my old records so I’m glad I ran these test again with the latest drivers. The performance increases are pretty minor across the board when going from 1600Mhz to 2000Mhz. However some of the numbers look much better with 2000Mhz. I actually hit 16767 in 3D Mark 11, not quite 17K that I'd like to hit [but that's close enough for me:thumb:]. I also finally broke 200,000 in 3Dmark Ice Storm with 1600Mhz RAM, 2000Mhz just wasn't doing it. Real-Time Benchmarks™ shows that there isn't a big difference I'm able to tighten the times on 1600Mhz and it shows when compared to 1900Mhz - 2000Mhz RAM in some of the benchmarks. Thanks for reading. I'll update as I continue to benchmark.