Kingston's HyperX Sets DDR3 World Speed Record

O.o

not to be negative, but is there a point to overclocking ram?

i mean on an outside of rendering applications that are ram heavy, is there even a need for faster memory?

keep in mind, i can see a need for faster cpu, faster gpu, faster hdd, faster ssd (not any more, not its price over speed they should focus on) but memory?

underclock your ram and let us know how it turns out

no bottlenecks, faster is always better. I want faster everything in my pc. i wish they sold RAM that fast but maybe later in life they will.

Thats not the same thing, other bottlenecks and limits of what is supported mean that anything past those levels doesnt do anything. You can get most any memory to go well beyond what you can use so goin for a record is mostly for fun, no practical use in a consumer pc. Not yet anyway. Lowering your clock tho will lower performance, upping it wont help unless you just have slower memory than what your system supports. And 3600 is way beyond that.

But can it play Crysis faster than with a Phenom II X6 ? I want to see the benchmark.

Wow so nice rig. How can i get one of these? May be when DDR 4 arrives starting at 2100mhz at starting speeds up to 4000mhz.

i would love to see some benchmarks. i know AMD apu love higher faster RAM speed.

Thats impressive ram overclocking. Pretty sweet, nice job Kingston!

So far, my A8 is excited.

Come on AMD, make HT faster and make better RAM!

Cheers!

Cool! Congrats...

Liquid nitrogen. Wish I had some.

AMD Fanboy overclocking a Yugo. LOL, why don't you overclock the i7 990X, that will blow your mind.

Looks like something changed in either the Processor or the CPUz program, the earlier screens when the world record was broken listed the Codename as Bulldozer and this one lists it as Zambezi. The rest is all the same though but it'd be interesting to know whether they made any changes in the processor.

The bottlenecks of performance in order are CPU -> RAM -> Disk. If your bottleneck is not your CPU for some 100% utilization computation, but instead CPU resources go unused because reading too / from memory is taking too long. RAM Frequency increases the rate the memory can be read, and how fast it can be read sequentially. So if you want to read out 500MB of ram to cpu and subtract 1 from every number in that chunk of RAM, if it isnt CPU bottlenecked, the ram speed is too low to load all that data fast enough. The CAS latency is how fast a random area of memory can be reached and read from, frequency is how fast data can be read from the entirety of the stick.

So more frequency means more bandwidth, lower CAS means faster random access. More frequency becomes more important as CPUs get faster because they can execute stuff faster, read and process instructions faster, etc.

So more bandwidth is important, but so is access time, which is why maximizing frequency while minimizing latency is important and a hard tradeoff.

Overclocking memory typically offers minimal benefits to performance compared to overclocking other components. Obviously, though, faster is always better, and even tiny boosts are welcome if they don't cost anything extra. A modest overclock that doesn't stress the RAM much is free and better than no overclock at all, even if its impact on performance is less than a 1% increase overall (referring to the RAM overclock, not counting the CPU overclock).

A setup like the one in this article is just for fun and bragging rights, though, and really has no purpose in a computer meant for real use. It is ridiculously expensive and would probably barely net you a 1-2% increase in overall system performance (referring to the RAM overclock, not counting the CPU overclock).

The fact that you can strap a top fuel jet engine to a Pinto and outrun a Porsche does not mean a Pinto is a good car.

Why is it 2CORE/2THREAD?

^I think turbo mode kicked in and disables two of the core to reduce heat, but not to sure though. Good question, I like to know why myself.

Hell yea that's why I choose Kingston products cause there epic~~~

Well done Matose and Lab501 Romania!

Suddenly, neckbeards. Neckbeards everywhere!

I bet teh ladies are lining up for these events.

Because AMD sort of cheated. Running 2 cores instead of all 8 means there is less heat to get rid of.

Eventhough Intel outperforms AMD at most, or probably ALL tests, why do these AMDs OC better? Any particular reason? And that must be why our Mr. Matose here used an AMD rig, and not an Intel one. Plus, these AMD boards OC better than Intel ones. Why is that???????

Clap hand for this

Depends very much on the architecture, not the gigahertz overclocking potential. You can take an old Athlon XP and overclock it to 2.5+ GHz, but an i7 at 1.6 GHz will mop the floor with it. Same if you overclock a Pentium 4 at 2.8 GHz and pin it against an underclocked Phenom II at 2 GHz; the Phenom II will come out on top (the Pentium 4 from back in the 2000's can roughly be compared against today's Intel Atom).

It's all about the architecture. It's not that AMD processors are better because they can reach over 9000 clocks per nanosecond.

Nice but again useless!

I'm not one to bother with overclocking RAM, but it seems to me there could be a larger impact on performance if you lowered latencies rather than raising MHz.

Personally, I'd be curious to see Tom's do a test of the impact on system speed/performance of changing these numbers. What gives the greatest real-world performance boost: lowering CL, lowering tRCD, lowering tRP, lowering CAS, or raising MHz? Help us decide if we should spring for faster RAM or lower-latency RAM, or if the money is better spent on a different component of the rig.

in order to increase speed, you need to sacrifice the latency.

No, but we are talking about the Bulldozer vs. Sandy or Ivy here, aren't we? My question is what is inside the AMDs that isn't inside the Intels. I'm specifically talking about the clocks. In addition, why do the RAMs OC better on AMDs? I mean, look at all the RAM speed records on CPU-Z. And the processor speed records. Why is AMD stable at way higher clocks than Intel? And why is RAM more stable on AMDs, even though Intel rigs are the winning combination? I mean, if Intels could boast more clocks, at least as good as the AMD counterpart(s) (or whatever is the nearest Apple to Apple), no one would even give two ***** about an AMD. Right?

lol thats clocked faster than my cpu is pc tech is moving fast