Results posted must have bandwidth tests

[dinos22]

im not opposing or caring or not caring about anything but rather just looking at this from what this section was used for over the years and that is frequency RAM overclocking. I dont really get into that sort of thing myself and prefer performance based OC but at the same time i like watching these guys put the effort into getting every last MHz. What this section has been about is that...every last MHz. Mhz hunting is contradictory to bandwidth hunting. This section is as much about RAM OC as IMC stregth on CPUs. This sort of stuff is not easy to achieve and is much more exciting to a lot of guys than looking at raw bandwidth screenshots which mean what....i donno i certainly wouldnt bother with everest unless i was troubleshooting and finding things out in initial stages of ram testing

It is the same deal with CPU MHz OC for example where you pick your strongest core to shoot for best clocks or you use crazy low QPI on clarkdale for example to be able to reach the highest frequency such as a recent cracker effort from duck on 655K....It's the same thing to me. You wouldnt ask duck-san to run vantage CPU test at that frequency or any other form of intense CPU testing to prove it not running in crippled more even though it is

[Afrokalle]

Well said Dinos

I prefer also perfomance based oc but i like to see that high numbers here too....

It is not easy to push memory to the last MHz and i think there is no need to show some crappy Everest Benchmark

CPU-Z is the official validation and needed to prove WR's.

If B2B (i never use) makes that big difference in max. mhz the guys should report (or show cpu tweaker) what B2B used but...

...it is similar to cpu mhz hunting: It is more easy to push a Core i7 920 to BCLK 250+ with low qpi speed but it will be a regular WR (even if the other guy is using "normal" qpi speed and can't get same mhz cause of this fact).

I like this section and i would love to see raw memory mhz power in future here too

[T_M]

For me, i felt this section was good to cover all bases of RAM overclocking and performance.
For instance the latest thread about where teh Hyper ICs might still be coming from, the death list of a particular IC type, and the max overclocking results of PCS do not really belong here anymore if we play by the new rules.

[Hondacity]

guys keep posting frozen rams.....

[damha]

Makes sense to me. Xtreme bandwidth = xtreme bandwidth. XOC section for super cooling/suicides sp32m whatever.

What doesn't make sense to me is why it's such a big deal. It's not like everest/sandra bandwidth tests take forever to run

[Kazi]

System specs are in sig.

[Ket]

Marc, it is sad to see people exploiting/clueless to B2B to fudge results.

Look at Mat's results of 3Ghz yet he stats bandwidth was bad...

Look at your results, you exploited B2B to the fullest without a single bandwidth test.

Lastly, this is the "Xtreme Bandwidth" section, not the "Xtreme B2B exploit" section.

QFT Its sad to see even now though some time after this thread went up people still posting saying X memory can do X speed, showing CPU-Z and the like, but no bandwidth results Such offenders that repeatedly do this should get a temp. ban IMO to learn them.

[Nikolasz]

Everest tools old and new makes diferent scores.
Is there any other tool to test bandwidth?

[fredeb]

Everest tools old and new makes diferent scores.
Is there any other tool to test bandwidth?

Aida bought Everest out , so one could use Aida64 Extreme edition . Si software Sandra have some mem bandwidth tests .

It's interesting to see comparative results eg. Lower speed / lower latency vs. Higher speed / higher latency ; dual channel vs triple channel vs quad channel memory . Would be cool if ppl posted uncore frequencies , ND Qpi Link Speeds .

Aida64 home

[railmeat]

:0 this thread went out the window I think.....looked great,topic header wise!

[alpha754293]

What do you recommend to use for benchmarking memory bandwidth in Linux (specifically SuSE Linux Enterprise Server (SLES) 12 SP1)?

Thanks.