Thuban,MaxxMem and Mem @2126Mhz

[demonkevy666]

I'm going to say this now.

not enough core speed at all.

4000mhz NB needs 5.8ghz core speed.

at least that's what I've concluded.

[chew*]

Other way around on the Asus mb (atleast with my mems/subtimings)
Noticed the bench a bit sensitive about right HT L/NB F combo, that's why i suggest ed you to up your HT L a bit

Problem with that is I'm tuning for ln2.....maintaining a high HT link on ln2 at high clocks usually you are mercy of cpu.....some can do it some can't, tune with it out of equasion and your better off.....if it likes it cool, if not at least you didn't waste time pretuning.

[El Gappo]

That is one hell of a crazy bootable northbridge for phase :O

[chew*]

That is one hell of a crazy bootable northbridge for phase :O

Especially considering something that you always seem to forget, All phase units are different, his is actually very very old as is mine so we don't get half as cold as other units......

If we see negative temps under load at cpu, we are lucky.....

[Dumo]

Especially considering something that you always seem to forget, All phase units are different, his is actually very very old as is mine so we don't get half as cold as other units......

If we see negative temps under load at cpu, we are lucky.....

QFT

Nice runs Mr. Icee

[pcnazz]

Wow look how off my mem scores are . little better Randy fixing the nb bottleneck , these runs are on a stock OCZ Cyro-Z gents , I know laugh at me , lol . I have a 300w on standby I just got , need some insulation before I set it up .

[MrIcee]

This is my last stab at the DDR3 class of the June MaxMem competition. I went back to lower speed with tighter timings and subtimings, with a bump of 200mhz on the NB. Don't even ask how I got the NB that high, as I consider it chinese phase torture and would not recommend the technique to anyone who values their warranty Needless to say, the changes were good for another 17 points. I guess not to bad for SS Phase.


Best regards,

Randi

[ViViD]

Wow That's Amazing Work
37ns :O

[MrIcee]

Welp, I just proved I was full of shiit again I couldn't leave it alone until I parked my arse behind friend and Team mate chew*. I made one last stab
in the DDR3 class, this time using the alternate strategy as illustrated earlier in the thread of high memory Mhz,primary and secondary timings
individually adjusted as tight as the memory would allow, kamikaze NB frequency and ended up with the following results. Great friend
Victor "DDTUNG" would be elated at this, his "less is more approach" in this SS Phase vs. our LN2 friends.



Best regards,

Randi

[chew*]

I see you like my timings

Would not have posted it here if i hadn't tested and knew for a fact it wasn't faster

[MrIcee]

Yes my man Its interesting how the score will drop if certain sub timings are too tight.
Never thought I'd see timings that could be "too tight". Anyways, I'm happy with this one. Could you imagine this tuning,
jack the multiplier on LN2 and the results? Especially with the NB at 4.5+ Ghz.

Randi

[chew*]

Could you imagine this tuning,
jack the multiplier on LN2 and the results? Especially with the NB at 4.5+ Ghz.

Randi

I have already done math as i always do before i waste time on ln2. No sense in doing it if it doesn't add up.

[pcnazz]

That's awesome Randy really man bravo ! I'm amazed at how well the Dominators do with AMD ( your timings ) , I couldn't come close to tight timings with i7 9xx / x58 like AMD is allowing . I pulled the plug until windows 7 hits tomorrow . Five years layoff Randy , you picked a good time to come back , would you ever believe AMD right about 4000MHz nb & over 2100MHz mem with timings like that , on your cooling , nasty nice my man .

[chew*]

Last run tired of this bench.

[ViViD]

WoW Chew Work Awesome
Low tRc Low CL
What's Ur Ram Model?

[pcnazz]

All around Nasty chew* !!! I like the lower HT myself . Fantastic run chew*

[Tomasis]

chew, thats awesome I wonder if u tried custom ln2 mem cooling like hiwa tried? would be nice with 2200 6-6-6 at thubans imc limit I need to see 32 ns

[chew*]

As always air to ln2 temps changes things. 2000 is the limit of board/cpu cold beyond -80....freezing ram won't remove a board/cpu limit.

[Lastviking]

To bad... inet didn´t work there we benched(cant upload score from a DDR3 platform?!, and to late to fix up the ddr1 stuff again.. and two more troubles(takes to long time to get DI, LN2 where i live..1 week+! , so it as to late), 2x SS did only go down to -25c! with 404 gas bumer.. we didnt find out what was wrong.

But here is the best one...to bad the save button did get bugged..



782.1 score was max that we did get saved..

302mhz 2-2-2-5 was highest with -10c temp on sticks stupid SS.

Thx to Ludran for nice bench!


Ps.. Top score looks buged on Hwbot juni...or do 2x512 give so much more?!..to bad i didnt use 2x512mb instead then.

[knopflerbruce]

Your sticks are nuts :p

What can these to on air cooling?

[chew*]

512 give more but they don't enhance latency......

I still say first score is bugged in ddr 1.........he said he changed hardware, alot of good when the culprit is software

[Lastviking]

512 give more but they don't enhance latency......

I still say first score is bugged in ddr 1.........he said he changed hardware, alot of good when the culprit is software

+1 on that one....it looks abit wrong if you compair it to the another scores...MaxxMem needs to be improved alot.

[chew*]

+1 on that one....it looks abit wrong if you compair it to the another scores...MaxxMem needs to be improved alot.

Yep the bench is horrid in currnet conditions.

Yah that score looks way way way way off......take a look at this.

We both had same ram speed but i had tighter timings + fastest mode + higher cpu which does boost score.

His


Mine


So we know latency lower is better.

The max pi memory score higher is better.

And also i think it calculates on time bench takes to run......

So i beat him on all 3.........yet lower score things that make you go hmm.