Something Interesting With TCCD

[MaxxxRacer]

I was messing around with my TCCD on my nf2 (since I have killed my winnie) and I set the voltage to 2.9 volts and I was able to get 2-2-2-6 timmings (didnt try 5) and its totaly stable.. As far as I know most tccd doesnt do these kind of timmings at this mhz..

I find this kind of interesting as Saaya recently posted about the volt loving TCCD. now I am not able to put 3+ volts through my memory (without vmod which im considering) but just a guess is that I have one of these dimms. I would do the vmod to see if i could get a higher fsb, but this particual nf7-s seems to max out at 225fsb.

[blinky]

yah tccd needs a little higher voltage to run 2-2-2 timings

[jikdoc]

tccd doing 222 timings with volts has been known for a few months now. it's recently had a revival but it's nothing really new as bh5, vx does 222 at higher htt

[MaxxxRacer]

i know i know.. but was just showing some more data here... kinda excited that mine can do it.. I just wish I had a board that could do over 225 fsb... I suppose its time for me to do some suicide runs with the memory to see if the board will love me for a few hours and run at higher fsb..

http://www.xtremesystems.org/forums/attachment.php?attachmentid=23918&stc=1

[goreblast]

MaxxxRacer - the speeds you are doing there are only slightly higher than I guess most people get at stock volts with their TCCD memory on an Athlon 64 system.

My OCZ Plat Rev2 will do 2-2-2-5 at 220Mhz at stock voltage, just so happens that with my set increasing the volts to silly levels using the OCZ Booster doesnt get me a better overclock. I have change to 2.5-3-3-7 timings to get a higher FSB. Now if you could push that to 230-240 range it would be very impressive whilst keeping the tight timings. Also the nf2 chipset prefers 2-2-2-11 timings, and with slightly looser timings you might get a better overclock.

What happened to the Winni - too many volts or did you crush it?

[MaxxxRacer]

I removed the IHS on the winnie... nuf said?

goreblast: that is interesting. I thought that most ppl were getting around 210-215 on average with 2-2-2-5 timmings..

I know the nf2 likes the 11 tras (i think its tras), but i was just trying to show what it would do.. ppl would complain that it was 11 and not 5 or 6..

Like I said, I will do some testing when I get my nf3 system back and running (hopefully venice comes out soon as I would much prefer that over another winnie).

I will loosten up the timmings and give the nf2 overclockign another run. we will se how high I can get it stable. If I can get 235-240 I will do a vio (3.3volt) direct mod and see how well the memory does at those speeds and report back.

I guess the reason that this is such a good thing imo is that the tccd can go sky high with the clocks, but usually wont go above 215~ unless you loosten the timmings. And wahts soo anoying about it is that from 215-275 it uses the SAME freeking timmings... why cant we get a middle ground with some lower timmings for the 220-250 range.

[MaxxxRacer]

Well Ive been doing that testing, and Im up to 240fsb and am about to do a few super pi runs to check for preliminary stability. it seemed fine and dandy at 235 and 230.. we will see how it goes.. timmigns at 2.5-3-3-6.

Well I have struck gold here.. at 2.9 volts, I got to 245fsb which seems to be the new limit of the mobo. And as you can see in the screenie, I got 2-2-2-6 timmings.. I am a happy camper... looks like i got the best of both worlds and not even at extreme volts!!!

http://www.xtremesystems.org/forums/attachment.php?attachmentid=23928&stc=1

[goreblast]

Looking good mate.
I was running the OCZ TCCD on an Abit AV8 (via chipset) with a NC3500 - I felt that combo wasnt giving me the best from the RAM as everyone raves about how high it will go with loose timings, but for me with the whole system maxxed out the highest I can get no matter what timings I use, either in the bios or using a64 tweaker, is 245. If I underclock the CPU i can hit 270, so me thinks it could be a memroy controller issue, or possible chipset/motherboard. Got Neo2 on the way to see if it helps.

I know what you mean re the tras, I guess a lot of peeps who are using a64 now may not have used the nf2 chipset and will think you are running loose timings at 11, when for that chipset as we know it beats tras 5 or 6. Not sure exactly why, but it does.

[MaxxxRacer]

It beats it becuase of the syncronization (sp) that you get with the tras of 11.

Well the tccd memory is 2.5-3-3-7 at 250mhz stock. those are the timmings that samsung sets for the memory out of the factory.. so its DEFFINETLY the board and not the memory....

[FallenAngel]

Very good taking in consideration that is a nf2

[goreblast]

Damn buggy Abit board !

[DevilsRejection]

wow way to go maxx, this makes me more confident that my rev 2 will scale well with voltage, hopefully to 280Mhz, at 2.9V, thats the max the shuttle provides.

[MaxxxRacer]

yah, im really liking this abit mobo... If I had the right chip i would break my prevoius 37 secon super pi 1m record. (that record is my own for nf2, on nf3 i hit 31)

Ive been running it at 245 for a while now and NO problems. I tried 250, but it was a no go either way.

[DevilsRejection]

Was this at 1T tho?

[MaxxxRacer]

no, nf2 cant do 1T with a gig.. it just cant do it. But the good thing is that 1T and 2T dont make a dam bit of difference between them when used in nf2. the difference is barely even noticable in benches thankfully.

here is a little benchmarking page i did a while back. I didnt realize it but all of them were cpc on (1t)...
http://www.angelfire.com/creep/maxxxracer/Bios_Mods_Scores.htm

[EnJoY]

The difference is very noticable in 3d mark and games though Maxxx. :/

Also, a point worth noting is that at 2T, that ram be doing things it won't do at 1T even in an nForce 3/4.

[MaxxxRacer]

for nf2 its what you had to do... and to be honest, the difference isnt noticeable. when i got off my but i will go do some benches at 200mhz with cpc on and off to test. if you check out the website i posted above, I will do benches similar to that.

[EnJoY]

Dude, the reason I said what I said is because I know it as a fact. It's been tested before. When the NF7-S 2.0 and the Infinity were wildly popular, a few tweakers ran tests...I think Saaya or SAE may have been one of them, that compared 3dmark2001, sandra and like super pi with cpc on and off. It made a big difference in 3d,,,equivilant to like 30mhz or so. I have a gig of KHX BH5 in my lanparty nforce 2 and they wouldn't do :banana::banana::banana::banana: above 205 for the same reason all gigabyte modules suck in nforce 2's. But turning cpc off, although yeilding greater fsb overclock, generally was not worth it unless you could gain around 30mhz or so and then it would really just equal out 200fsb with cpc on...so most never saw the point.

[justwOo]

Nice work!!
Have you done mods on the board?

[MaxxxRacer]

no mods on teh board.. no need.. the board does 245mhz at 2.9volts 2-2-2-6. the board tops out at 245mhz despite any timming changes.. i set it to 2.5-3-3-7 and it iddnt help.

Enjoy, I am in the middle of testing, and I have data that shows otherwise. I am using tictacs D26 black mantaray L12 cpc on and off. with it off I am seeing better everest (picked it instead of sandra as i hate sandra), and better super pi by one second. the differences are very small. I have also ran a test of 3dmark01 and it yieleded the same thing.. the score was inbetween the two scores that i got with cpc off.. I will post in about a half hour with the results.

[MaxxxRacer]

here is the benchmarking that i promised. A note that with the everest benchamrking, i removed the top and bottom score and took the average. the numbers listed are the ones that are in the average.

http://www.angelfire.com/creep/maxxxracer/CPC_Benchmarking.htm

EDIT: Ive also been messing around with 1T timmings and I am at 2-2-2-5 at 235mhz right now.. I will see how high I can go.. it seems that my NF7-S just decided to heal itself and love fsb speed now..

EDIT again.. appreas that wasnt 1T.. i will make sureits 1T next time... will do that now.

[MaxxxRacer]

Ok here we go. here is the testing with the real 1T command rate.. IT seems that when 1T is enabled that it blows in 3d01 and in super pi2m by a rather disturbing margin compared to 2T. I think it might have something to do with the 2T command rate being L12modded. But both of the tests were done with the exact same settings besides that. (I made sure of this by not resetting the boot record when i flashed or the ecsd)

http://www.angelfire.com/creep/maxxxracer/CPC_on_and_OFF_Benchmarking.htm

[MaxxxRacer]

Well I did some more testing with 1T and L12 mod mantaray XT bios and found that its 2 seconds faster in super pi2m than the L12modded 2T bios.. The everest bandwidth test show the two 1t bios to be the same, but the one mantaray1t bios had some issues that made it suck really bad in super pi.

[HMB]

On NF7s V2.0 with CPC on and 2x512 Geil ultra X i get around 220MHz 2-2-2-11.
With CPC off i get 240MHz+. Performance sucks though.

[MaxxxRacer]

hmm.. yah im getting NOWHERE with cpc on 2x512... ohwell.. I will be going back to cpc off.. I cant get it stable at any speed with 1T timmings.

[MooCow2]

lol my gskill LE does 2-2-2-11 at up to ddr 238. if i want to do 240 i have to make it 2-2-3-11. both at 2.9v

cant get it to work at 2.85 and under though.

on nf7 2.0 before the board died.

[MaxxxRacer]

mice man. looks like some of us are getting the magical tccd.

[FallenAngel]

I just got mine today, though no board neither cpu, waiting my DFI and money for the cpu