Thx for clearing that up, will be interesting to see how far you can scale tRCD 8 vs tRCD 7 @ CAS 6.
I'm only just starting to play with mine,
http://img300.imageshack.us/img300/9...5268981788.jpg
Thx for clearing that up, will be interesting to see how far you can scale tRCD 8 vs tRCD 7 @ CAS 6.
I'm only just starting to play with mine,
http://img300.imageshack.us/img300/9...5268981788.jpg
what s the difference between 8-9-8 and 8-8-8 for your kit ?
for voltage or 32m?Quote:
Originally Posted by Pt1t
for 32M with same VdimmQuote:
Originally Posted by raju
My CPU's IMC does not seem to like tRCD 8 over/at 2400MHz for 32M at all (tried 2 different kits now from 2 vendors). I think the jump is 0.04V+ at the same freq if I guess. Can you try on yours and see how much the voltage increases?
I'm having some fun with these sticks cold (I have not cross compared to the other kit I have yet so this will be interesting).
http://img40.imageshack.us/img40/4070/gs26008981788.jpg
Last one for today;
http://img246.imageshack.us/img246/5...6228981788.jpg
did you run 2k6 at 213 bclock/2560 mem and then bumped it up before you took the screenshot or is fm still not reading cpu speed properly?
very impressive either way!
I did not close 06 when I ramped BCLK so yeah the real clock was what is shown in CPU-Z (it did run the bench at the speed shown).
here's the everest (i need to update this to the latest beta too).
http://img2.imageshack.us/img2/7987/...m2622898gs.png
so it IS 2k6 stable at 2600 but on this screenshot the 2k6 run is actually from 2560? does 2k6 scale at all with more mem bw?Quote:
Originally Posted by raju
how is real world mem scaling in 1156 seeing as its only 2 channels?
at 1366 it pretty much stops at 1333, how about 1156? is there notable real world scaling above 1333?
No the run in the screenshot is from 2622MHz (you can cross confirm with the CPU score). 06 was left open for every run from first boot, I ramped BCLK then re-ran the bench every time - so the score you see is from 1311. I just wanted to show that you can bench at these speeds, I'm sure you've seen people do that before.Quote:
Originally Posted by saaya
Dunno about real world effects, I have not tested that kind of stuff (it's not what I do). I'd imagine its the same as it's always been, in that it makes jack difference..
later
Raja
hey raja was your RAM subzero?
very nice clocking
have you done any 32M SuperPi at those speeds :up:
hey Thomas nice kit of RAM
:up:
Very nice stcks raja,were them under cold? :up::up:
2622 even... hmmm its a real shame uncore cant be ramped up on 1156...
any word from intel if theyll fix that with 32nm or future skus? :/
Yeah guys, the sticks are cold - cascade was at -87 but contact is poor so I'd imaging the sticks are at -40.
Need to get one of the pot builders to make a good DDR pot (a platter with sliding heatspreaders will suffice).
Dinos, 32M is beyond my CPU up this high. 3D is fine. Aq3 will be interesting for the guys that like it.
Aha... it's under cold.....really niceQuote:
Originally Posted by raju
Do you have a picture of cascade evap on ram? i need to try also this king off ghetto stuff :DQuote:
Originally Posted by raju
Lol, I'll send you a pic when it all dries out. Get some nice big zip ties :)Quote:
Originally Posted by Pt1t
you didnt see it in the graveyard thread? :DQuote:
Originally Posted by Pt1t
raja, i still dont get it, i think fm is still not reading cpu speed properly...
it says 4050, and mem is 2600/12=217 bclock...
4050/217=18.7 cpu multi :D
2622/12=218.5 bclock
4050/218.5=18.5 cpu multi :D
so either you took this shot after just upping mem and before you ran 2k6 again and the 2k6 run was 213x19=4050 and mem at 2560, or fm detects cpu speed wrong, STILL :rolleyes: and it was actually 216x19=4100 or 218x19=4140
Ignore the reading in FM Sascha - BCLK is around 218.5 approx X19 CPU clock on 2:12 - this gives the 2622.
I upped the BCLK by 1-2 each time and then reran 06. The first time I opened 06 was when I first loaded windows. I eneded up at 218.5 when I re-ran 06 - I did not close 06 between ramping BCLK (apart from on the very first run) that is why it is still reoprting the CPU speed from the 213 run (I did not post all screenshots in between). WHat you see in CPU-Z in each screenshot is where 06 was run each time. Can't put it any simpler. I hope that is starting to make sense for you now. :)
the GPU tests have a habit of jumping around, the CPU test does not.
Keep reading this until it makes sense as I'm not going to explain it all again :)
The initial run which I did not post here as I got it higher was at 213, that is what FM read when 06 was loaded at 213, it just does not pick up when you change BCLK (it does not probe clocks again after it's launched).
geez clam down...Quote:
Originally Posted by raju
that really sux then, fm promised theyd get hardware detection right with their new patch... then again if they detect the speed correctly when its launched the first time thats good enough, if thats really reliable :D
I'm calm dude, no worries.Quote:
Originally Posted by saaya
yeah, it is accurate on first launch (at least it has been so far).
how well does this stuff with cas5? not worth it compared to cas6 i guess?Quote:
Originally Posted by raju
id die if id see 1800 or even 1866 cas5... that would be so sweet... :D
oh and what about ddr3 cas4? i remember some people saying its possible in theory a while back? there was a bug of a giga board showing cas4 when it was cas5, but several people saif cas4 is actually possible in theory?
i def prefer tight timings over high speed... :D
Not sure on CAS 4 or 5 really. I did fool around for a while with CAS 5 on some older sticks and was benching around 1400. I have no idea if these modules will go higher than that at this point (the Corsair seems better at CAS 6 than the GSkill (per mv) so might be the better choice).Quote:
Originally Posted by saaya
2600 CAS 8, is like CAS 4 1300 for real latency so there is some progress for DDR3 I guess. Just goes to show how well Intel have nailed IMC setup, it's all automated and works very well.
regards
Raja
I reran just for you Sascha. Ambient are a bit higher now so had to up Vdimm by 0.02V due the the warmer evap temp (1.808V real);
Anyway, I hope this clears up the 06 reading the CPU speed stuff etc. Its from a fresh boot as the scores always seem best on the first run..
http://img36.imageshack.us/img36/5556/gs26228981788.jpg