Thuban + Hyper speeds + D9 vs PSC?

[BeepBeep2]

Callisto, Deneb, or Thuban, doesn't matter to me!



Why must I ask? Because my DDR3-2000 CL7-7-7-20 kit has been running...

Prime95 Blend for 50 minutes at DDR3-1870 (935) 7-7-7-20-27 1T at 1.5625v set...(reads 1.550 in bios!)

EDIT: Failed at 52 minutes :p

On my 565BE!



1884 failed after one minute, didn't suprise me, but I was on the edge of my CPUNB clock stability XD

Same kit that won't boot over 1910 with Thuban at any CAS Latency

What's the highest you have achieved?

[NEOAethyr]

2040 is the highest I've primed stable so far (I think, pretty sure it was 255, if not then 252, 254, something like that, just over 2ghz).
255x8 = 2040mhz mem (1600mhz div, 8x).

I wasn't able to get much higher.
I might of been able to get to say 257x but it wasn't worth the effort.

Right now I just leave it be at 2000 with my main's set and the sub timings on auto.
Previous set of mem did 8-9-8 I think, and this set does 8-9-7 at those speeds.

I have to get either 2x2gig or 2x4gig set later on, seems I need to focus on getting a new cpu cooler 1st though.
I need one of those corsair aio water cooler thingys so I can direct the air out my case, plus it's probably 10x better then this stock pos.
Those coolers as so expensive though, $80+, wish it was $40 lol.

I wish the cpu had more in it in terms of mem clocking.
I'm able to do 4.1ghz on the stock cooler, right after redoing the tim (tim's never last for me :\ ).
And over 3.1ghz on the nb.
So she's a good cpu.
But it's a shame she doesn't do 2.2ghz mem lol .
That would of been nice lol.

[chew*]

Yawn........

8 gig 64 bit + more than 1 hour stable Bull.......at reasonable voltage ( aka lower than spec ).

[Daveburt714]

The best STABLE clocks I ever got with mem was the same one I used for Partcle's OC chart.

I've gotten higher speeds, but rarely stress my system to that point...

1090T, Mushkin Blackline Lovo's, 1750 6-9-6-18-30 1T @ 1.60v.

Here's the post if you want more info:
http://www.xtremesystems.org/forums/...&postcount=678

[slaveondope]

Someone give Beep some LN2 XS needs some more boints

[BeepBeep2]

Yawn........

8 gig 64 bit + more than 1 hour stable Bull.......at reasonable voltage ( aka lower than spec ).

I usually test 1 hour for personal use, 4 hours if I'm releasing results in a blog etc.
I just want to see the max of what people can get out of their Deneb and Thuban IMC's through Prime Blend

How much harder is 2x4GB 8 gig? I understand that 4x2GB would be harder on IMC but considering results you got in 32M 2x4GB looks like no problem.

I'd still be glad to send you this kit for testing, its not performing any better freq. wise on Thuban than Deneb and Callisto, tested it on all three CPU's, 965, 565, 1100T

Someone give Beep some LN2 XS needs some more boints

Ha, I wish I had the cash for LN2. I've got the time and materials, just bought a thermometer, multimeter, pot and a crapload of eraser some towels and pipe insulation. I dont have spectacular CPU's or the money to buy more CPU's though to bin.

Apart from that I don't have a quarter of the skill or knowledge someone like chew* has, I'm just a kid that's still attempting to grow up

[chew*]

8 gig on thuban is not to bad, deneb gets a touch harder as well as adding 64 bit to the mix which is a requirement for 8 gig.

3 hours is the min I would use to claim stable, just my 2c.....credibility is everything, mislead endusers with results and well you give yourself and a brand a bad name.

Example if i used 32m pi alternatively for the review or tests I did for that ram I think some people intending to use this ram for 24/7 would be peeved, considering obviously not even a retard is going to use this particular set to break some 32m record.....why then use 32m to denote stability......

There's quite a discrepancy from what people use in there reviews to reality of a 24/7 rig........using this to prove stability = lying to endusers that stable is 40mhz higher than reality........

[BeepBeep2]

8 gig on thuban is not to bad, deneb gets a touch harder as well as adding 64 bit to the mix which is a requirement for 8 gig.

3 hours is the min I would use to claim stable, just my 2c.....credibility is everything, mislead endusers with results and well you give yourself and a brand a bad name.

...
...
...

Makes sense. I'll stick to 4 hours then from now on.

I swapped out 565 for 1100T...

Began Prime95 Blend at 1886 7-7-7-20 @ 1.6v (same freq you show on that 560 above, a step above what made it 52 minutes with the 565 at 1.565v) and my system hardlocked immediately.

~1880 is fine for booting and light benching, don't think it would even pass 32M.
1870 1.565v set somehow made it 52 minutes in Prime95 Blend on the 565 BE, yet failed loop 9 (I'm really starting to wonder about 32M on this setup, it never finishes) in 32M

I'm about to start Prime95 Blend on the 1100T at 1870...albeit at 1.6v set to be safe.

I got to the logon screen at 1940, then a reboot. 1920 made it into windows but startup services were crashing after I logged in.

Would you know of anything that could explain this type of behavior?

[chew*]

ram not stable.......to much volts, ram is heating up.

[BeepBeep2]

ram not stable.......to much volts, ram is heating up.

There's no way.

You can't tell me 1.565v to 1.6v is too much, I've got a 120mm fan directly on them...
If it made 50 minutes at 1.565v you would think 1.6v would be enough to make it stable...

I've Primed these sticks at 1.7v before, 1800 C6, I'm running 1880 C7 now, the ram isn't overheating.

[chew*]

prime 95 heats up hypers, hypers run hot period.

[BeepBeep2]

prime 95 heats up hypers, hypers run hot period.

http://www.techreaction.net/2010/06/...3-2200-review/
I ran all those tests on that kit 5 to 8 hours, each one passed. I'm pretty sure this kit doesn't run all that much hotter, correct me if I could be wrong.

Literally speaking, whats next, running them at 1.4-1.5v?

Like I said, trying to boot at DDR3-1920 7-7-7-20 on thuban was not stable enough to last 5 minutes under no load, startup prog's were crashing.

My D9's boot on Thuban up to 2020 9-9-9

[chew*]

I'm not saying you didn't run tests but as far as I can see or anyone else can see you proved in that blog you can take cpu-z screenshots at XXX speeds and nothing more.

I don't know what to tell you other than hitting decent speeds and furthermore proving it requires time and energy. There is no and are no shortcuts.

[BeepBeep2]

I'm not saying you didn't run tests but as far as I can see or anyone else can see you proved in that blog you can take cpu-z screenshots at XXX speeds and nothing more.

I don't know what to tell you other than hitting decent speeds and furthermore proving it requires time and energy. There is no and are no shortcuts.

Wonderful, looks like I oughta start a 4 hour 1800 6-6-6 P95 run, if I can pass that will you offer to take these sticks off my hands, or help think of other possibilities?

I literally just BSOD'ed at 1880 9-10-9-27 1.5v.

[charged3800z24]

Prime95 never uses all the RAM, I can't see it as a 24/7 stable MEM tester. What ever happened to Memtest pre windows?

I'll put my G-Skills back in tomoorw maybe, see what they have. They would bench at decent settings, but doubt they do stable much over stock.

[chew*]

Prime95 never uses all the RAM, I can't see it as a 24/7 stable MEM tester. What ever happened to Memtest pre windows?

I'll put my G-Skills back in tomoorw maybe, see what they have. They would bench at decent settings, but doubt they do stable much over stock.

True but for ram to be stable the IMC must be also and nothing I know of really hammers the IMC and ram like blend.

Memtest is always first and tbh memtest stable is not stable either.

[charged3800z24]

True but for ram to be stable the IMC must be also and nothing I know of really hammers the IMC and ram like blend.

Memtest is always first and tbh memtest stable is not stable either.

hmm, how about run both... get a combo of the two Should get you a good idea of stable.

[BeepBeep2]

True but for ram to be stable the IMC must be also and nothing I know of really hammers the IMC and ram like blend.

Memtest is always first and tbh memtest stable is not stable either.

I used to use memtest, I had memtest86+ on my USB drive...not sure why I quit using it, partially due to the fact that it didn't show true stability nothing really does, but an hour of blend usually never BSOD'ed or locked up through anything, it was stable enough for what I need it for, which is actually mostly gaming and video/photo editing (which suprises me... I work with large 20MP+ images (2GB+ in ram) in photochop and gimp all the time without issue)

Running Memtest for Windows is probably the dumbest thing I ever did. It makes your system so unresponsive I just couldn't live with it. Simply stopping the tests can take an hour or more, and taking a screenshot over half an hour.

Not to brag, but 1800 6-6-6-18-24 1T 90ns 1.7v set (reads 1.69 in bios) is 22 minutes into P95 Blend right now with the 1100T. It's not stable, not even close yet, but it's a lot better than crashing at boot at 1900 9-10-9-27 1.5v (I felt 1.5v would have been PLENTY?...kingston LoVo are MGH-E rated 1866 C9 @ 1.35)

[chew*]

hmm, how about run both... get a combo of the two Should get you a good idea of stable.

Heh I always do that anyway.

[BeepBeep2]

Heh I always do that anyway.

so here's the "One hour stable bull"


Two hours, 2 minutes in, looks like heat got the best of them - I would still run this 24/7, it's stable enough for me personally, maybe not for others.
Core 3 was first to go, then the others just went down promptly one after the other.

Still doesn't explain why my 1100T is no better than my 565BE as far as frequency scaling at CAS 7.

[NEOAethyr]

You ever try 9-9-9 and go from there?
Once you reach a good spot, lower the timings back down.

Btw, you guys should really post aida64 or everest benchmark scores, the generic one that shows all 3 mem benchs on the same screen.

I'll post my results when I can.
My pc is on da fritz right now lol.

[Royalk]

2000 6-9-6 on C4E, not tested too much
1.7v Vdimm, PSC chips

[chew*]

so here's the "One hour stable bull"
[IMG][/IMG]

Two hours, 2 minutes in, looks like heat got the best of them

Really the quote says it all......

Seriously if you can pass 3 hours at one speed and not 3 hours at another speed do we need to paint a picture?

If it looks like a duck it quacks like a duck guess what it's probably a duck.

[BeepBeep2]

Really the quote says it all......

Seriously if you can pass 3 hours at one speed and not 3 hours at another speed do we need to paint a picture?

If it looks like a duck it quacks like a duck guess what it's probably a duck.

What I'm saying -
1900 8-8-8-24 1.55 to 1.7v hardly boots into windows. (What happened to "You can bench on Hypers up to 2000") I would like to know why. I'm not looking for stability at 1850-1900 6-6-6 or anything absolutely stupid like that. What I want to know is, why, at CAS 7, 8, AND 9 I can not find ANY IMC headroom AT ALL with thuban over Deneb.


The only ****ing reason I ran that test in the first place was because it seemed to you like my problem with booting windows at 1900 9-10-9 was because the sticks sucked.

[NEOAethyr]

1900 @ 6t and ducks are the devil...
So momma always says lol, just messing .