3000+ / Neo2 Platty 311x9 = 2800mhz!!

[bjc112]

Originally posted by: Zebo
BJC are you saying board won't post 300x9 with the 133 memetting? D told me you are using PC3200 Hyper X that's not "low end" by any means.

That's correct.

Regaurdless of timings, anything over 241 won't even post..

Drop it to 100 and boots up just fine @ 300x9

I actaully think I could get this thing to post @ 2.8 but haven't tried yet..

Right now I am

295x9

Cas 2/3/3/6

2.7V

Vcore 1.56-1.58

If everyone thinks 1.65 is safe I could definitely get 2.7ghz stable, don't know if it's worth it.

 

[Zebo]

133. One stick. 2.85Vdimm. 2.5-3-3-8 2T and work up from there. Lemme know how it goes.

Me thinks it's still board. Or worse mem controller. What was max mem in isolate mem phase?

 

[bjc112]

Originally posted by: Zebo
133. One stick. 2.85Vdimm. 2.5-3-3-8 2T and work up from there. Lemme know how it goes.

Me thinks it's still board. Or worse mem controller. What was max mem in isolate mem phase?

Well on the CBBID chip was unable to get it past 240 with 1 stick. Reguardless of voltage or timings..

I just know for a fact that this memory can run 200mhz 2/3/3/6 -> did in my nForce 2 1700+@2.4

Worked like a champ.

I have went through multiple boards and PSU's as well.

Not sure what it could be..

Just passed the SuperPI 32M as well..

I think 2.655 is damn near rock stable. Need a few more hours.

 

[Duvie]

I also suggested you do that in the PMs BJC.....


I think it is memory controller but it is related to a lot of the chips cause i see this all the time.....

Do not be hng up with the fact of "what it did" in other boards, or "what it is rated for".....When you OC and OC to levels of HTT you are at that all goes out the window....

Do not be stuck with cas 2......You should see if cas 2.5-3-3-7 can run at 133 which at 295 is only 380ddr will run and then test it....See if running 100 divider at 295 which is really low at 294ddr even with cas 2-3-3-6 will be slower and if so by how much...My guess would be if you could run 2-2-2-6 at that speed maybe it might cross cancel with the ram at 380ddr running cas 2.5-3-3-x, but at cas 2-3-3- I think the 133 divider would be better ofcourse if you can get it stable.....

 

[bjc112]

Do not be stuck with cas 2......You should see if cas 2.5-3-3-7 can run at 133 which at 295 is only 380ddr will run and then test it.

When usingthe 133 divider, nothing above 241 works.

Nothing.

Cas 3/4/4/8 whatever you want to run. No luck.

Reguardless of voltage.


I might drop the multiplier and see how I can go before the ram hits the wall..

 

[Duvie]

Do it in theory if you can get 133 divider to work with 9x241 it would have made the memory 9/.67 = 13.4 rounded to 14 = 9x241=2169mhz/14=155mhz or 310ddr....

Drop it to 8x313.....at that speed it would place the ram at 310ddr again.....

 

[Zebo]

241???

I'm still confused as to what happends here:

3. Memory MAX
- set mem setting to 200 and 2.5-3-3-7 1 T and give it boards max volts, or for LL ram set mem setting to 200 and 2-2-2-7 1 T and give it boards max volts. (up to 2.9Vdimm is safe with all ram)
-Set CPUs multi to 6-7, again to take it out of equation.
-Raise HTT up to boards max or until you start failing memtest86 test 5 and 6 http://www.memtest86.com/
-When you start failing memtest86 test 5 and 6, back down 5%, (HTT * 0.95), thats your max MEM speed.

 

[Duvie]

I am not 100% sure but the Hyper X 3200 stuff is CH5 modules, right??? I have heard the 1.36mod to 1.37mod bios works well for them and winchesters but I dont know about venices.....

May also want to try 1.51 beta, 1.52mod, 1.6, 1.62mod, 1.7 (I think) and 1.8beta.....Yours may be more of a compatabiity issue....


Another example of how ram speed and timings deterioate is my Geil PC3500 Ultra sticks...I believe they were CH5 as well....

They would do 240 1:1 with 2.85v and cas 2.5 for 480 ddr...However when I got to 290 range I could not run a 166 divider for roughly the same 474ddr range...Actually it would not boot and would not boot anything at cas 3...I had to run 133 divider but that allowed me to run cas 2-3-3-8 just fine at 372ddr...When I tried 8x330 and a 133 divider for about 440ddr I could not get cas 2.5 to be stable....

So it did 480ddr cas 2.5 and by the time I got to the end I couldn't get 40mzh lower to run and had to ultimately drop 100+ddr to get stable....

Gskills are the same....600ddr at 7x300 with cas 3, 1t.....8x300 need 2t to run....9x300 no boot....8x283 does cas 2.5-3-3-8....9x283 no go not even at cas 3-4-4-11, 1t.....


You damn near need decent to high end memory just to have some room left at the end...Now this may be more critical or realfor 3000+ owners cause we need such a high HTT to achieve same 2.6-2.7ghz speeds...However I have seen to many threads were ppl complain their meory wont run rated speeds as they OC only to have them drop the divider and push on....NOw if you start with crappy cas 2.5 PC3200 memory you wont have anything left.....BH5 and CH5 appear to be stuff that needs plenty of vdimm to hold tight timings so maybe this ram would have done better on boards with up to 3.0v options.....

 

[bjc112]

Originally posted by: Duvie
Do it in theory if you can get 133 divider to work with 9x241 it would have made the memory 9/.67 = 13.4 rounded to 14 = 9x241=2169mhz/14=155mhz or 310ddr....

Drop it to 8x313.....at that speed it would place the ram at 310ddr again.....

8.5x310 boots

8x330 does not, reguardless of timings.



Zebo will try that in a bit..

I know the memory is capable of 200mhz @ 2/3/3/6 (2.8v)

I know that in this Neo2 anything over 240-241 doesn't work.



Currently @ 296x9 (2/3/3/6)

 

[Zebo]

Originally posted by: bjc112




Zebo will try that in a bit..

I know the memory is capable of 200mhz @ 2/3/3/6 (2.8v)

I know that in this Neo2 anything over 240-241 doesn't work.



Currently @ 296x9 (2/3/3/6)

You keep reapting this stuff but what does that mean? In another board? with what dividers is this 241 failing?

Yes you need to isolate this memory on A64 first like I suggest. Then I can get an idea if it's mobo or memory.

 

[Duvie]

I am not so concerned with having same clock speed as seeing if it was the memory that is why I suggested 8x313...i was just trying to achieve near sam ddr speed as that of 9x241.....

If you are using 100 divider a 8.5x310 will be 310ddr so that is good I guess...should try a few notches higher like 312-314, etc see as you push past the 310ddr...

What slots are you using 1-2??? Have you tried 3-4???

 

[bjc112]

I believe 1 and 2..

Closest to the socket.

We switched that when messing with the CBBID chip.

 

[bjc112]

Originally posted by: Zebo

Originally posted by: bjc112




Zebo will try that in a bit..

I know the memory is capable of 200mhz @ 2/3/3/6 (2.8v)

I know that in this Neo2 anything over 240-241 doesn't work.



Currently @ 296x9 (2/3/3/6)

You keep reapting this stuff but what does that mean? In another board? with what dividers is this 241 failing?

Yes you need to isolate this memory on A64 first like I suggest. Then I can get an idea if it's mobo or memory.

With the 133 memory divider neither chip CBBID/LBBLE are stable over 240mhz.

REGUARDLESS of timings or voltage. In multiple Neo2's.

It must be the boards considering this ram is tested @ 200mhz (DDR 400 ) with tight timings.


With the 100 divider I can boot into windows all the way up to 310 ( for now )

 

[Zebo]

It must be the boards considering this ram is tested @ 200mhz (DDR 400 ) with tight timings.

Guessing

I have'nt seen you properly test the memory yet. Yes it's true the 133 mem setting inside MSI maybe fluxered which I'm leaning twards too, but at least see what your memory is truley capable of on that board!!! Screw CPU speeds right now We already know it's a good chip.

 

[Duvie]

Originally posted by: bjc112

Originally posted by: Zebo

Originally posted by: bjc112




Zebo will try that in a bit..

I know the memory is capable of 200mhz @ 2/3/3/6 (2.8v)

I know that in this Neo2 anything over 240-241 doesn't work.



Currently @ 296x9 (2/3/3/6)

You keep reapting this stuff but what does that mean? In another board? with what dividers is this 241 failing?

Yes you need to isolate this memory on A64 first like I suggest. Then I can get an idea if it's mobo or memory.

With the 133 memory divider neither chip CBBID/LBBLE are stable over 240mhz.

REGUARDLESS of timings or voltage. In multiple Neo2's.

It must be the boards considering this ram is tested @ 200mhz (DDR 400 ) with tight timings.


With the 100 divider I can boot into windows all the way up to 310 ( for now )

Noooooo!!! I dont care if it did this on a NForce 2 or any other board...that means squat!!!

I bet you does fine at stock 200 right??? Cause 400ddr at 241 would be using a 166 divider...133 divider was aleady taking a serious downgrade.....


I can tell you for a fact my ram will run (was tested thoroughly by Gskill to run 283 1:1 at cas 2.5-3-3-7....And it did on my board at 8x283....BUt it wont run at 9x283...I cant rma it cause it run the speed...It is clearly the AMD64 memory controller....It is not the mobo cause I see this all the times with others...It is not just my chip cause I see it all the time with other as well....

 

[Duvie]

Originally posted by: Zebo

It must be the boards considering this ram is tested @ 200mhz (DDR 400 ) with tight timings.

Guessing

I have'nt seen you properly test the memory yet. Yes it's true the 133 mem setting inside MSI maybe fluxered which I'm leaning twards too, but at least see what your memory is truley capable of on that board!!! Screw CPU speeds right now We already know it's a good chip.

It is not for me!!! I can run 133 at 330HTT fine...Did it with both Geil sticks and MarkFW900's Hyper X Pc3200 at 290 HTT....Nothing is wrong with it and I highly doubt you would get 3 Neo2 boards to do the exact same thing....

 

[Zebo]

Noooooo!!! I dont care if it did this on a NForce 2 or any other board...that means squat!!!

Exactly here's why: http://www.mushkin.com/doc/techSupport/papers/athlon64.asp

"Athlon64 3000+ and Higher
The Athlon64 itself uses a single memory channel and only one controller on the CPU. Because of this, the Athlon 64 (non-FX) can actually use plain vanilla, unbuffered, non-ECC modules. There are, however, still some peculiarities of the Athlon64 systems in how they are handling memory which effects memory compatibility.
One of the most critical issues in this resepect relates to the first step in the memory access sequence, that is, the opening of a memory page and subsequent steps until a read command can be issued. All of these steps fall into the time frame referred to as RAS-to-CAS delay or tRCD. In a typical system, the memory controller on the chipset is running at bus speed. On the Athlon64, it is running at CPU speed, which means that it is A LOT faster. That means that the internal sequence of commands also goes faster, and that's where all systems are getting hiccups if the memory chips used can only run at a high tRCD.
A very trivial explanation is that the controller expects that the memory is as fast as it itself but we don't have memory yet running at a 2 GHz and beyond clock rate. What it comes down to is that, depending on the BIOS tuning (by the manufacturer), the read command is issued a bit too early and violates the tRCD. The workaround is to add one additional cycle to the actual memory latency, that is, a 2:2:2-rated DIMM will run stable at 2:3:2; a 2:3:2 rated module will, most likely, need to be set to 2:4:2 in order to run stable (where the latencies are CAS:tRCD:tRP). Bottom line is that when shopping for memory for the Athlon64, it is highly advisable and in most cases mandatory to only consider memory with a tRCD of 2 or, to make it simple, memory that is rated at 2:2:2 a the designated speed. Alternatively, registered DIMMs can be used as well without problems, even though they are slightly more expensive and will give lower performance.
"

I repeat again you need to find mem max on this board.

 

[Jeff7181]

bjc112, that's the EXACT problem I've had with my CBBID Winchester and CBBLE Venice. I bet at 250 HTT it won't even POST and you have to reset the CMOS, don't you?

*EDIT* In regards to what Zebo posted just above mine... my RAM is rated at 2-3-3-7 at 217 MHz... I've run it at 2-3-3-10, 2-4-3-10, 2-4-4-10, 2-5-4-10, 2-5-5-10... and all those combinations once again with CAS 2.5 and CAS 3... and HTT above 240 is still unstable unless I set the memory divider at 100, and the multiplier low enough to keep the core speed under 2 GHz. I've tried to get help or advice on other boards, and people just assume I don't know what I'm doing since other people are getting HTT up to 330 on the same board... maybe they're right. But no setting I've tried has helped... and I've tried A LOT of settings.

 

[bjc112]

Originally posted by: Jeff7181
bjc112, that's the EXACT problem I've had with my CBBID Winchester and CBBLE Venice. I bet at 250 HTT it won't even POST and you have to reset the CMOS, don't you?

@ 133 it won't..

up to 310 with the 100 divider..


Anything @ 312 HTT 313HTT locks even at the 100 divider..

I am currently testing @ 295x9 =2655 using Cas 2/3/3/7/ 1T

 

[Jeff7181]

Originally posted by: bjc112

Originally posted by: Jeff7181
bjc112, that's the EXACT problem I've had with my CBBID Winchester and CBBLE Venice. I bet at 250 HTT it won't even POST and you have to reset the CMOS, don't you?

@ 133 it won't..

up to 310 with the 100 divider..


Anything @ 312 HTT 313HTT locks even at the 100 divider..

I am currently testing @ 295x9 =2655 using Cas 2/3/3/7/ 1T

Yeah... if I could get HTT to about 400 MHz I wouldn't mind using the 100 divider... but that's not going to happen.

 

[Duvie]

Well in his case 9x295 = 2.65ghz and I am sure even running 295ddr 1t cas 2 is better then running the cpu at 9x241=2.17ghz with 133 divider at 310ddr....Heck you could run the 241 at 1:1 and it would still be faster for the 2.65ghz....If you say he takes a 10% hit running low DDR (now 1t) which I think would be very aggressive then he would still be faster to be at this speed then take a drop to 2.4ghz with 267HTT and run 1:1.......

You do what you got to do to get the clock speed where it really matters anyways..ram bandwidth not as much as timings...if he can hold 1t then he doesn;t take that hit so he probably at 2.65ghz can hang with my system at 2.57ghz.... think he would be happy....Since he couldn't get to 2.57ghz without using the 100divider...


If he can hold 2.65ghz at 1.57v range means you likely are stable at default vcore in the 2.5ghz range....That would put it on par with my cpu....

 

[bjc112]

Originally posted by: Duvie
Well in his case 9x295 = 2.65ghz and I am sure even running 295ddr 1t cas 2 is better then running the cpu at 9x241=2.17ghz with 133 divider at 310ddr....Heck you could run the 241 at 1:1 and it would still be faster for the 2.65ghz....If you say he takes a 10% hit running low DDR (now 1t) which I think would be very aggressive then he would still be faster to be at this speed then take a drop to 2.4ghz with 267HTT and run 1:1.......

You do what you got to do to get the clock speed where it really matters anyways..ram bandwidth not as much as timings...if he can hold 1t then he doesn;t take that hit so he probably at 2.65ghz can hang with my system at 2.57ghz.... think he would be happy....Since he couldn't get to 2.57ghz without using the 100divider...


If he can hold 2.65ghz at 1.57v range means you likely are stable at default vcore in the 2.5ghz range....That would put it on par with my cpu....

Well SuperPi finished in 30m 28sec @ 2655.

http://pics.bbzzdd.com/users/bjc112/suerpi.JPG

System defniitely seems stable so far @ 1t as well..

 

[Duvie]

Not bad that is not the optimised one..You could take off at least a full minute plus with it....May want to search for the SSE2 version and run it again...I am trying to find where you relate to me at my tested 2.66ghz to give yo an idea how much that 100 divider and lower bandwidth hurts you...Superpi is one that like bandwidth a bit more...Most apps except games are not this nutty for it....

Good job I am happy you are finally out of the sub 2.2ghz range....

 

[Jeff7181]

Originally posted by: bjc112

Originally posted by: Duvie
Well in his case 9x295 = 2.65ghz and I am sure even running 295ddr 1t cas 2 is better then running the cpu at 9x241=2.17ghz with 133 divider at 310ddr....Heck you could run the 241 at 1:1 and it would still be faster for the 2.65ghz....If you say he takes a 10% hit running low DDR (now 1t) which I think would be very aggressive then he would still be faster to be at this speed then take a drop to 2.4ghz with 267HTT and run 1:1.......

You do what you got to do to get the clock speed where it really matters anyways..ram bandwidth not as much as timings...if he can hold 1t then he doesn;t take that hit so he probably at 2.65ghz can hang with my system at 2.57ghz.... think he would be happy....Since he couldn't get to 2.57ghz without using the 100divider...


If he can hold 2.65ghz at 1.57v range means you likely are stable at default vcore in the 2.5ghz range....That would put it on par with my cpu....

Well SuperPi finished in 30m 28sec @ 2655.

http://pics.bbzzdd.com/users/bjc112/suerpi.JPG

System defniitely seems stable so far @ 1t as well..

Yeah, I got about the same results... but with RAM at 217, and my CPU at 2475 with my Winnie, it finished in about the same time. So the ~125 MHz gain in CPU speed was negated by the running my RAM at 186 MHz instead of 225.

 

[bjc112]

Originally posted by: Duvie
Not bad that is not the optimised one..You could take off at least a full minute plus with it....May want to search for the SSE2 version and run it again...I am trying to find where you relate to me at my tested 2.66ghz to give yo an idea how much that 100 divider and lower bandwidth hurts you...Superpi is one that like bandwidth a bit more...Most apps except games are not this nutty for it....

Good job I am happy you are finally out of the sub 2.2ghz range....

Any where to get the patched version? NM, i'll check the thread..

I really appreciate all the help. I was pissed this afternoon when I was stuck @ 2.16

If I can take a minute off that time ending around 29M, that still beat's quite a few systems, which would make me ok with this lowly 100 divider..



Just played Hl2 for quite awhile with no issues..

Still not leaving 2700mhz completley out, but again, don't think the added Vcore is worth it.



Thanks again everyone!