RAM type/speed for new system

[Guest Nick_N]

I see absolutely no reason why you won't match what you saw with pushing the DDR2 to the max, and then someWith the processor you are using that will also play into it. Your not bottlenecked any where near as bad and your sticks are not going to get red hot either

[Guest D17S]

The hobby of Flight simmin' will inevitably draw a user toward the additional (but separate) hobby of computer tech. That's a great hobby too. Integrating these two hobbies is a challenge as a user finds these priorities increasingly difficult to separate. If a hobbiest retains simming as the priority, the effort / dollars / time / frustration expense (by the example of memory subsystem optimization) is a vast over-expense relative to its benefit. Reasonable advise would recommend the use the DDR2 platform and wait for this advanced computer tech to drop-in to the high-end mainstream. Then spend the bucks for stabilized results. The best advise to a builder now Continues to be this: Sit tight with any modern core/ddr2 system and see what Nehalem offers (this is 4 months away now, for goodness sakes!) With the little bit of reason I've been able to apply on my own (over belligerent objection, I might add), Nehalem might be able to provide a significant improvement in memory subsystem performance. Wait for that, rather than trying to force it on an ill-equipped FSB platform. That's Still good advise. If a hobbiest retains computer tech as a priority, this massive expenditure is all well spent. . . actually, this expenditure is without option. Simming becomes a more of a proving ground for these technology experiments, relighted to the role of a secondary beneficiary gaining only marginal rewards. However despite all else, we appreciate ya'll blazing this technology trail (especially with Your hard earned dollars!).

[Noel 4,507]

Here's some real comfort from a Corsair forum for someone who wants to go 1800Mhz with their 2x2 sticks you linked, with a QX 45nm CPU. It was not on an X48 board tho:"It's not a given. You would need to raise the FSB to 450Mhz to attain the 1800Mhz for the DRAM and the 45nm Quad Cores are not easily taken that high. The 45nm Dual Cores do handle that FSB, but the Quads are a hit and miss with more missing than hitting. You have 2 X 2048MB Modules?I would advise you to move to 1600Mhz C7 DRAM for your system. Then you can just raise the multiplier for the overclock and leave the FSB at 400Mhz with a 1600Mhz DRAM speed.400 X 10 = 4.0Ghz with 1600Mhz 7-7-7-20400 X 10.5 = 4.2Ghz with 1600Mhz 7-7-7-20 400 X 11 = 4.4Ghz with 1600Mhz 7-7-7-20Now you might be able to make 425 or so, maybe lucky enough to make 450 also and then would need the 1800Mhz DRAM. But it's not a given.425 X 9 = 3.825Ghz with 1700Mhz with 7-7-7-20450 X 9 = 4.0Ghz with 1800Mhz with 7-7-7-20450 X 8 = 3.6Ghz with 1800Mhz with 7-7-7-20See? The issue is not the speed of the processor. The issue is getting the FSB up to a ratio that will work with the DRAM. Use a good aftermarket cooler with that CPU. Now to get the CPU FSB up that high, you need to raise the voltages. The CPU does not need very much of a voltage raise but the CPU VTT, PLL and additional voltages will need quite a bit of tweaking to make higher than 400Mhz on the FSB."______I very much like the idea of keeping the FSB as close to 400 as possible. Too close to the bleeding edge at 450. Would I be able to run the Corsair 1800 at 1650 @ CAS7? A crap shoot? That would get me to CALWI 7.5, with bunches more bandwidth. And a more relaxed FSB. This is what I'd like to do, 1650 at CAS7. Perhaps nothing will do that reliably.As I say, I don't have the stomach or time for living on the bleeding edge. Nor do I care if it competes with Nellie. Just want the fast enough PLUS plenty durable system for my QX to work with. This system needs to last til software catches up for multicore, that's all.

[Guest Nick_N]

Then go for the DDR3 1600 product if you would feel more comfortable.I did run some underclocked tests with the 1800 product and was not that impressed other than the fact that they did run the same a 2x2GB @ 1440 1T and MCH6http://forums.avsim.net/user_files/189206.jpgThats also 400FSB and my 2GB's were at 450 so there is a defined increase above even though its not a huge one. I do not know how far you could push that at tRD6 with your processor however if you had to take a 7 it would probably give you some solid room to push past 1600 @ CAS7As you reduce CL on those 1800's MCH becomes harder to maintain 6 and stable. They will also run 6-6-6-15 but not at tRD6.I can not comment on their 1600 product. I have never used it.and yes Sam, we are blazing a trail... blazing would be a good word, so much so I have to play the warrantee replacement game on those 1800 Corsair sticks because I did Tony at OCZ a favor today and checked out a BIOS change for him since my systems do not exhibit the high FSB/2x2GB issuesThat

[Guest Seadog]

Nick - After an initial taste with the Anand article, I've been scrupulously avoiding learning anything about this stuff you guys have been discussing, tRD's, straps and syncs, and what all.I've done that in the earnest and prayerful hope that it won't matter a hill of beans when the regime changes and Nellie and the X58 are king of the hill, my next way station up from Abit IP35 Pro and Q6600 with 4GB of PC 6400.Please tell me whether I'm wisely conserving mass quantities of neurotransmitters.-Doug

[Noel 4,507]

>>>I see absolutely no reason why you won't match what you saw>with pushing the DDR2 to the max, and then someNick, that is all the news I needed to hear. I did call two techs at Corsair, and they confirmed the 1800 2x2gb CAS-8 are samsung chips. They also both took the position that it is always better to run 1:1 divider. I guess this just insures less timing conflicts or something. They also seemed to want to discourage going over 400FSB, saying it "put too much stress on the CPU." I have a post over there about this very topic.Anyway, this was what I asked them: which would probably go better:1. Dominator 2x2GB 1600Mhz 7-7-7, or2. DHX (i think) 2x2GB 1800 8-8-8Their take was: the 1800 Dominators. I asked if they thought they would do 1650Mz @ 7-8-8 or better, and they both though for sure. I think for stability's sake, I might just aim for FSB of 410 or so, 1:1, 1640Mhz, and that will give me CALWI of 8.5, and alot more bandwidth. 1:1 seems like a natural sync-favorable state, but of course that's just a hunch.Anyway, I'm ordering the Doms.

[Guest Nick_N]

Your conserving cash Doug and yes it will be easier to deal with all this with Nellie however there will be a change in how its clocked.. a bit more like AMD works now. I am scheduled to work with the x58 and a ATi 770 variant coming up this summer so it should be an interesting season and as for my burnout... LOL Turns out this old fhart forgot the reset the CMOS after flashing back from Tony's mod'd BIOSDuhhhhhhhhhhhhhTough sticks.. I like em! for 2x2GB they sure take a pounding

[Guest Nick_N]

Well, when I said that I assumed you would be at 450At 400 and 1:1 your latency is not going to be as lowBut if you ordered the 1800 sticks with the intention of clocking them down, then you can always play with the higher end and see how stable it is. Regardless you should see improvement without stress, how much I do not know. I skipped 1600 and went right to 1800+The values when I down-clock tested those Corsairs did not include FS testing but the numbers looked decent none the less.I did mention the reports of the larger memory installs loading down the NB with the 45nm QCPU, which is what they are talking about.. its a hot subject and they are working on a BIOS fix for it to see if it can be corrected in code, one of which I installed today to find out if it would have a negative effect on those not showing the problem.Its not that you would burn anything out... as long as you remain within spec its just a matter of working a few settings however I understand the desire to not want to work BIOS settings you may not understand completely. And remember, I dont care if its DDR2 or DDR3 -... 2x2GB is not going to have a lot of headroom @ 1T. Even 2T is not going to let you clock 150+ over because of the design. Thats the downfall right now. The 1GB sticks are hitting some wild numbers and low CAS. The 2GB's are still a few months out before they start getting into 'zone' values.

[Noel 4,507]

>And remember, I dont care if its DDR2 or DDR3 -... 2x2GB is>not going to have a lot of headroom @ 1T. Even 2T is not going>to let you clock 150+ over because of the design. Thats the>downfall right now. The 1GB sticks are hitting some wild>numbers and low CAS. The 2GB's are still a few months out>before they start getting into 'zone' values.Ok, I told you I was going to quote this:">Whats nice about 1800+ is you dont have to push them like that>and can get a solid latency without 1T but with higher>CPU/speed"Now Nick, what the heck are you talking about, what 'zone' values? If the DOMs do 1640 @ CAS 7, then that's one zone, CALWI of 8.5ns. Bandwidth is decent (everything is relative, and this bandwidth is alot greater than my current 800Mhz memory).I would go to 1800 too, but I think it may be hard on the board, as it must be on your board. If not, why not?

[Noel 4,507]

Interesting stuff from Derek T at corsair:The issue will be the CPU before it is the DRAM. Test with a low multiplier. Set the multiple to 7 and the FSB to 500. Set the DRAM low to isolate it from the MAX FSB of the CPU. That will give you 3.5GHz which you know the CPU is capable of but I doubt it will work because I don't think you will find that the CPU is capable of the 500MHz FSB. Even the dual cores have a hard time to reach such a spot. The 45nm chips are more of an anomaly than any other Intel made CPU. The FSB gives out before the thermals do. In other words, the fail of overclock is not due to great thermal output or voltage need. It has been found that the 45nm needs far less voltage to get where it is going, but far greater VTT and PLL voltages and it has also been shown that if you raise VTT too high, drastic issues of CPU longevity can result, leading to CPU damage. There are new electrical characteristics in the 45nm processing that are just now being documented because they are anomalistic. To put it bluntly, we just do not know the long term results or even the characteristics of the electrical reactions. Suffice it to say, 45nm product is being damaged with low voltages and high FSB's. Triply so with quad 45nm product which leads me to believe that the effect is both cumulative (multi-core additions) and degenerative.I personally would not go to the 500Mhz FSB. I would go to 400 --> 425 --> 435 --> 445 --> 450 and stop there. Keep in mind that 450Mhz will give you 1900Mhz DRAM but even that is not a given. Most are reaching 400, some not. Many are reaching 425 but more are not. Most are not reaching 450 but some are.500? Only under very extreme and short term conditions. Remember, it is not the CPU speed, it is the FSB that is at issue, so:400 X 10 = 4.0Ghz with 1600Mhz on the DRAM400 X 11 = 4.4Ghz with 1600Mhz on the DRAM400 X 12 = 4.8Ghz with 1600Mhz on the DRAM

[Guest Nick_N]

I looked at it like this.. You said you have no intention of upgrading to the N platform when it

[Noel 4,507]

I run a QX9770 @ 500FSB and DDR2 2000 but have>swapped the 2GB for 4GB from time to time and back again. As I>pointed out yesterday the 45nm voltage issue can be dangerous>and on 4GB+ VTT and PLL may need to be adjusted, and, you can>not exceed a certain level of Vcore either so that is why I>passed on the info about the problem they are seeing with the>NB and Q45nm processors.I'm tellin' you, I am not running vCore over 1.40. And max temps on the CPU is 62C, rarely 64C, and usually 57C at full load. It creeps into the low 60's if I forget to turn up the AC enough. So, you're saying it IS the NB. Ok, maybe the X48 will handle that FSB better. Still, there is controversy on what role higher FSB has on the CPU when vCore and temp are fine.>the issue of stress is not the board, it's the Quad processor>and mainly large memory use but none the less any quad can be>a bit finicky and the 45's are definitely worse than the 65's>in that respect.>>I switch to a dual core from time to time for high FSB testing>and run 535-565 on that board on 2 or 4GB>>>450 may need to be trimmed in on the proc however I never saw>any details around problems hitting that with 4GB+ until a few>days ago. Here, the auto settings work just fine at 450 but I>do have to stabilize above that with a quad. I just don't have>to go as far as Tony posted. Thats also a 1600FSB processor I>am running which we had surmised was why I do not see the same>problem with stability others do. >>>>>Regardless, I am sure you won

[Guest Nick_N]

Still some confusionI just helped Anthony Voz with his new system. He

[Noel 4,507]

>Still some confusion>>I just helped Anthony Voz with his new system. He

[Guest Nick_N]

Go back up and read the links I posted NoelYou dont blow cold air into a tower from a source that is not designed for IT use and even then there are steps which are taken to ensure damage does not occur=======================================Today's technology rooms require precise, stable environments in order for sensitive electronics to operate optimally. Standard comfort air conditioning is ill suited for technology rooms, leading to system shutdowns and component failures. Because precision air conditioning maintains temperature and humidity within a very narrow range, it provides the environmental stability required by sensitive electronic equipment, allowing your business to avoid expensive downtime.======================================you could never get yours past a certain point or to 450FSB according to your posts so something is wrong and always was, unless you never tried it at the formula values I posted