:-spacecraft Ooops, maybe this is a not so good thread to follow.Byeeeee:-eek

>Its harder to overclock with 2GB Modules so instead of>getting a 4.7GHz overclock you will only reach 4.5.Please expand: Are you advocating 4 single gig modules?Or are you advocating 2 2gig modules?RhettAMD 3700+ (@2585 mhz), eVGA 7800GT 256 (Guru3D 93.71), ASUS A8N-E, PC Power 510 SLI, 2gb Corsair XMS 3-3-3-8 (1T), WD 150 gig 10000rpm Raptor, WD 250gig 7200rpm SATA2, Seagate 120gb 5400 rpm external HD, CoolerMaster Praetorian

Rhett,He's advocating 4 1GB modules.

>Hi john i mean i tried out fsx on 4.7GHz and got 30FPS on my>5.7GHz i got 45FPSWow, I think he found the sweetspot! For a 21% increase in clockspeed, he gets a 50% increase in FPS! Oy! The rules have changed!

I agree. That's one sweet spot. But the rules haven't really changed. Here they are: This is the Yorkfield's world record page: http://www.ripping.org/database.php?coreid=YorkfieldNote the absolute world record it 6.1Ghz. It takes Liquid Nitrogen to get even to 5.7Ghz. What's that? Here's an old Tom's Hardware example: http://www.tomshardware.com/2003/12/30/5_ghz_project/You actually pour LN2 (Liquid Nitrogen) onto the CPU. It only allows a single benchmark (or until you run out of LN2) and is for the serious hobbiest only. This is EXTREMELY dangerous and can cause SerIouS personal injury in a moment. You are dealing with temperatures of -170C. Forget about it.The first "Cascade" runs shows up at 5.7. What's that? Here they are. I like the 'Lill' Red Wagon.' You can tow it behind your bicycle.http://www.under-the-ice.com/forums/viewtopic.php?f=26&t=625.7Ghz will take either LN2 or a Cascade. Still, these temperatures are below -120C. These temperatures will permanently destroy even the toughest hide in a microsecond. Less lethal water cooling run starts at 5.0ghz.Any of these setups will:1) Cost several thousand dollars, 2) Provide only short runs at these speeds and 3) Subject the user to a very real possibility of serious personal injury. These "rigs" all generate temperatures below -100C.Please folks, be real careful with this stuff. It's dangerous and has nothing to do with increasing FPSs or flight simming. It's a whole different hobby.

$3,731.82On Tiger Direct!

>I agree. That's one sweet spot. Yer kidding, I was just . . . kidding!If this is true, you can o'c clockspeed 21% get 150% fps increase, what the heck is that all about? That is completely discordant with every other testimony I've ever read. Who knows, but it sure isn't compatible with other scaling.

Yes everyone i do use a -130 degree cascade to get to 5.7GHz. Also i dont think i got my performance increase by percentage increase in clock speed but actually how much of an increase of speed there was. (1GHz). What do you mean by personal injury. I got someone else to build and connect the cascade and such. I dont get into the "I b uild ice machines, im so great i dont get hurt or anything" thing.

Yea, I was kidding too. The disconnect seemed to me also, entirely self-evident. Also, ANY piece of equipment that operates at those temperatures also operates at Significant and therefore - dischargable - pressures. This thing is Locked (pressurized), Loaded (-100C) and belongs kept somewhere out of harms way. Lock it up in a gun cabinet. Keep it away from those you care about. Like a Ferrari at 170MPH, everything will be OK as long as the wheels stay on. Like the Ferrari, if that thing blows it's cork the result - most certainly could be - instantaneous and dramatic. Say goodnight Irine. Folks, these are not toys in ANY sense. They are serious pieces of equipment to be used by serious hobbiests that completely understand - and respect - their functionality. PLEASE be careful with this stuff.

>Rhett,>>He's advocating 4 1GB modules.Is that really a good idea? The mb usually only have 4 memory slots and I have heard that you should never fill all memory slots since you will lose dual channel and the memory frequency will be lowered.Also with only two modules you can have all modules matched pair. With four modules you don't get four matched modules. Also with all slots filled with 1 Gb modules you can never add more to get >4 Gb.

Aye, methinks this thread is infested with bovine scatology.

>Yea, I was kidding too. The disconnect seemed to me also,>entirely self-evident. Also, >>ANY piece of equipment that operates at those temperatures>also operates at Significant and therefore - dischargable ->pressures. This thing is Locked (pressurized), Loaded (-100C)>and belongs kept somewhere out of harms way. Lock it up in a>gun cabinet. Keep it away from those you care about. >>Like a Ferrari at 170MPH, everything will be OK as long as the>wheels stay on. Like the Ferrari, if that thing blows it's>cork the result - most certainly could be - instantaneous>and dramatic. Say goodnight Irine. >>Folks, these are not toys in ANY sense. They are serious>pieces of equipment to be used by serious hobbiests that>completely understand - and respect - their functionality.>PLEASE be careful with this stuff.Sam, I love your prose! Hey, I'm going crazy on this decision. Tell me again what you recommend for one of the cheaper Q9450 or 9550s as far as oclockability goes. Which one to you predict will oc better, and with what memory.I almost pulled the trigger this morning on the QX9650 at $1109. That is an insane amount of $$, tho I can afford it right now np. If it's a case of waiting for 2-3 mo for a Q9x and the real price diff will be perhaps around $550 more for the QX, perhaps I should go for it. I understand some steppings or individual QX's may not perform as well as people might hope. I would like to run my sys at very close to 4GHz, and I use air conditioning enhanced air cooling only. I know many are able to do this on minor voltage tweaks, but yikes I'd hate to blow out a core at these prices!Noel

To build a more perfect (. . . at least less lethal) system you Can fill all 4 ram slots. You will not loose dual channel mode. You need to use matched pairs in each set. However the sets do not have to match. Two matched 1 gig stocks and two matched 2 gig sticks will get six gigs and work just fine in dual channel mode (you're going for a 64 op system, right?!) I currently have two sets of (separately) matched pairs of 1 gig sticks w/ my 64bit Vista. With high sliders, I'm constantly into that 4th gig. Wish I'd bought 2 sets of TWO gig sticks. The stuff is soooo cheap right now. Might as well get it. On the Q9 front, 9450 has a 8x multiplier. The entire question then becomes "how far will the selected FSB go on the selected mobo?" The FSB is now the limiting factor, Not the CPU.To get 4.0, the FSB has to get to 500mhz (8 x 500 = 4.0). That happens all the time in the O/Cing forums, but I would not recommend it as a 100% stable, 24/7, 'set and forget' FSB speed. 500mhz is right on the edge of a standard cooling setup. Like any other chip, you can hit the Northbridge with LN2 and get it to the moon, but that might not be what we're after. The P35 (and X38/X48) will run at 400 all day with a bios reset and leave the rest to Otto (Auto). That's where I'm at. 400mhz is fine. I've played at dropping the CPU's multi and raising the FSB some more. 'Puter wasn't happy. I'm sure if I figured out all those bios settings, I too could get to a stable 450, but I'm really not That interested. There is No performance benefit from a higher FSB and Otto's doing a fine job at 400. But let's be real optimistic and say Intel tweaked the X38/48 so Otto could handle a 450mhz FSB (500 will still be for the pros). 450 x 8= 360. Opps. Best case (for us normal folks), the Q9450 will goto 3.6Ghz. Remember too, the ram will be running at 900mhz. DDR2-800 will not (reliably) run there. You gotta pony up for DDR2-1066. Might as well get the Q6600. It's 9 multiplier X that 400 FSB = a 3.6Ghz CPU with DDR2-800.This is no accident. Intel knows how to do arithmetic too. It's their CPU and their chipset. The QX at 1100 bucks is what it takes to get that 400-600 mhz bump over the Q66 without insane cooling. Is it worth it? Of course not. It's ridiculous. But there on the other hand, . . . If not now, when?