E6850 or Q6600?

[lmponte 0]

That's the question, which is better? I know the 6850 is probably faster on fsx and the majority of apps and games right now, but I am worried that might change in the future as applications are written to take advantage of the 4 cores (FSX SP2 maybe?).What do you think?

[vgbaron 3,621]

Well out of the box, the E6850 is faster but if you overclock it the q6600 is a better choice. Also, the quad is still closer to the cutting edge than the duo. As more and more software is written to take advantage of the quad, you will be better off in the long run.The only reason for buying a duo was the price and since they are equal, IMHO, it's a no brainer.VicQ6600 G0 CPU 2.4 o/c 3.6Evga 680i A1 with P30 BIOS 2G XP2-8500 DDR2 1066FSB Mushkin 996535 RAM 5-5-4-12-2T320G 7200 HD partitioined for XP/Vista/Programs 2 - 74G Raptors in RAID0 500G 7200 HD for backup SATA DVD burner Evga 8800GTS 640 PCIx Kandalf LCS case w/ built in liquid cooling 850W Thermaltake power supplyVisit the Virtual Pilot's Centerwww.flightadventures.comRadar Contact Supporter: http://www.jdtllc.com/

[lmponte 0]

Ok, now the tricky million dollar question: I have never overclocked a single thing in my life. I know you will tell me to read a lot of articles on the web and that's what I am doing. But is it easy enough so that a beginner like me can overclock a processor without problems? I think I've read all about it, and I think I have a pretty nice idea of how it's done, but I am fearful about unexpected surprises...Luis

[boris4356 4]

I bought an E6850 thinking I would be able to overclock it into the 4 gig range, but I have only been able to get it to 3.7 stable on air with a tunig tower. I am a bit disappointed with this, as I have seen a lot of people getting the Q6600 to 3.6 on air. If I were to do it all over again, I would go with the Q6600. It is a crap shoot though as to how high you can overclock. It varies with each chip. It is the luck of the draw.Overclocking is pretty simple, but you have to make sure you read up on it, as there are a lot of things you should know, and different tests to check stability.I am pretty happy with the performance I am getting in FSX with my E6850 despite the lower overclock. I have most sliders to the right, and FPS locked to 30. It runs pretty smooth, and rarely dips below the 30 mark. Bill

[Guest D17S]

A builder might want to consider their own upgrade roadmap. Intel has one. Why not us? Certainly, this will not be a final build. The question then becomes, 'What hardware matches current -and projected - FS versions and will last far enough to allow significant hardware improvements to evolve and become available for the next upgrade.' It appears the reality of the (now apparent) multi-core future became clear too late in FSX's development. By then, the FSX team may have been essentially locked into a single core engine. The big dog, still, will be getting those 3rd and 4th cores fully on board. Sadly, it really seems that's not gonna happen 'till FS11. You may have heard Phil addressing performance concerns with FSX with the comment, 'We gotta put FSX to bed so we can get focused on FX11.' Quad or dual for optimal FSX performance now? The dual will O/C to 3.7ghz. The quad will O/C to 3.6ghz. If anyone can measure a FPS difference between 3.6 and 3.7, they get a prize! There IS no difference in performance. Wait for the Penryn? Getting the quad now could be a bit of economic future-proofing. Penryn or Kentsfield, there's not much FSX advantage, Remember the quad's pricing cycle? Like the Kentsfield quad, the initial Penryn quad will be $1000. Also like the Kentsfield, it might take the Penryn quad a full year to get back down to $275. And all that waiting will net a 20% FSX single/quasi-dual core advantage, at best.Considering FX11 will be fully multi-core optimized, at that point the quad will be a must-have. If a person is building a system now anyway, a prudent financial choice might be to just go ahead and get the quad now. That way it will already be sitting in your socket when FX11 hits. Otherwise, buying a dual now and a quad 2 years from now is just throwing away the price of the dual.This kind of forecasting is becoming possible because Intel's CPU road map is becoming extremely predictable. Intel wants us to buy Much more often, but it seems to me a user can stay quite up-to-date with a 3 year upgrade cycle. Skip the Penryn and Nethalem cycles entirely. Intel doesn't like to hear this, but the next upgrade we really might need could be Intel's 32nm Sandy Beach in 2010. Hopefully, once again (as it was in 2007), this will be an overkill for FS11 and a 'future-proof' for the expected requirements of FS12. Tick, Tock.

[Guest ChrisE]

I wonder if FSXI will be "fully multi core optimised"? I guess that there is an enormous amount of legacy code in FSX & it would all need to be substantally rewritten to achieve "fully optimised"I do so hope that you are right & that the resourses are there to fully sort out the code for FSXI - not just the multi core issues - but I wonder if cost & lead time will mean the same sim with a few more headline features. Chris

[vgbaron 3,621]

>Ok, now the tricky million dollar question: I have never>overclocked a single thing in my life. I know you will tell me>to read a lot of articles on the web and that's what I am>doing. But is it easy enough so that a beginner like me can>overclock a processor without problems? I think I've read all>about it, and I think I have a pretty nice idea of how it's>done, but I am fearful about unexpected surprises...>>LuisAlthough I've been bujilding computers for over 20 years, I also have never overclocked a thing. It really is simple and as long as you move in small increments you really can't hurt anything.I could have gone higher with my quad but 3.6 was my goal so I stopped there.As someone said, it all depends on the CPU and the MoBo combination but there are plenty of options.I wouldn't let the lack of O/C experience hold you back but remember, small steps and test and repeat. May take a few weeks to get it where you want depending on your free time.Good luckQ6600 G0 CPU 2.4 o/c 3.6Evga 680i A1 with P30 BIOS 2G XP2-8500 DDR2 1066FSB Mushkin 996535 RAM 5-5-4-12-2T320G 7200 HD partitioined for XP/Vista/Programs 2 - 74G Raptors in RAID0 500G 7200 HD for backup SATA DVD burner Evga 8800GTS 640 PCIx Kandalf LCS case w/ built in liquid cooling 850W Thermaltake power supplyVisit the Virtual Pilot's Centerwww.flightadventures.comRadar Contact Supporter: http://www.jdtllc.com/

[lmponte 0]

There's one thing I am little worried about: voltages. How do I know when it should be increased? I know about increasing the FSB, and about the relation with the RAM speed (that's why I am going for good memory), but I'm not so sure about voltages, as they have no direct relationship with CPU speed... Any tips?

[jackcnd 0]

I increased the FSB until I got errors running prime 95. Then increased the voltage within the processor max voltage on listed Intel.In my case I decided running lower voltage at 3.2Ghz made more sense than a high heat producing voltage just to get it to 3.5. Thats the key watch the CPU heat as you run P95.Instead of finding the max clock speed and needed voltage I tried to find the min voltage to run at 3.2GHz. System has been rock solid since and runs cool with a 120 Extreme on the CPU.I found a Q6700 overclocked much easier than a Q6750, as the default FSB is lower on the Q6700. For FSX once you hit 3GHz your fine, going to 3.6 really does not make much diff in the game but it does on your system heat.

[Mace 1,926]

>>Wait for the Penryn? Getting the quad now could be a bit of>economic future-proofing. Penryn or Kentsfield, there's not>much FSX advantage, Remember the quad's pricing cycle? Like>the Kentsfield quad, the initial Penryn quad will be $1000.>Also like the Kentsfield, it might take the Penryn quad a full>year to get back down to $275. And all that waiting will net a>20% FSX single/quasi-dual core advantage, at best.>Sam, I think you're referring Intel's "enthusiast" edition, and $1000 is where they have been pricing those over the last few years.But don't you think the $400-$500 range is more likely initial price for the "normal" Penryn? Don't you think? And then it will drop down to the $300-400 range after production ramps up? ...if the past is any guide...this what I see happening.What have you seen to conclude $1000 for a mainstream (non extreme edition) Penryn/Kents?>Skip the Penryn and Nethalem cycles entirely. Intel doesn't>like to hear this, but the next upgrade we really might need>could be Intel's 32nm Sandy Beach in 2010. Hopefully, once>again (as it was in 2007), this will be an overkill for FS11>and a 'future-proof' for the expected requirements of FS12.>Tick, Tock.So do you think, that if a person gets a G0 Q6600 right now, that they will be in a position to skip Penryn and Nehalem entirely, with respect to FS?After all...a 10-20% perf improvement is STILL a 10-20% perf improvement...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

[lmponte 0]

OK, sounds easy enough for me. My goal is also 3.2GHz. and I have been thinking about the 120 Extreme cooler too... do you recommend it? Does it require any fan, or is it completely passive (it looks like it from the pictures)?Thanks for your help!

[jackcnd 0]

>OK, sounds easy enough for me. My goal is also 3.2GHz. and I>have been thinking about the 120 Extreme cooler too... do you>recommend it? Does it require any fan, or is it completely>passive (it looks like it from the pictures)?>>Thanks for your help!I love the Extreme, took my temps down a lot.I use the Scythe 1600 rpm fan on it, and two 1200rpm intake fans on the front of the case, one tri speed fan on the back, and one in the PS. Lots of fans. The quads run hot. If your going with a duo 6850 then I'd think you could drop to 1200 rpm on the Exteme and a single intake and outtake fan.

[boris4356 4]

The extreme 120 does not come with a fan, and you do really need to have a fan when you overclock. That is why I bought a Tuniq Tower 120, because it was highly rated, with only the extreme 120 barely beating it out in most reviews. The Tuniq comes with a fan, and is cheaper then the extreme 120.Bill

[Guest D17S]

It's interesting to note that the first 775 quad Penryns will be their enthusiast additions running at top clock speeds. Remember, these are all Identical chips. Intel 'bin sorts' these things like Lucy and Ethel sorting chocolates on that assembly line. The reason we are getting such massive O/Cs is that there are no bad ones! Intel has this down to a science. It's called 'yield." We now have the GO/XX50 series. The best of this series might get another 100mhz on a maxed out O/C . . . but they are all ReaL close. Intel needs SKUs for the various market segment, so they just stamp a price point (6550/6750/etc) on em and send them out the door. They are all the identical CPUs.The Penryn is just the next refinement to the GO/XX50 Core2 CPU. They are leading with a 3.0 (3.2?) clocked CPU. Heck, that's nothing special for us, but Intel needs this to run even on an ECS board! Once again, all these Penryns are identical chips coming off the same production line. Clocking their lead offering at that enthusiast speed it telling me the Entire Run will run will be enthusiast speed capable. They will ALL be that good. Intel will need SKUs for us po folks so they will just stamp some with a lower rating. Intel will burn in lower multilplers so they will default to slower speeds. We will buy these cheaper CPUs and just turn them back up. This is the marketing sequence we saw with the initial Quad Kentfield CPU. It took a 8-10 months for that origional Kentsfield to get marked down from $1000 to $275 . . . Stepping updates will come and go, but the Q6600 IS the QX6800/6850, just with a lower price tag. My guess is that Intel will use this same marketing strategy with the Penryn. To the non-geek world, this is a whole new CPU, but we know better.As you say, 10% is 10%. However my strategy has been to wait for a 10X increase. It's more a matter of how much does a user want to spend on commuter equipment. Intel and I really battle it out. They say "Spend more!" I want to spend less. I really think the Q6600 will let a user bypass Penryn. The Nehalem will have 8 cores, so that will be getting a bit tempting. That would get me a double(++) from my Q6600 . . .but still, just a double. We'll have to see how well FS11 optimizes for multi-cores. If my quad is still getting the job done in FS11, the "GO" and "Penryn" updates to the Nehalem will not make much difference. Remember even the Penryn is just 10%-20% faster than the original Core2 14 months ago. So it will go with Nehalem and its updates.The drop to 45nm will allow Nehalem to go to 8 cores. I expect the next drop to 32nm will allow at least 16+ cores. Intel is also talking about having room for their much vaunted Larabee (Intel's discrete Vchip project) to be integrated into that same chip. I say leave the Vchip out and gimee more cores. There will be lots of room. Remember I'm after a 10X? I'll be at 4 cores then. 16 cores is a slam-dunk. Ditch the vchip and 20-30 cores (40? That's 10X and I'm home!) on the Sandy Bridge is entirely possible. 2010 . . . That's 3 years.I think it's entirely possible for a build right now (sans Vcard) to last 3 years . . . but its gotta have the quad.