RAM type/speed for new system

[EasyPC 0]

Hi, I,m just about ready to get my credit card out for my new system which will consist of :-Core2Duo E8400 (2x3.0GHz)Asus P5K MotherboardVista H.P. 64bit750GB SATA2 HD for O/S and other junk300GB WD VelociRaptor HD for flight sims512MB GeForce 8800GTSX-Fi Xtreme gamer sound card600w P/SThe question I would like to ask is, should I go for 4GB of Corsair Dominator 1066MHz RAM, or 8GB of Corsair XMS2 800MHz which I can get for around the same price. Any comments regarding other components of this setup would be welcome too.Many thanks.Pete.

[jackcnd 0]

No comment on the ram, but if you can wait a month the new nVidia video cards are due.They have very large 512/448 bit memory bus which is key for FSX.The 8800GTS 512 has only 256. Search for GTX280 and GTX260 info. Something to consider if you can wait, it will make a much bigger difference than ram type.

[Guest D17S]

If you are not going beyond a 400Mhz FSB, go for 4 gigs of DDR2-800. The difference between 800 and 1066 ram speeds is entirely indistinguishable. Same deal with the HDs. Drop the Raptor and go for a 2nd 750 for backup. Ditch the addon sound card and use the mobo's onboard sound. A person would need dog-ears to tell the difference! Bump the Mobo up to an X38. All you really want over a P35 is the PCIe-v2 feature. Vcard performance is about to double. Bandwidth issues may come into play sooner than later. A lab assessment - would - (almost) certainly be able to distinguish performance increases that will occur by using the Raptor, DDR2-1066 and an (offloading) addon sound card, but a gamer never will. VelociRaptors, X-Fi Xtreme, X48, DDR2-1066 ram? Sadly, this is all just exuberant marketing hyperbole aimed at an under-informed audience. They get to use cool dinosaur names and lots of Xs. I second the earlier motion. Save those bucks for something that could actually provide a performance gain, that new GTX280 Vcard.

[sargeski 16]

If you are building the computer for FSX I would not be so quick to drop the soundcard. If you would like to have FSX sounds on one device and ATC sounds on another you will need two devices. THat could be your onboard sound plus an additional soundcard or your onboard sound and a USB headset. Just something to think about as it really increases the immersion factor within FSX to have sound coming out of your speakers and ATC only coming over a headset.

[EasyPC 0]

Many thanks for the advice, greatly appreciated. Obviously a bit of a re-think may be necessary. Maybe I will bit longer for the next generation video cards but there again I've been waiting for months already waiting until I could afford this system and for things to settle down a little with FSX and Vista, but as as been said in these forums many times you could wait for ever.As regards the soundcard I really do need this for my musical interests where I need soundfont support.Once again thanks for taking the time to reply.Pete.

[McCrash 3]

>If you are building the computer for FSX I would not be so>quick to drop the soundcard. If you would like to have FSX>sounds on one device and ATC sounds on another you will need>two devices. THat could be your onboard sound plus an>additional soundcard or your onboard sound and a USB headset. >Just something to think about as it really increases the>immersion factor within FSX to have sound coming out of your>speakers and ATC only coming over a headset.sargeski, did you have any problems mounting that Zalman onto that motherboard? Did you use that tension clip?thanks.

[jackcnd 0]

If waiting is not ideal, then check into eVGA video cardThey offer a 90 day step up program, so if the GTX280 does come out as expected you will at least have an option to upgrade,but still have a system now.Note Sam's comments on the x38 MB offering PCIe v2 that may be needed for best performance from the GTX280 video card.

[sargeski 16]

>sargeski, did you have any problems mounting that Zalman onto>that motherboard? Did you use that tension clip?>>thanks.McCrash,Good morning!! No issues were experienced at all installing the Zalman fan on this motherboard and yes, I did use the tension clip.

[ryanbatc 12,867]

I'd stick with the X-fi, or Auzentech I think it's called. I must have good hearing, cause I can tell the difference, especially in games like Battlfield, Half life 2, command and conquer etc. If you only play FSX then yeah, might as well lose the sound card.I get 4GB of RAM. Instead of the raptor, I'd look into Seagate's latest 32mb cache hdd's.Will you be Overclocking this system? If you get a decent cooler like the tuniq tower or thermalright ultra 120

[Guest Nick_N]

Rock my worldhttp://www.clockmehigh.com/TestingAtCorsai...ingleHyperPi32MI can finally announce that Corsair has broken the DDR3 2280 barrier @ 1T CMD with a 41.2LtcyBeen running those sticks @ CAS 7 DDR3 2000 1T for almost a month.. soon to be available if you have a DDR3 system but you have to have the cash too

[Guest D17S]

That's some quick stuff and brings to mind an interesting pozer. I'm curious that if the present day FSB (Off-board memory controller) architecture can take full advantage of the capabilities of these memory speeds, what (really) will be the advantage of Intel's next-gen onboard memory controllers. As it seems, there Is no bottleneck, yet.The OP probably already knows that in less that 6 months, this current generation of FSB based computer systems will become obsolete. Welcome Nehalem. But on the other hand, consider: I'm courteous to see the results of the memory subsystem with this DDR2000/T1 stuff at a 250Mhz FSB at 4:1. When do these current front side buss (FSB) bottlenecks actually start to appear? Said another way: How much headroom does this current FSB platform really have left . . . and how many - Decades - worth of future-proofed capability will Intel (soon) be telling us we just gotta have, Right Now?

[Guest Nick_N]

I would also wait for Nellie if cash is an issue... but I would make that move for sure none the lessSam's point and suggestion about the platforms is correct. there is a bottleneck on the northbridge in relation to FSB even with DDR2 and Intel. Be that as it may, going faster does provide a better platform however the higher you go, lower latency becomes more and more important in order to compensate for the loss and make the DDR3 purchase worth the investment. That will no longer be the case with Nellie.As for the suggestion of how far and efficiency within the limits of the current platform design.. my estimates place that at right around CAS4 or 5 @ DDR3 1800 so DDR3 2200 @ CAS 9 is closing in on the same barrier but not quite there. Right about the time that point is reached in memory technology the platform to unleash the DDR3 ability and remove the bottleneck will be on the market.For a pro a 60+% TRUE increase over today is quite possible by the second iteration of Nehalem as it matureswhich by the way, the first iteration of Nehalem will be unleashing the ability to run a 1M test of SuperPi in just about 8 seconds without any clocking on standard DDR3 1333 memory. If you dont know anything about SuperPI... that 1M test @ 8.32 seconds on a UNCLOCKED X58 Nellie @ 1333 is INSANE to try and get with even the best LN2 system on 5+GHz so that should give you an idea of just how much you are truly losing with Intel and DDR2/DDR3 today.

[Noel 4,507]

>That's some quick stuff and brings to mind an interesting>pozer. I'm curious that if the present day FSB (Off-board>memory controller) architecture can take full advantage of the>capabilities of these memory speeds, what (really) will be the>advantage of Intel's next-gen onboard memory controllers. As>it seems, there Is no bottleneck, yet.>>The OP probably already knows that in less that 6 months, this>current generation of FSB based computer systems will become>obsolete. Welcome Nehalem. But on the other hand, consider: But Sam, what does Phenom offer over C2D in terms of meaningful performance effects in FSX? There is obviously more the the story than memory controller latency, and it seems you are again seeming to focus mostly on bandwidth (bottleneck) over wait times (latency). Isn't it about both, or what did I miss?

[Guest D17S]

It appears the initiating premise has been this: "Ram speed MusT make a difference. Since it's not, it must therefore be the buss that is holding it back." That has never been my position. My point has always been there is No buss bottleneck. Ram speed just makes no (little) difference. I guess I should consider this a big mystery. However thankfully, an event that I simply observe and have (intentionally) not taken the time to fully understand really doesn't confuse me. (That's gotta be some sort of defense mechanism, for sure!) The current FSB system is capable of distributing whatever "current" (bandwidth output) their little "generators" (latency driven efficiencies) can produce. That's why I'm questioning the need for Quickpath. Nick showed the potential of lower latency's productive capability, I was just surprised (even disappointed) it took So much to get sooo little. I was kinda hoping faster ram might actually be able to start providing some productive assistance, rather than just being a necessary passenger in the CPU's 1st class cabin. Oh well. It appears to me Nehalem / Quickpath will proper exactly the same strategy that as was presented with SATA II's introduction. SATAII can transfer at 375MB/s, but its (then) user HDs could only transfer at 50MB/s. We're up to 100MB now, but still only 1/3 of the way there. That's why I went with that 3 drive raid. Bandwidth is a terrible thing to waste and I can transfer at 300MB/s. But my results would rival Nicks for verifiability! It IS better, but really only results a mother could love. What can I say. I wanted 750G of storage and the three 250s only cost a couple of bucks more than a singe 750 (back then). Cheap toy. Yet SATAII's goal has Always been to intentionally mislead a lay customer base to believe that a single SATAII drive is faster than any SATA I drive (or a PATA-66 for that matter) because it can run on that buss. This all gets quite annoying.It appears they are doing the same thing with this Quickpath stuff. It will be needed eventually, but not before they misled millions of customers that it will provide a huge performance advantage. As far as I can tell, the Core2 was it. Nehalem will simply be the Penryn schleped onto a totally useless Quickpath system architecture. It Will be necessary someday, but probably not until well into the next socket "upgrade" (forced re-buy) or 2. Skip the 1st-gen Nehalens, at least. The future will still be about numbers of cores. A bump in core speed from 3.6 to 4.0 become entirely irrelevant when Nehalem's shrink goes to 16-24 physical cores (HT'd to 32-48). . . . and the software guys will finally "get it." Memory subsystems will remain the last frontier for the serious hobbyist that will still only provide that last, sweet drop of O/Cing satisfaction (but not much else). There on the other hand, remain hopeful. I could be wrong.

[Manny 1,293]

>> "The current FSB system is capable of distributing whatever "current" (bandwidth output) their little "generators" (latency driven efficiencies) can produce. That's why I'm questioning the need for Quickpath."Sam,Currently when I OC, the FSB does come into play. Its quite often a weak link in my OC chain. Eliminating this link from the chain is a positive benefit thats realizable today for OCers don't you think?Manny