Nick, thanks for your thoughts! You are right 250 more on a CPU means alot more time to do what the sim was made for .... FLYING. The xtra on the mem. is no big deal.So that being said, what VC? Here are my choices within my budget:9800 GTX9800 GTX Plus9800 GTX KOGTX 260I am leaning toward the 260, what are your thoughts?Thanks!

>Hi JK.>>>The 45nm Penryns have been shown to be around 10% faster>clock-for-clock than the 65nm Kentsfields. Go do your>research.<>>Interesting remark. Can you go into some details on what you>base your statement? TV>For starters, just google 45nm vs 65nm performance. The concensus from many different sites is around 7% to 8% faster clock for clock. OK, 10% might be generous.The 45nm process is smaller (that's just physics), less power, less heat, more L2 cache.My point is that the 45nm and 65nm are NOT the same in terms of performance, whatever the difference the 45nm is just plain better.And the 45nm E0 stepping is a masterpiece. Even a E0 Q9550 will get over 4GHz on the right board and kick the snot out of a Q6600 at any clock.-jk

260 or 280

>Hi JK.>>>The 45nm Penryns have been shown to be around 10% faster>clock-for-clock than the 65nm Kentsfields. Go do your>research.<>>Interesting remark. Can you go into some details on what you>base your statement? TV>=========================================Intel

Hi, Nick.Thank you, for trying to explain. It's been a few years since I've actually designed things, and not with the newer CPUs / architecture, and I understand the logic / explanation with regard to " reason a 2.2GHz Core2 will clean the clock of a 3GHz AMD Sandy in FSX", but I am having trouble with the statement when it comes to 45nm Penryns vs 65nm Kentsfields. The only difference, correct me if I am wrong, is the additional set of Instructions set, which MSFS does not use, and mostly nobody, and the Cache, who's value is negligible. The rest is the same when it comes to external interface. Am I wrong? If I am not wrong I have trouble self explaining the difference. Other than the above mentioned Real differences, and the fact that they can put more Transistors, and lower power dissipation, disregarding the Marketing verbiage, what am I missing? When it comes to MSFS direct benefits. Just as an additional bit of info I've actually experimented, with the above combinations, clock per clock, and None of the existing BenchMarks showed any performance difference between them. Given that fact, if someone can save a few $$, assuming that they can do that, why get the more expensive set? I know that there are people out there making that statement, mine is faster than yours, but where is the beef?My advice would also be to go with more efficient / lower power, but that has nothing to do with what we are discussing. TV

The beef is simpleIf you dont understand the brief and the impact to performance through that explanation... which most wont unless they are an EEI have run the 96 and 97 procs and they will clobber a Q6600 I can confirm the difference in FSXNeed real world beef from another source?Look here:http://www.simviation.com/cgi-bin/yabb2/Ya...?num=1223693603Started out on Q6600, ended up on Q9650 with the SAME supporting partsNuff saidIf you need more beef look it up because there are plenty of engineering sites that will confirm the increase over 65nmjk is right

Congrats on your 1000th post!:-beerchug

" ... but where is the beef?"That's been my observation too. Any clock for clock difference between the 45nms (Penryn) and the 65nm Core2s in FS performance is only in the eye of the beholder. There's no objective difference. For the FS purposes of This forum, Super-Pie benchmark runs and/or EE analysis are very interesting and certainly welcome, but entirely irrelevant.The Penryn's real performance claim to fame is that it Will clock faster. Any of the 45nms will cruise to 4.0Ghz on air. But there's a big "But." Intel rescinded the free lunch we had been enjoying with the 65nm chips. With the Penryn generation, Intel increased the FSB speed. The increased FSB speed was 'hung out there' to induce a lay audience to believe that faster was better. However tacit, this was a misrepresentation. In reality, FSB speed increases provide No performance advantage. In the Penryn's case, the only reasons the FSB was increased was so they could Reduce the multiplier and require more expensive (and equally useless) faster rated ram. The power reduction will save a dollar a month. This is significant to a server farm, but not to a single user. We really got hosed on this one.How can a reduced multiplier increase sales revenue? For we overclockers, the Penryn's 4.0+ overclock must now be achieved with a faster FSB. 400Mhz is the P35/38/45/48's rated speed. In all cases, the highest fixed multiplier is ~ 9X. The Q9650 and the Q6600 share this multiplier. At the current Intel chip set's rated speed of 400Mhz, they Both run at 3.6 . . . and the performance will be indistinguishable.So, if a user wants 4.0 from the Penryn, they gotta pay up. The QX CPU has an unlocked multiplier . . . but is offered at a huge price premium. By paying Intel's price user can get that 4.0+ by running a stable 400Mhz FSB and simply increasing the CPU's multiplier. But those QXs are costy. The other route is to get the cheaper Q9 and - take a chance - by increasing the FSB beyond its rated 400Mhz. In a nod to Intel's ram constituents, a user also Must run faster rated ram or chance running slower rated ram beyond its rated speed. All that just adds expense (i.e., revenue) . . . but can get costy in both terms of both smoke and stability issues. Intel knows this. They know any savvy O/Cer will pay up for the more expensive unlocked QX. So much for our free lunch. It's gone. The Nehalem will present the same set of decisions. So far, the question is how far will that 133Mhz FSB (BCK) go before it explodes? That will tell how far we will be able to drive the cheaper low multiplier'd i7s. The more expensive i7s will have either higher or unlocked multipliers (Just like the core2s and the Penryns). As before, this will be the Only difference between the i7 SKUs. The other issue is this dumb triple channel ram. Now we have to buy ram 3 sticks at a time. Maybe this bandwidth will be necessary eventually, but not yet. The conversation is not about outright performance. The Penryn will go faster than the older Core2 . . . if the user is willing to pay the price in either dollars and/or system stability. This is Only a cost/benefit analysis. As always, "Speed costs. How fast do you want to go?"

Beef?He was so impressed he threw down the money on DDR3 and replaced the DDR2 system from the first thread moving the Q9650 to that platform because of the difference he saw going from Q66 to Q96http://www.simviation.com/cgi-bin/yabb2/Ya...=1225569427/120wow.. money talks, BS walks and stays on a P35, hardly stable at 400MHz and swears they know all about the difference and its only 10% for thousands of dollarsrubbisha few hundred, not thousandstheres the real beefthat and screen shot compares fo a Q66 to a Q9 clocked righthttp://forums.avsim.net/dcboard.php?az=sho..._id=45716&page=wOW!thats a really well defined 10% differencesome just continue to post more BS cause it reads good to them and they want to stir the pot to get people like j/k and myself to post back... that what its all about to themI dont know of any other way to show you the beef Avcomware. I can go into engineering of it.. but how do you transfer that to FSX and real world results? I think what I posted in small sample should present enough beef to digest I can post thread after thread but you should see the beef now :) @ j/k... let this and future stuff like it go please and dont give it any more audience to work off of. I think by now anyone who can read knows where the beef and the bull[/] are.

The Q66 works over Rhine valley farmland inbound to Frankfurt as well as cities inbound to JFK. http://forums.avsim.net/user_files/194657.jpgAnd FYI, if ya stay in that left turn, your gonna cross the ILS to the 13s and ATC gonna want to talk to ya on the land line. I think ATC citation would read "uncontrolled maneuver in a controlled airspace." But just so we don't air-to air, look up for a minute. I'll flash my landing lights! - Just trying to relate all this back actually flying airplanes. Isn't that the real goal?

Hey... great shot!bring that view angle up so we can see a bit further out than 2-3 miles.. you know, 2 miles out where the autogen starts to fade out in the top of the shot and the ground turns to mud?How are those stuttering frames?Gotta be at least 10!Thats great you can angle in a clean and sharp closeup. Being a primary designer for GEX EU and knowing exactly how much default autogen displays in the Rhineland Farm rurals per slider tick I think its just a plain a miracle how that scenery and autogen slider grew with that Q6600 from the farmland to NYCI would think one could run at least 10x sharper with absolutely no blur at least to the first 1/3rd or center of the shot in Rhineland farms with little or no AG showing but gosh darn it... I must be wrongI should have stayed with the P35 DDR2Im such a dopeImages like that prove me completely wrong@ j/k.. thanks for that... happy holidays :-beerchug

Agreed. And it's the holiday season too and I need to be more...cheery. Deck the halls & all that. :-beerchug I will seriously try to curtail my comments from now on. I realize it doesn't help the original poster's question. Maybe smash my hand with a hammer when I read something I don't like :-lol -jk

I should have said something before I hijacked your thread, sorry. Just to add my humble opinion...GO GET THAT 940! I have the Q9650 and it's a great chip. But I'd gladly trade up to a 940 & rampage II extreme (I have the DDR2 rampage now) if I had the money. If you do decide to go core2 than just don't get the C1 stepping Q9550, get the E0.The choice between the Q9650 and 940 i7 would be a no-brainer. Same price, 30% better performance. And FSX appears to LOVE the i7.-jk

Hi Nick.I am not sure why the sudden burst of hostility?My post was simple and to the point "clock per clock". I never argued the point of who can be over clocked more or get a better performance out of. I've looked at the block diagrams, and the App. notes and I could see no real difference on clock per clock, especially when you have to live with the restrictions / boundaries of the Buss that we are dealing with.I would gladly look at your explanation on how you think they are saving cycles / instruction and it translates / propagates all the way to the system's trough put, otherwise I think it's best to move on. You have an opinion, I am not trying to change it, I am trying to understand it. Maybe JK, to whom the question was originally directed, has some real / quantifiable data to support his statement. Something we can duplicate? TV

I think you read what I posted the wrong wayI was not being hostile in answering you! I thought the best 'beef' is in the real world result :)The answer is right here for the first part