I just ran a dynamic comparison of dual versus single core (using AffinityMask) performance with SP1, and found FSX performance was actually WORSE with 2 cores versus one. :-eekHere are the FRAPS results from a 100 second recording I did around TNCM with all sliders right and checked (except bloom off) at 3360x1050 resolution to really max out the workload:Single core (AffinityMask=2):2007-05-19 11:23:58 - fsxFrames: 1537 - Time: 100000ms - Avg: 15.370 - Min: 12 - Max: 252007-05-19 11:26:33 - fsxFrames: 1524 - Time: 100000ms - Avg: 15.240 - Min: 12 - Max: 25Dual core (AffinityMask=3):2007-05-19 11:30:26 - fsxFrames: 1319 - Time: 100000ms - Avg: 13.190 - Min: 11 - Max: 202007-05-19 11:33:05 - fsxFrames: 1329 - Time: 100000ms - Avg: 13.290 - Min: 11 - Max: 21 ie. FSX + SP1 on single core is nearly 16% faster than with dual core.And in case anyone is thinking that perhaps both cores weren't getting a good workout in this scenario, see below for what task manager showed at the end of the second test for dual core.http://forums.avsim.net/user_files/171962.jpgZoomed in so you can read it:http://forums.avsim.net/user_files/171963.jpgLooks like we got our multi-core support with SP1 alright but what, apart from slowing things down, is it actually doing for us? Can anyone else replicate this behaviour?GaryPS. As a sidenote, this is pretty impressive performance compared to RTM regardless of how many cores are being used. I would have been lucky see over 10 FPS, let alone average it with these settings before. I'm just curious as to why multi-core actually takes us a bit backwards from this otherwise great leap forwards. :-hmmm

AffinityMask=2 is simply forcing every thread to use core #2, Gary.What are your results usingAffinityMask=1which means "run all threads on core #1"?Interesting, yet very puzzling results nonetheless!

Yes I know that AffinityMask=2 makes FSX core 2. The reason I like using core 2 is that the little fester hardly ever gets to take on the full load and that, more importantly, it is very easy to confirm that FSX is running purely on core 2, because everything else tends to default to core 1. I did some other tests last night using core 1 and core 2 individually, and the result was exactly the same. Nonetheless, I will do another couple of runs with core 1 only and post back here.Gary

Okay, I ran the test with just core one and got the following result:Core 1 only:2007-05-19 12:26:03 - fsxFrames: 1432 - Time: 100000ms - Avg: 14.320 - Min: 11 - Max: 232007-05-19 12:28:35 - fsxFrames: 1434 - Time: 100000ms - Avg: 14.340 - Min: 11 - Max: 22This puts core 2 as the winner, then core 1, then core 1 and 2, with about 1 FPS average between each position.One thing I did notice was that my FPS were locked at 30, which is a bit of a no no when benchmarking, however since the highest frame rate I was getting was 25 I didn't think it would affect the results very much.Was I wrong! After unlocking the FPS the results immediately boosted across the board, and changing the core performance pecking order. The results were:Core 1 only:2007-05-19 13:00:46 - fsxFrames: 1695 - Time: 100000ms - Avg: 16.950 - Min: 13 - Max: 262007-05-19 13:03:10 - fsxFrames: 1711 - Time: 100000ms - Avg: 17.110 - Min: 13 - Max: 27Core 2 only:2007-05-19 12:54:04 - fsxFrames: 1797 - Time: 100000ms - Avg: 17.970 - Min: 14 - Max: 292007-05-19 12:56:32 - fsxFrames: 1802 - Time: 100000ms - Avg: 18.020 - Min: 14 - Max: 28Cores 1 and 2 together:2007-05-19 12:47:11 - fsxFrames: 1824 - Time: 100000ms - Avg: 18.240 - Min: 14 - Max: 292007-05-19 12:49:50 - fsxFrames: 1812 - Time: 100000ms - Avg: 18.120 - Min: 14 - Max: 29Now dual core is in the lead, but only by 1% over core 2 and core 1 is the slowpoke.Whilst these new results obviously extinguishs my original concern that dual core was actually slower, they still beg the question as to why is there only a neglible gain (1%) for all the extra effort (roughly 70% extra CPU use) that the second core is chiming in? Perhaps 100 seconds isn't enough to fully appreciate the pre-processing that the second core is probably doing for scenery I have yet to encounter??Probably the biggest thing I have got out of all this is how much the frame rate goes down by locking FPS, even if youR FPS are already under the cap you set. I did a quick test at the end of the last test when paused and found that unlocked my FPS hovered around 21 and located at any value over 20 (even 99), the FPS went down to 17 (ie. nearly 20% slower.) No more FPS lock for me and I might keep that second core running afterall!Gary

So you think dual core machines are a waste of money?

Not if the second core is doing something constructive. One of these days I am going to create a longer distance and time benchmark (say 15+ mins in a jet crossing over a city then off across a mountain range) to make sure the second core gets a chance to show its wares.Having said that, I wouldn't like to go back to single core from dual core for general computer use, as I find windows and apps are so much more responsive with two cores running. I still spend quite a bit of time flying FS9 and I can do lots of other things while it consumes one of the cores. When I had a single core system, I often had to minimise FS to get any decent performance out of something else I was running.Gary

>Not if the second core is doing something constructive. One>of these days I am going to create a longer distance and time>benchmark (say 15+ mins in a jet crossing over a city then off>across a mountain range) to make sure the second core gets a>chance to show its wares.>>Having said that, I wouldn't like to go back to single core>from dual core for general computer use, as I find windows and>apps are so much more responsive with two cores running. I>still spend quite a bit of time flying FS9 and I can do lots>of other things while it consumes one of the cores. When I>had a single core system, I often had to minimise FS to get>any decent performance out of something else I was running.>>GaryLol indeed interesting going to see if I have the same results... Andr

"1" for 1 core"3" for 2 coresAm I wrong? I don't think affinity = 2 should ever be used.It looks like from Phils explanation on his blog, this is a bitmask value.1 = 001 hex (only use core 1)2 = 010 hex (only use core 2) (I think this is essentially equivalent to 1 on a dual core)3 = 011 hex (use core 1 and 2)15 = 1111 hex (use cores 1 - 4)

Hi Gary,I have a Core 2 Duo, but when I hit ctrl alt del, and click the performance tab, I have only one window where it says CPU usage history, though everything is running fast as lightning, I wonder if I am supposed to have two windows there, one for each core or not?Jeff

I definitely have 2. Seems like you should too.

Interesting take on the FPS locking issue I find I get the best results by locking FPS just above what is achievable during normal flight ops. I currently have mine locked at 40 because I don't see any activity above 35 with unlimited lock. I have also found that if I set my FPS lock to an achievable figure such as 20 or 25 then I never seen a higher frame rate during flight than about 5 FPS below the locked setting. John

>Hi Gary,>>I have a Core 2 Duo, but when I hit ctrl alt del, and click>the performance tab, I have only one window where it says CPU>usage history, though everything is running fast as lightning,>I wonder if I am supposed to have two windows there, one for>each core or not?>>JeffJeff with C2D you should see at least two lol I'm running a E6600 @ 2,7Cheers,Andr

Yeah, this is weird.Jeffhttp://forums.avsim.net/user_files/171989.jpg

>"1" for 1 core>"3" for 2 cores>>Am I wrong? I don't think affinity = 2 should ever be used.>Why not? That's what a mask does - allows you to pick which specific cores FSX will use. Try the different values and see which cores get used if you want to confirm for yourself.Gary

You need to set the following option in Windows Task Manager:http://forums.avsim.net/user_files/171994.jpgGary