I5 2500K Vs. RAM frequency and latency

[idahosurge 2]

I see what you mean now but I don't have the 2000MHz 9-10-9-27 since they were my friends rams, he came over one day for me to test his rams with memtest cause he was building a new rig and he did want to make sure they were good.They did pass memtest so I tested them with FSX also, that's where I saw a difference.

Alain, you could just get into your BIOS and change your ram timings to 9-10-9-27, run the test, get back into BIOS and change them back to your 7-7-6-20 and run the test. I do not think that you necessarily need the ram that you borrowed, just try yours at 9-10-9-27.

[whitav8 98]

First off, could we please pin this thread - it should be renamed "Terrific Stutter Analysis"! I haven't seen any better math/statistical analysis on our FSX performance - great work you guys! Make sure you also run through the thread in this AVSIM forum topic "Sandy Bridge Benchmark Testing Results" because several folks discussed their FSXMark07 results there and one of them from HLJames focused on memory speed - here is it in summary:FSXMark07, Global High Cfg:Core i7 2600 K @ 4.8 GHz1333 MHz 79 fps avg1600 MHz 82 fps avg1866 MHz 83.7 fps avg2133 MHz 86.0 fps avg Core i7 2600 K @ 5.2 GHz (increased CPU speed)2133 MHz 90.873 fps avgThis data shows a measurable improvement versus memory speed - it doesn't show the timings (CLxx) though. FSXMark07 is a different load on the CPU from what you have been using perhaps. In general, the results for FSXMark07 are about twice that of FSXMark11.

[whitav8 98]

Another thing to try here, especially seeing how wide (5->10 fps) the bell curve is for the results so far, is to see what happens to the overall results when you either limit the frame rate by the external limiter addon and/or by using Vsync=ON and setting the FSX cpu load (as it seems to be already) so that most of the frames could be at least 30 fps which is a typical sync point for a 60Hz monitor (60,30,20,15). Many have reported that although the fps were limited, the system had less stutters. Hopefully, this would tighten the bell curve considerably. I realize that this might get away from the original concept of this thread which had to do with varying memory speeds, but what you all have shown here is some remarkable analysis tools in the depiction of frame rate deviation.

[dazz 69]

Another thing to try here, especially seeing how wide (5->10 fps) the bell curve is for the results so far, is to see what happens to the overall results when you either limit the frame rate by the external limiter addon and/or by using Vsync=ON and setting the FSX cpu load (as it seems to be already) so that most of the frames could be at least 30 fps which is a typical sync point for a 60Hz monitor (60,30,20,15). Many have reported that although the fps were limited, the system had less stutters. Hopefully, this would tighten the bell curve considerably. I realize that this might get away from the original concept of this thread which had to do with varying memory speeds, but what you all have shown here is some remarkable analysis tools in the depiction of frame rate deviation.

I tried limiting frames ingame and it made things worse. I'll try the external limiter and switching Vsync ON

[whitav8 98]

@David Roch, Can't we PLEASE pin this thread - it is one of the best analyses - with nice Excel spreadsheets that can be downloaded - of frametime evaluation!!!!!!!!!!!! I hate to lose this to time.Dave

[simon747 185]

Hey guys,I have a 2600k @ 4.7 with a GTX 580 SC and am loving it in FSX. I came from a 920 @3.8 with a 285 GTX. In all aspects, as discussed here with other users on this forum, the sim runs soooooo much better than I could of imagined and much better than my old set-up. So very happy indeed.The only area that bothered me in the new system was as the fps are so much higher overall with this set-up, when I went to JFK for example with 100% ai, fsdt, JFK aerosoft mann., all sliders maxed, maxed weather etc etc the fps would drop to 20 sometimes (not all) and I noticed small micro stutters that were not present on my 920 in fps <20 fps situations. So, I swtitched my ram from 8GB 2000 Corsair xms3 running at 1866 mhz cls 9-10-9-27 to 4gb mushkin radioactive 1600 mhz 6-8-6-24 (lowest cls setting I could find) and voila microstutters gone. FPS are the same, but MUCH smoother for me on the low end fps situations where the cls 9 ram would cause micro stutters. The sim is much smoother overall despite the bandwith drop from 1866 to 1600mhz and going from 8 to 4 gig:)I don't think it is a placebo effect as there is no real way to measure this besides naked eye, but the smoothness has improved greatly on low end maxed-heavy situations. For some reason FSX likes low latency ram of 1600mhz on high oc's and is a much much smoother overall set-up for me now+:)) Weird game indeed as most games cas latency makes no difference!

[dazz 69]

Do the stutters come back if you set your new mushkins to 1600MHz 9-10-9-27 Simon?

[simon747 185]

Hey Dazz,Good thing I am off today:)Good idea. I went back to bios and set timings to the old 9-10-9-27. Fired up the sim went to my lowest fps test spot JFK all maxed as above, and yes I am happy to say there were much more miro stutters at cas 9 vs cas 6 only evident when the fps were low at 20fps or the occ. dips below. The cas 6 setting smooths out the microstutters in the above scenario for me. Now myvboot issues are fixed (i hope) and the low end stutters are gone, it seems like a nice cool well balanced system. That was the only complaint I had about the switch going to Sb were the more evident mirostutters in the worst case scenario fps and the cold boot which the b3 and new gig ud5 bios updates seem to have fixed... This has fixed it for me. I am not missing the 8 gb either as I only use one app at once on this system. Overall though the fps are so high that mirostutters don't exist on sb, but when using all add-ons at once with all bells and whistles on at jFK, LHR etc etc, the cas 6 stuff seems to do the trick on my set-up.It is nice on SB that memory can be tweaked seperatly on an oc system. The 920 was not as easy to alter memory settings whatsoever.

[Kosta 122]

I have a 2133Mhz CL9 G.Skill F3-17000CL9D-8GBXLD sitting here, waiting to be returned to Amazon.I didn't make a test yet, but compared to my current Corsair CMX8GX3M2A2000C9 1866Mhz CL9, would it be worthy to keep? Price G.Skill is 50€ more.

[simon747 185]

I am not a real memory guru, but the bandwith of the 2133 may help over 1866? Doubt you will see much diff between cas 9 cas 8 at those speeds though. Each set-up is different, we all fly with different add-ons etc. Some people have issues running at 2133, others do not. For me, the lower cas of the mushkins compared to the 1866 xms 3 cas 9's made a big impact on microstutters that were present on heavy low fps areas, smoothed out the low end considerably for me so the whole sim is more fluid. FPS are always in the 35-40 big cities and airports and ORD and JFK can get in to the 20's where this new ram helped.

[SpiritFlyer 987]

Hi Simon,A part of me has always felt that there was something about this machine that gave me smoothness and fluidity more than most, and suspected it was the Mushkins. I too do not have the microstutters of others. The only real difference seems to be the very low cas. FSX is indeed a strange beast. Thanks for bringing this subject up.MushkinKind regards,Edit: I never tried testing it, but what do you fellas think might be the difference in FSX performance between 8g or 4g of Mushkin 6-8-6-24 with high end Sandy Bridge systems?

[simon747 185]

Hi Simon,A part of me has always felt that there was something about this machine that gave me smoothness and fluidity more than most, and suspected it was the Mushkins. I too do not have the microstutters of others. The only real difference seems to be the very low cas. FSX is indeed a strange beast. Thanks for bringing this subject up.MushkinKind regards,Edit: I never tried testing it, but what do you fellas think might be the difference in FSX performance between 8g or 4g of Mushkin 6-8-6-24 with high end Sandy Bridge systems?

Hi Stephen,Yes the cas 6 helped with low-end stutters. So I am happy:) Noticed no difference going to just 4 gig, but my pc is just for FSX so only have one open app going at a time. Not sure if you would get a noticed difference with 8 gig, but you never know:)Simon

[dazz 69]

Simon, what A/C and addons are you using? I can't for the life of me reproduce it

[Gary A 8]

Hi Simon,A part of me has always felt that there was something about this machine that gave me smoothness and fluidity more than most, and suspected it was the Mushkins. I too do not have the microstutters of others. The only real difference seems to be the very low cas. FSX is indeed a strange beast. Thanks for bringing this subject up.

Hi Stephen, could you please confirm that you are running 8GB RAM and your timings are as posted in your signature please. I may upgrade my RAM at some point but want validation that 8GB will run @ 6 8 6 24

[SpiritFlyer 987]

Hi Stephen, could you please confirm that you are running 8GB RAM and your timings are as posted in your signature please. I may upgrade my RAM at some point but want validation that 8GB will run @ 6 8 6 24

Hi Gary,Yes, I can confirm that. I run 4 sticks of 2g Mushkin Redline 1600MHz @ 1600 @ 6-8-6-24 at 1.65v without any kind of problem.Kind regards,