December 14, 200619 yr I currently am using an Athlon64 3800 single core (2.4Ghz) with a PNY Nvidia 7900GS. I am using 2 sticks of 512Mb DDR400 speed memory and want to upgrade to 2Gb of RAM. I understand that in some situations, there is a disadvantage of using four sticks of memory 512x4 for 2Gb over 1024x2 for 2Gb. How much of an actual performance hit will I take in FS9 by saving money and upgrading using 4x512 over having to purchase two entirely new 1024Gb sticks of RAM?Do memory timings matter in FS9?Thanks for any help in clarrifying this for me.RH
December 15, 200619 yr You can try before you buy. Set you memory speed to DDR333 and see if you get any substantial performance decrease in a saved scenerio or two (preferably complex) that you have benchmarked before you reduce your memory speed. In my experience it makes neglible performance difference (<5%), but you need to see this for yourself.Gary 9800X3D | 4090 | 64GB | 2+1TB NVME | 2TB SSD | 2TB HDD | 85/50/43” TVs | Quest 3 | DOF H3 Motion Rig | Buttkicker | T.16000M Flight Kit MSFS @ 4K Ultra DLSS Performance FG 80 FPS | VR VDXR Godlike 80Hz SSW | MSFS VR DLSS Quality, Ultra Preset - Windows 11 Acer Nitro 5 | i5-11400H | RTX 3060 6 GB | 32GB DDR4 | 15.6" FHD IPS 144Hz | 2 x 512 GB SSD | Windows 11
December 15, 200619 yr Having 4x512 does not necessarily mean your RAM speed will be lower but that the DDR Command Rate will default to 2T instead of the optimal 1T.Command Rate in A64s has a profound impact on memory intensive applications so people try to make sure they run at 1T, that is only one pair of DIMMS. If I remember correctly one of the members of ACES had insinuated that FS being a memory intensive piece of software profits from low memory latencies.Having taken a similar upgrade path myself i would advise you to sell your 2 DIMMS and buy 2 larger ones. At worse it will keep you from agonising over the possibility of poor RAM performance ;)Regards,Vassilis
December 15, 200619 yr It would be more agonising to upgrade without having done the simple test I suggested, then find out later it made neglible difference.I have done this very upgrade on my son's A64 computer, going from 2 x 512M @ DDR 400 and 1T to 4 x 512M @ DDR 333 and 2T. Sure, his sisoft sandra benchmarks went down by 30%, but there was neglible performance impact to FS.Gary 9800X3D | 4090 | 64GB | 2+1TB NVME | 2TB SSD | 2TB HDD | 85/50/43” TVs | Quest 3 | DOF H3 Motion Rig | Buttkicker | T.16000M Flight Kit MSFS @ 4K Ultra DLSS Performance FG 80 FPS | VR VDXR Godlike 80Hz SSW | MSFS VR DLSS Quality, Ultra Preset - Windows 11 Acer Nitro 5 | i5-11400H | RTX 3060 6 GB | 32GB DDR4 | 15.6" FHD IPS 144Hz | 2 x 512 GB SSD | Windows 11
December 15, 200619 yr If I might add to this:I have experimented with 1T vs. 2T in FS(x), and the 1T command rate along with 2.5-3-3-8 gave an 8% boost to frames.I need to change my sig, because I'm running 2.5-3-3-8 stably now.RhettAMD 3700+ (@2530 mhz), eVGA 7800GT 256 (Guru3D 93.71), ASUS A8N-E, PC Power 510 SLI, 2 GB Corsair XMS 3-3-3-8, WD 250 gig 7200 rpm SATA2, CoolerMaster Praetorian Rhett 7800X3D ♣ 96 GB G.Skill Flare ♣ Gigabyte 4090 ♣ Crucial P5 Plus 2TB
December 15, 200619 yr Hi Rhett, those 8% are what 'the books' suggest. Well done! :-) @ Gary, it would be interesting to learn with which sim the observations you described were made? FS9? 4x1GB/400 @ 1T BSODs over here, no chance. :-( I haven't tried hard though. @ Robbie, you got the picture, not much I can add apart from a summary: 2x1GB = 1T & 400MHz; 4x512MB = 2T & 400MHz (400MHz shouldn't be a prob). Delta performance between 1T & 2T is ca 8%. That's hardly noticeable and likely only measureable. Is the money worth the extra frame @ 12fps? Well, you decide... :-) Cheers and kind regards Jaap
December 16, 200619 yr In addition 2T might allow to sqeeze some more Mhz out of the DIMMs to help offset the performance loss.Regards,Vassilis
December 16, 200619 yr Hi Vassilis, assuming it's so, I wonder whether i.e. 216MHz would be sufficient to offset the difference 'according to the books'? Wadda ya think? Cheers and kind regards Jaap
December 16, 200619 yr Jaap,I meant I was comparing different speed settings of the 2x512M sticks before I put the other two in because, as you say, 4x512M @ 1T (or even 400MHz) doesn't work on my son's computer. Gary 9800X3D | 4090 | 64GB | 2+1TB NVME | 2TB SSD | 2TB HDD | 85/50/43” TVs | Quest 3 | DOF H3 Motion Rig | Buttkicker | T.16000M Flight Kit MSFS @ 4K Ultra DLSS Performance FG 80 FPS | VR VDXR Godlike 80Hz SSW | MSFS VR DLSS Quality, Ultra Preset - Windows 11 Acer Nitro 5 | i5-11400H | RTX 3060 6 GB | 32GB DDR4 | 15.6" FHD IPS 144Hz | 2 x 512 GB SSD | Windows 11
December 16, 200619 yr 216 aka 8% right? Sorry Jaap I am too sloppy to bench such things :) However,most DIMMS can probably do more than 230 at 2T with a bit of Vdimm increase. If yours could climb to that you probably would have offset your 2T losses I guess. OTOH, My cheap as chips DIMMs are even good for 240@3-4-4-8/1T:-wedge Regards,Vassilis
December 17, 200619 yr You know, speaking of this, I was running 230x11 = 2530 at 3-3-3-8 2T with stock voltages on the ram.Now I am running 220x11 = 2420 at 2.5-3-3-8 1T with stock voltages on the ram. This gives me an 8% boost to fps in FSX.I have not tried upping it from 220 back up to 230 where I used to be running when I was at 2T.RhettAMD 3700+ (@2530 mhz), eVGA 7800GT 256 (Guru3D 93.71), ASUS A8N-E, PC Power 510 SLI, 2 GB Corsair XMS 3-3-3-8, WD 250 gig 7200 rpm SATA2, CoolerMaster Praetorian Rhett 7800X3D ♣ 96 GB G.Skill Flare ♣ Gigabyte 4090 ♣ Crucial P5 Plus 2TB
December 17, 200619 yr Rhett if your CPU is good for 2530 why settle for 2420? Even if your memory can't climb up there you could run it on a divider. I am currently doing 270x10 on a 166 divider i.e. CPU@2700 and DDR@225 on a lowly x2 3800.Regards,Vassilis
Create an account or sign in to comment