September 15, 20169 yr Commercial Member Guessing? 1365=01,01,01,01,01,01 this is taking one LP from each core pair of the six cores. There are not twelve cores only six. So this 1365 AM uses all six cores without sharing them to multiple sim processes. The sim makes six parts. With the four core +HT we can make Six parts with 245=11,11,01,01 and four with 85=01,01,01,01. All cores with HT disabled are single Logical Processors and with HT enabled there are two LPs per core. Those two LPs are still only one physical core. Steve Waite: Engineer at codelegend.com
September 15, 20169 yr Guessing? 1365=01,01,01,01,01,01 this is taking one LP from each core pair of the six cores. There are not twelve cores only six. So this 1365 AM uses all six cores without sharing them to multiple sim processes. The sim makes six parts. With the four core +HT we can make Six parts with 245=11,11,01,01 and four with 85=01,01,01,01. All cores with HT disabled are single Logical Processors and with HT enabled there are two LPs per core. Those two LPs are still only one physical core. Confused, as if you type 101 0101 0101 (11 digits) in the binary calculator you end up with 1365, by cores i meant logical processes/cores.. so wouldnt 1365 represent 11 of 12 logical with only 5 turned on? Similarly with 85 (7 of 8) Asus Strix z790-e; 1000 watt evga SuperNova Plat; 14900k AC_LL 0.55 adp -0.050 253/253/355 CEPoff (CB-1pass 39200 80c, msfs peak 92,avg 60-78c, astrorender 95c,room76F); 64GB(dual 32) cl32 6400 at 6400 xmpII F5-6400J3239G32GX2-TZ5RK, Asus Ryuo III 360mm; Thermaltake v51 Case; Gigabyte 4090 OC; VR-Crystal; Dofreality H6; Astrosite
September 15, 20169 yr Commercial Member Very. You are actually saying you are only accounting for 11 LPs because there's no zero in front of binary 10101010101? Steve Waite: Engineer at codelegend.com
September 15, 20169 yr Hi, When you type in 1365 decimal to the calculator 1,01,01,01,01,01 binary will be displayed (note I have inserted the commas for clarity showing the six cores). However, you need to remember that core 5 (first core is core 0 with logical processors 0 and 1 reading from right to left), is only shown as 1 LP when you should be assuming 2. The second LP of core 5 has not been flipped to binary '1' and therefore would be '0', but is not shown in the result by the calculator. So, in fact the result for 1365 decimal is 01,01,01,01,01,01 binary. Using Affinity Mask = 1365, logical processors 0, 2, 4, 6, 8 and 10 of your 6-Cored CPU are selected for P3D to run on. Mike
September 16, 20169 yr Hi, When you type in 1365 decimal to the calculator 1,01,01,01,01,01 binary will be displayed (note I have inserted the commas for clarity showing the six cores). However, you need to remember that core 5 (first core is core 0 with logical processors 0 and 1 reading from right to left), is only shown as 1 LP when you should be assuming 2. The second LP of core 5 has not been flipped to binary '1' and therefore would be '0', but is not shown in the result by the calculator. So, in fact the result for 1365 decimal is 01,01,01,01,01,01 binary. Using Affinity Mask = 1365, logical processors 0, 2, 4, 6, 8 and 10 of your 6-Cored CPU are selected for P3D to run on. Mike Yeah that was the ticket, that first core/entry.. thanks for the clarification, makes sense. Asus Strix z790-e; 1000 watt evga SuperNova Plat; 14900k AC_LL 0.55 adp -0.050 253/253/355 CEPoff (CB-1pass 39200 80c, msfs peak 92,avg 60-78c, astrorender 95c,room76F); 64GB(dual 32) cl32 6400 at 6400 xmpII F5-6400J3239G32GX2-TZ5RK, Asus Ryuo III 360mm; Thermaltake v51 Case; Gigabyte 4090 OC; VR-Crystal; Dofreality H6; Astrosite
September 17, 20169 yr Commercial Member In the binary representation, last core is on the left, first core (core zero) is on the right. Steve Waite: Engineer at codelegend.com
October 2, 20169 yr HI SteveW Since the last update from LM, i ve noticed a difference with AM setting i normally use am=116, but now with this setting p3d give me a lot of blurry during en route so i need to add core 0 in the process windows task, plus the vas growing more than version 3.3.5 now when i finish a flight 1h30, vas is 500mb, before it was 1000 or more, maybe you are aware of something about certain change made by LM in their core process, maybe have you got a solution for this problem? Thanks Frédéric Giraud
October 2, 20169 yr Given that 1365 has worked well with my i7-4930k hex.. what have others found to be a good value for the i7-6700k? (the search option for this thread seems to be missing, at least in my browser?) edit: i think its 252 based on the calculator in windows.. is this a good one, have others found it to be the best? edit2: NO actually, 1365 is 11 cores alternating, not 12.. not sure how/why that one worked best on the 12 core.. so i guess the equivalent 8 core might be 85 (7 cores alternating), which is in perfect sync with the OP first post and findings.. ? Did you try 340 too ? I have tried 1365 , 4085 , but at the end I always end up with 340 ( 5820k 6-core ) 5950x3d 5.4-5.7 GHz - Asus ROG 870 Crosshair Apex - GSkill Neo 2x 24 Gb 6000 mhz / cas 26 - MSI RTX 5090 Gaming Trio OC - 1x SSD M2 6000 2TB - 1x SSD M2 2800/1800 1Tb - Corsair 5400 case - Corsair 360 liquid cooling set - 3x 75’ TCL tv. 13600 6 cores @ 5.1 GHz / 8 cores @ 4.0 GHz (hypterthreading on) - Asus ROG Strix Gaming D - GSkill Trident 4x Gb 3200 MHz cas 15 - Asus TUF RTX 4080 16 Gb - 1x SSD M2 2800/1800 2TB - 2x Sata 600 SSD 500 Mb - Corsair D4000 Airflow case - NXT Krajen Z63 AIO liquide cooling - FOV : 200 degrees My flightsim vids : https://www.youtube.com/user/fswidesim/videos?shelf_id=0&sort=dd&view=0
October 2, 20169 yr Did you try 340 too ? I have tried 1365 , 4085 , but at the end I always end up with 340 ( 5820k 6-core ) 340 even with hyperthreading enabled?
October 3, 20169 yr 340 is with HT enabled Quote from Steve "With 340 core 0 (LPs 0 and 1) and core 5 (LPs 10 and 11) will not be used by the sim. The sim uses cores 1 and 2 do the main work of the sim, cores 3 and 4 to do the main scenery loading. 00,01,01,01,01,00 = 340 xx,x0,x0,00,00,00 = 7,9,10,11 available for addon exes and leaves an entire core free to the system (core 0) xx,x0,x0,00,00,xx = 0,1,7,9,10,11 available for addon exes can try using core 0 as well if you have a lot of exes. if you have a very aggressive exe addon can also try on core 5 use two LPs 10,11, or even more bandwidth core 0 and core 5 LP 0 and 10 " 5950x3d 5.4-5.7 GHz - Asus ROG 870 Crosshair Apex - GSkill Neo 2x 24 Gb 6000 mhz / cas 26 - MSI RTX 5090 Gaming Trio OC - 1x SSD M2 6000 2TB - 1x SSD M2 2800/1800 1Tb - Corsair 5400 case - Corsair 360 liquid cooling set - 3x 75’ TCL tv. 13600 6 cores @ 5.1 GHz / 8 cores @ 4.0 GHz (hypterthreading on) - Asus ROG Strix Gaming D - GSkill Trident 4x Gb 3200 MHz cas 15 - Asus TUF RTX 4080 16 Gb - 1x SSD M2 2800/1800 2TB - 2x Sata 600 SSD 500 Mb - Corsair D4000 Airflow case - NXT Krajen Z63 AIO liquide cooling - FOV : 200 degrees My flightsim vids : https://www.youtube.com/user/fswidesim/videos?shelf_id=0&sort=dd&view=0
October 3, 20169 yr Did you try 340 too ? I have tried 1365 , 4085 , but at the end I always end up with 340 ( 5820k 6-core ) I'm now on a (silicon lottery) 4.7 6700k with hyperthreading enabled and using AM of 85 as of now.. Side note: I did a comparison of the 4930k and 6700k in this thread in terms of performance. Asus Strix z790-e; 1000 watt evga SuperNova Plat; 14900k AC_LL 0.55 adp -0.050 253/253/355 CEPoff (CB-1pass 39200 80c, msfs peak 92,avg 60-78c, astrorender 95c,room76F); 64GB(dual 32) cl32 6400 at 6400 xmpII F5-6400J3239G32GX2-TZ5RK, Asus Ryuo III 360mm; Thermaltake v51 Case; Gigabyte 4090 OC; VR-Crystal; Dofreality H6; Astrosite
October 3, 20169 yr With a 6700 you cannot use 340 as that is for a hex core. Why didn't you buy a 6800 hex core or do you run not much addons ? 5950x3d 5.4-5.7 GHz - Asus ROG 870 Crosshair Apex - GSkill Neo 2x 24 Gb 6000 mhz / cas 26 - MSI RTX 5090 Gaming Trio OC - 1x SSD M2 6000 2TB - 1x SSD M2 2800/1800 1Tb - Corsair 5400 case - Corsair 360 liquid cooling set - 3x 75’ TCL tv. 13600 6 cores @ 5.1 GHz / 8 cores @ 4.0 GHz (hypterthreading on) - Asus ROG Strix Gaming D - GSkill Trident 4x Gb 3200 MHz cas 15 - Asus TUF RTX 4080 16 Gb - 1x SSD M2 2800/1800 2TB - 2x Sata 600 SSD 500 Mb - Corsair D4000 Airflow case - NXT Krajen Z63 AIO liquide cooling - FOV : 200 degrees My flightsim vids : https://www.youtube.com/user/fswidesim/videos?shelf_id=0&sort=dd&view=0
October 3, 20169 yr Commercial Member Those new hex cores are quick, and moving addon exe activity away from the cores running the sim can help more than expected even with less aggressive addons. Steve Waite: Engineer at codelegend.com
October 3, 20169 yr Those new hex cores are quick, and moving addon exe activity away from the cores running the sim can help more than expected even with less aggressive addons. My direct comparison between the 4930k and 6700k even with a ton of addons showed an improvement going back to 8 logical (HT).. i basically went from 32 fps 4930k to 49 fps on the 6700k at 4.7 with ddr4 at 3200.. mainly because i was able to get the clock rate up to 4.7 ghz and crushed my old 4930k passmark cpumark as well. Granted, i realize having the extra cores would benefit or at least in theory.. but going to DDR4 was also a huge boost (at 2900+ on the ram clocks). Perhaps the 6800k + ddr4 would have crushed these results, but i'm happy with the change so far, unfortunately i didnt do a comparison on that. And i run with AM=85 so far with these results. Asus Strix z790-e; 1000 watt evga SuperNova Plat; 14900k AC_LL 0.55 adp -0.050 253/253/355 CEPoff (CB-1pass 39200 80c, msfs peak 92,avg 60-78c, astrorender 95c,room76F); 64GB(dual 32) cl32 6400 at 6400 xmpII F5-6400J3239G32GX2-TZ5RK, Asus Ryuo III 360mm; Thermaltake v51 Case; Gigabyte 4090 OC; VR-Crystal; Dofreality H6; Astrosite
October 3, 20169 yr Commercial Member Agreed, I think the latest RAM and motherboards are making great improvements. Comparing the 4930 vs 6700 is like old dog vs young whippet. But the new six cores could take over from the old fours, and they show a faster turbo than the eight and ten due to heat. Steve Waite: Engineer at codelegend.com
Archived
This topic is now archived and is closed to further replies.